AWS Certified Solutions Architect Associate 2023 (Full Free AWS course!) - Day Two

Aug 16, 2023

AWS Certified Solutions Architect Associate 2023 (Full Free AWS course!) - Day Two

Welcome to our latest Live FREE AWS Bootcamp! In this Livestream series AWS Certified Solutions Architect Associate 2023 (Full Free AWS course!) we provide free AWS certification training, for the AWS solution architect certification (SAA-CO3 exam). Whether you are wondering which aws certifications should i get, or you know what you want this is for you. We offer this AWS certification course free to you as part of our mission. We provide this free aws certification course online to assist individuals in building their cloud architect or cloud engineering careers.\r
\r
Get our FREE eBook and Lab Demos to go along with this http://bit.ly/41fCEJE\r
\r
This AWS cloud computing full course will help with your AWS solution architect certification, build your cloud architect career skills, and is an essential part of your cloud computing career training. That’s why we provide this AWS full course tutorial, and other AWS training free to help the cloud architect community!\r
\r
Please use this FREE AWS course to elevate your cloud computing career, achieve your certification, and ace your AWS solution architect interview questions.\r
\r
If you enjoyed this content from Go Cloud Careers and want to learn how to get a cloud job, please subscribe to the Go Cloud Architects channel.\r
\r
More Content for you!

Learn more about our award-winning training programs at https://training.gocloudcareers.com

Take the Cloud Architect Challenge, May 2nd, sign up here https://bit.ly/3HCTevZ

FREE Webinar, learn how to become a cloud architect https://my.demio.com/ref/01ppStEkIUVx…

Free Interview Webinar, register here https://bit.ly/3IP4B4M

FREE Interview Guide get yours today, http://bit.ly/3KKRVx0

FREE How to Get Your First Architect Job Guide, get yours today, http://bit.ly/41rixJl

Learn about our Elite Cloud Architect Bundle, https://training.gocloudcareers.com/b…

Learn about our How to Earn More Money in Tech Program, https://training.gocloudcareers.com/c…

Sign Up for Our Updates, https://www.gocloudcareers.com/form-o…\r
\r
At Go Cloud Careers and Go Cloud Architects we are focused on helping you be the best at your dream cloud career.\r
\r
Every day we speak with people looking to build their cloud architect careers. Unfortunately, many people are confused by the various cloud computing job roles, such as the difference between a cloud architect vs cloud engineer. This makes it hard to build your cloud architect career development program, which is necessary to get the right cloud computing career training.\r
\r
You can find some more of our content below:\r
\r
What Is A Cloud Architect | Cloud Architect Job (What Does A Cloud Architect Do) -    • What Is A Cloud Architect | Cloud Arc…  \r
\r
Can I Be A Cloud Architect? | Cloud Architect Career Advice (Get Cloud Hired) -    • Can I Be A Cloud Architect? | Cloud A…  \r
\r
What Does A Cloud Architect Do | Cloud Architect Job (Cloud Architect Roles and Responsibilities) -    • What Does A Cloud Architect Do | Clou…  \r
\r
How To Start A Career In Cloud Computing | Tech Career Transition Coaching | Changing Career To Tech -    • How To Start A Career In Cloud Comput…  \r
\r
\r
Our goal at Go Cloud Architects and Go Cloud Careers is to make technical training as accessible as possible to as many people as possible. That is why we create FREE technical content for the community. Make sure to check out our multitude of free content below.\r
\r
FREE CCNA Networking Bootcamp –    • CCNA Networking Bootcamp  \r
\r
FREE AWS Advanced Networking Bootcamp –    • AWS Advanced Networking Bootcamp Nove…  \r
\r
\r
Learn more about us here: https://www.gocloudcareers.com\r
\r
Please follow, like, or subscribe to us on our other platforms: \r
\r
Go Cloud Architects Facebook Page: https://www.facebook.com/gocloudarchi… \r
\r
Mike Gibbs LinkedIn Page: https://www.linkedin.com/in/michael-g…\r
\r
Go Cloud Architects LinkedIn page: https://www.linkedin.com/company/go-c… \r
\r
Twitter: https://twitter.com/Gocloudcareers\r
\r
#awstraining #cloudarchitect #cloudcareer #cloudjob

Content

78.799 -> Hello everyone this is Alonzo once you know how to get your first cloud job please register

83.09 -> for our webinar will teach you everything that you need to know and answer your questions

87.67 -> along the way hope to see you there

166.31 -> Hello everyone this is Alonzo wants to know how to get your first cloud job and please

170.36 -> register for our webinar will teach you everything that you need to know and answer your questions

173.69 -> along the way hope to see you there?

198.92 -> My name is Richard Im cloud hired I can see I am cloud hired that yes, come and join and

299.37 -> get cloud hired.

300.37 -> I cloud tired. I'm cloud hired. I'm cloud hired. Hey. Go Cloud Architect family. I'm

308.5 -> cloud hired. Oh, guys. I'm cloud hired. I'm cloud hired thanks to go cloud architects,

321.039 -> it worked for me a now I’m cloud hired because because of go cloud architect’s program.

327.46 -> I am cloud hired.

329.26 -> I am cloud hired. Thank you, Mike, and the go cloud team

366.699 -> Welcome back, everyone, we're here for day three of the completely free, AWS Certified

373.72 -> Solution Architect Associate 2023 course. My name is Michael Gibbs, and I'm the founder

378.47 -> and CEO of go cloud careers and I'll be your instructor throughout this week.

383.139 -> I have my master producer here is day two, which dated I call it. You said day three,

391.3 -> you are too excited you yourself.

394.32 -> It's day two. Apologies, everyone. There's been a lot of things going on in our world

398.32 -> right now. So even I get my days confused periodically. So day two of the AWS Solution

404.03 -> Architect certification course. You know what some of the things we'll cover here will help

409.879 -> you with some AWS Solution Architect interview questions. And this is a complete AWS full

415.039 -> course tutorial, and free AWS training. And it's a free AWS certification course online.

422.129 -> So with this free AWS course we're going to help you pass the AWS Certified Solution Architect

426.879 -> Associate the exam just to you know, is SAA does SEO three, it used to be the SAA does

433.02 -> co2 But this is a more modern exam, the SAE ACO three, and we're gonna have a lot a lot

437.69 -> of fun here. Now, I want to make sure that you all know some things we're going to do

445.31 -> to help you, I want you to all sign up for the how to become the ultimate cloud architect

451.46 -> or how to get your first cloud architect job webinar. It is tomorrow evening. And we will

456.75 -> tell you everything you need to do to get your first cloud architect job certifications,

462.12 -> maybe three to 5% of what you need to do. And I want you to all get these elite cloud

466.32 -> architect or AWS Solution Architect jobs, or Azure Solution Architect jobs or Google

471.69 -> a Solution Architect jobs, I want to know that you've got the best cloud computing career

475.71 -> because that's what we're all about. So join us tomorrow on the completely completely free,

481.03 -> AWS Certified, I'm sorry, on the completely free how to get your first cloud job. As a

487.09 -> cloud architect, it will be well worth your time, and will be on a platform where we can

491.56 -> speak live and talk to each other so we can make sure we can answer any questions you

495.53 -> have to help you build the best cloud computing career. While we're at it, I want you to completely

502.63 -> download the completely free AWS Certified Solution Architect, associate and professional

508.479 -> ebook and labs, you know, we're going to be focused heavily on the concepts here. Why?

515.36 -> Because you get hired based on the concepts if you're a cloud architect or a Solution

519.32 -> Architect, guess what you design, present and sell, you don't even touch the technology.

523.95 -> But it's more than that. If you're a cloud engineer, you're going to be performance tuning.

529 -> And the challenge is knowing what could do not how to do it how to do it is nothing,

534.029 -> you can go straight to the AWS website. And they've got complete step by step instructions.

538.61 -> So in the labs, which you can download, you can watch on your own time, you can practice

542.76 -> those as well. But guess what, we're going to focus heavy, heavy, heavy on the concepts

548.93 -> because that's what it's going to take to win the interview. And that's also what it's

552.26 -> going to take to be able to know how to do any of these jobs. I've interviewed 1000 AWS

556.97 -> Certified people that took the courses out there, which is why we do ours free. And while

560.55 -> they all knew how to configure, they didn't understand what they were doing and why they

563.49 -> were doing it, which meant they were unemployable. And I went to getting the best jobs, because

567.29 -> we're all about giving getting you all cloud hired. So make sure make sure that you attend

575.53 -> the completely free how to get your first cloud job in America and get the free additional

579.95 -> AWS resources because I want you all hired, passing your certifications and having the

584.16 -> absolute best career. If you miss day one. Don't worry about it. You can catch day one

588.649 -> on our YouTube channel each day stands alone. And I go back and watch it tonight. And enjoy

595.06 -> day one and today we're going to cover day two I know I called it day three but it's

599.209 -> really day two today. I was all excited about all the things we're doing. Because let's

602.649 -> face it, that's just me, I love doing these things. Many people ask us which AWS certifications

608.62 -> to get as a rule, the AWS Certified Solution Architect Associate is the starting point.

614.839 -> Now which certifications you should get is completely dependent upon the career you desire.

618.52 -> But the AWS Certified Solution Architect Associate is a basic intro to cloud computing. And that's

624.47 -> what we're going to do. So let's start talking about the AWS Solution Architect, training,

634.649 -> because this is an AWS cloud computing full course at least as it pertains to passing

638.26 -> the Certified Solution Architect Associate exam. So yesterday, we talked about the orchestration

646.13 -> of the cloud and getting our data to the cloud and storage. But today, we're going to begin

650.82 -> by talking about computing on the cloud. But before we do that, can you guys all give me

656.779 -> a hashtag that says AWS Solution Architect Associate Certification in the chatbox. And

663.519 -> that way, I know you're awake, alert oriented. I'm a medical person, we medical people like

668.96 -> to know that the people are know where they're at. That's why when somebody falls down, we

672.279 -> say, You know what time it is, you know who the president is, you know where you're located.

675.68 -> So I like to know that you're all awake, alert and oriented. This is medicine, we would call

679.76 -> times three, which was my old fun career internal medicine. Before I went into tech 25 years

685.54 -> ago, and let me tell you, I will never go, I've never looked back. And I love Tech Tech

690.64 -> Tech Tech tickets, the greatest thing. So for all you guys that are putting Cisco Cisco

695.31 -> over there, I love that name. I spent about a decade at Cisco, AWS Solution Architect

700.26 -> certification, fantastic. A diverse Solution Architect certification, I love it. I'm seeing

704.089 -> all this AWS Solution Architect certifications out there. So I know you're awake. I know

708.88 -> you're here. And I know you are ready to go. Love it. So fantastic. And if you're a nurse,

718.079 -> you can that's wonderful, wonderful, wonderful. I was a nurse and then a nurse practitioner,

722.089 -> and I was a firefighter paramedic even before that. And health care, people become great

727.3 -> architects, because they're used to asking people questions, and they're used to communicating

730.721 -> with people. And for the nurse that's here. If you ever had to sell your patient into

734.3 -> taking a medication, which I know you have, or following healthy lifestyle guidance, guess

738.821 -> what? We do that with technology all the time as architects, so let's have some fun, and

744.1 -> let's get into the content. So a cloud or cloud computing is nothing more than renting

750.96 -> space in somebody else's networking data center. That's it. It's somebody else's data center,

755.449 -> the red space. And again, what is involved in these things, it's routers and switches

760.529 -> and servers and storage, physical load balancers, firewalls, intrusion detection, intrusion

768.449 -> prevention systems and cabling. That's it. That's what the data center is, and guess

772.511 -> what the cloud is. It's renting space in that. So when we talk about computing, guess what,

778.49 -> we're going to be talking about the same thing, because cloud computing is nothing new. First

783.121 -> cloud I worked on was in 1996. And it hasn't changed very little, even though the marketing

787.35 -> people does. So renting space and somebody else's, even though IKEA, I am thrilled to

791.649 -> see you here along with Lady guitar. So let's talk about computing. In the data center,

800.82 -> we've got these physical servers, everybody, okay? physical servers. And in the cloud,

807.43 -> they still have the physical servers. Now in our data center, typically speaking, we

812.61 -> take these servers, and we either put VMware on it and virtualize the servers. On today's

819.09 -> modern world, we either use Nutanix, which is one of my favorite hybrid cloud solutions,

824.05 -> or we use IBM OpenStack, which is more commonly used than Nutanix. But they're all great private

829.639 -> clouds. And what are these things? Do they enable us to take our servers, virtualized

834.44 -> the servers and reuse them to full capacity and scale up and scale down just like the

840.18 -> cloud? And when you're on the cloud, guess what? You still need servers, even serverless

845.3 -> uses servers, it's a marketing term no matter what you're on servers. So in the data center,

851.46 -> we call it a virtual machine. And guess what, what are we going to be talking about today

855.68 -> on the cloud virtual machines, nothing's new. So if you've used a VMware virtual machine,

861.68 -> or a Microsoft Hyper V virtual machine, or if you've used a Citrix virtual machine, or

869.64 -> if you use the KVM or QEMU virtual machine, guess what we're gonna be using the same identical

874.61 -> stuff with a new name on the cloud, like the husana Meet the new boss, same as the old

879.089 -> boss, it's the same stuff. So AWS is going to call your virtual machines acpl instances,

885.12 -> because that's what they paid their marketing people to do. Google is going to call them

889.27 -> Compute Engine instances. And Azure and Oracle are gonna call them virtual machines, which

893.55 -> is really what they are. They're a little little more you know, by a glass of water,

897.459 -> they call it what are not a hydration system, whereas AWS has to throw the term elastic

901.88 -> and come up with all kinds of funny names for everything. So when you pick virtual machines

907.63 -> on the cloud, how do you size them? The same way, you've done it for the last two decades

914.339 -> in the data center, you size it based upon CPU cores. you size it based upon DRAM. you

923.44 -> size it based upon storage capacity and performance, which are typically going to be using block

928.62 -> storage for like we talked about yesterday, unless it's a dedicated server. And we size

933.24 -> them based upon networking performance. That's it. So if you need to figure out what you

938.681 -> need, how many cores do you need, and how much DRAM do you do? Now I want to make it

942.459 -> very clear. bare metal server performance, like the data center, and cloud server or

948.569 -> any virtual machine performance is not the same. For example, if we use my favorite servers

954.529 -> right now, which are using these beautiful AMD epic processors, and we have a server

960.019 -> that says has 128 cores, and four terabytes of DRAM, that's 128 physical cores, physical

968.44 -> little CPU chips on the CPU. Now on the when we deal with the servers, they can do something

973.71 -> called hyper threading, which is where then the core can split into multiple cores.

979.759 -> Now, you've got a car that can drive at 90 miles an hour, and you split it into two cars,

986.51 -> each car can drive 45 miles an hour, seeing things so in the data center, if you've gotten

992.12 -> 128, physical cores, they split into 256 virtual cores. So when you buy 128 core server, say

1000.139 -> from Dell, or IBM, you're getting 128 physical cores. When you buy 128 core server on the

1007.75 -> cloud, you're getting 120 virtual cores. So it's basically 50% of the performance of your

1013.959 -> actual server. So keep that in the back of your mind. Now, the reality is, in many cases,

1019.58 -> those virtual cores aren't being used. So it's closer to the performance of say, 75%

1023.3 -> of the actual server. But gotta keep that in mind, physical cores in the server and

1029.22 -> virtual cores. So when they're selling you, virtual cores are not exactly the same as

1034.35 -> physical cores. And you're going to have to test and kind of get the performance that

1038.199 -> you actually need. Now, the marketing department decided to take something, basically, that

1045.179 -> is simple, and efficiency, and come up with a bunch of silly names. Look, here's the reality,

1051.42 -> if I need a server, I'm going to basically see which has the right cores, which has the

1055.53 -> right d RAM and which has the right network performance. And I don't care about any of

1058.549 -> these silly letters, I'm going to google it. But you know, AWS has pre made servers specifically

1065.679 -> for pre made uses, which may or may not match your needs. And they basically came based

1070 -> with ARM based workloads, ARM based workloads are great for web servers, they're super low

1074.159 -> power draw. They're not the highest performance thing either. They've got your compute optimized,

1080.02 -> which you can, which you can work with, for example. They've got your G based servers,

1086.64 -> which have GPUs in them, basically use them for machine learning, if you're going to use

1089.669 -> your build your own machine learning environments using say pytorch, TensorFlow or any of the

1094.02 -> machine learning tools that you might use, you've got your eyes for high speed storage,

1097.75 -> you've got your M fi for general purpose to do use it for databases, your M six again,

1102.69 -> or general purpose, but more application servers or gaming servers, you got your R Series,

1107.179 -> which are really memory optimized, when you need to lots of memory like a cache, for example,

1112.12 -> you got your T three, which are basically burstable, which I think they get a little

1115.429 -> performance and you compressed a little more out a little capacity when needed. Test Environment,

1120.45 -> realistically speaking, you got your ex ones, which are really low price for DRAM and you

1125.28 -> need these high servers, huge in memory databases, for example, you need four, six terabytes

1131.03 -> of RAM. These are the systems that you're actually using. So kind of keep that in the

1133.39 -> back of your mind. They pre made these. And they be may or may not be exactly what you're

1139.9 -> looking for. But really, you're going to be sizing no matter what these things say based

1144.11 -> upon CPU, DRAM and network performance. Now, realistically speaking, your traditional virtual

1157.32 -> machines on AWS support Linux on Windows, like things you're going to stick on, whether

1163.88 -> it be Red Hat blonde to their Amazon Linux is an offshoot of Red Hat Linux, just like

1171.08 -> Oracle Linux is an offshoot of Red Hat Linux. The most common operating system on the cloud

1176.14 -> is not Amazon Linux, it's Ubuntu Linux, which is what most people use, but you can use any

1184.591 -> of that you want. Any which your applications are going to perform better. Typically, machine

1187.28 -> learning applications, for example, are better on a button to super stability. Things often

1191.47 -> are better on Red Hat, but it's up to you, your systems administrators to determine what's

1195.64 -> the best operating system for you. For the most part, you can fill the plan I In this

1200.25 -> many Linux servers, but somebody is gonna get closer on the Linux side, the Linux engineer

1205.59 -> is gonna help you select that. Now, AWS also has an easy to instance that runs the Mac

1212.67 -> operating system, they call it mission critical, but it's not mission critical. And here's

1216.14 -> the reason why it runs on a Mac Mini. And a Mac Mini is not a mission critical system,

1221.71 -> it does not have a Xeon type CPU or an epic type CPU, it does not support error correcting

1226.86 -> RAM. And it doesn't have any fault tolerance in it whatsoever. So it's not mission critical.

1231.549 -> But it's great those Mac EC two instances for, say, an application developer that needs

1236.14 -> to compile code, and they don't run on Mac or use a Mac. Normally, you can use pre built

1241.299 -> virtual machines like many people do, or guess what, you could use a custom virtual machine,

1247.059 -> you can create it just like you would in the data center. pre built virtual machines are

1252.36 -> available from Amazon as machine image. A machine image is basically an image of any

1256.93 -> server no different than a VMware image of a server that you've been using for decades.

1260.82 -> Azure has their own images, Oracle has their own images, Google has their own images. And

1267.73 -> typically, your machine instance is going to need a compute system to run on. And it's

1272.07 -> going to need some block storage to store your data. Now, like I said, you can build

1279.95 -> them, you can use a stock thing like the stock operating system and build on top of it, or

1286.19 -> you can buy one. So let's say you have some real security, you're not going to be using

1290.669 -> AWS WAF. And we'll talk about that when we get to the security section, you're going

1294.409 -> to be using an industrial grade firewall, if it matters, say something from Palo Alto

1298.45 -> or from Cisco, or from Fortinet or from checkpoint, you're going to be getting that from the marketplace.

1306.59 -> And it's a prebuilt virtual machine optimize operating system is fully hardened. And it's

1311.87 -> gonna with the precision firewall fix services. And we can get, we can take one of our virtual

1319.74 -> machines in the data center and convert it to an Amazon virtual machine image and relaunch

1323.69 -> it in the cloud. And that's what we're talking about. Basically, we're just uploading our

1329.559 -> fifth our servers. And we can also just like we do in the datacenter, VMware gives you

1335.19 -> an environment to take a physical server, a bare metal metal server, and then turn that

1340.61 -> into an image that you could then run on the VMware server, we do the same thing on the

1344.38 -> cloud, no different and there's tools to do that. So let's talk about the Amazon machine

1353.5 -> image, I mentioned has an operating system that's going to have lunch permissions, and

1357.429 -> it's gonna have a block device mapping, which basically says, which block storage, you're

1360.3 -> not going to know when you've got this image of a system. It's a single file that contains

1368.559 -> everything on the hard drive. It's really cool. For many of you guys that are a little

1372.48 -> older, like me, if you've used semantic or Norton Ghost knowledge, semantic ghost, and

1379.08 -> we could take a whole window system, we copy it to an image, and then we could push that

1382.6 -> to 100 different computers that were configured identically. That's an image sync thing. The

1389.243 -> image is that snapshot of a machine, we can stick that in different regions, we can stick

1394.48 -> it in different clouds if we want to. Now AWS would tell you to take that image and

1399.64 -> stick it in a different region for disaster recovery. It's insanity, I'm going to tell

1403.83 -> you why it's insanity. You don't backup your stuff to the same place. And the single cloud

1409.33 -> is the same place. Even if it's a different reason. You Back it up to another environment.

1414.32 -> I wouldn't I wouldn't put like if I was worried about my bank going bankrupt, like Silicon

1418.95 -> Valley Bank, just, you know, kind of had some financial issues. I wouldn't put my money.

1422.9 -> And back in Silicon Valley Bank, I use a different bank. So if you're going to do some disaster

1427.76 -> recovery, and we'll talk about that, don't stick it in the same file. That's as ridiculous

1431.84 -> as you're worried about your bank putting your extra money in the same bank and hoping

1435.16 -> that the banks, okay, it just doesn't make any sense. Now, when you're dealing with virtual

1441.6 -> machines, what we're really talking about is as follows. You got your operating system.

1451.58 -> Now, if we wanted to basically set up a virtual machine to configure itself, like we've done

1457.43 -> forever, we'd write a script, typically a bash shell script, or Windows, we'd write

1462.42 -> a PowerShell script. Systems admins have been doing it forever. We can do the same thing

1467.561 -> on AWS. Let's say we've got auto scale and we wanted to configure our servers coming

1472.679 -> up. We can basically write a little script many bash shell script that for example, they

1478 -> call AWS COVID. A bootstrap script remember bootstraps route for your for your exams.

1483.23 -> And realistically speaking, we can say update the operating system. So let's say it's an

1490.19 -> Ubuntu system sudo apt apt get update sudo apt apt get upgrade? We could have it immediately

1495.85 -> do that, and then we could have it so install our web environment. I'm an architect and

1501.72 -> architects don't touch the technology. But I think from I think I remember with a bunch

1506.2 -> of system, it's sudo apt install httpd or Apache, we could set that up to install the

1513.24 -> web server, for example. While we're at it, and basically, they're just simple scripts.

1522.5 -> Now, we're going to talk about the way you rent your systems. A cloud is is like a hotel,

1531.13 -> you rent a room at a hotel, you rent space in the cloud in somebody else's data center.

1536.049 -> That's it. So how do you rent it? Well, there's a lot of ways that you can rent it. And we're

1544.11 -> going to talk about the renting options. The first is on demand. What does that mean? And

1552.24 -> when when I use this, I started a new website for my cat, Cindy, I have absolutely no idea

1560.46 -> how many, how many how many hits per day she's gonna get. And if I knew, I would specify

1567.11 -> it, and I'd get a cheaper rate. But if I don't know, I could stick it on an on demand server.

1572.48 -> I'm going to pay by the second. And what will ultimately happen is, if I need capacity,

1582.1 -> I'll add other servers. Now, let's convenient to not know what I'm going to pay for. So

1590.02 -> guess what? I pay extra, because I don't know. So kind of keep that in the back of your mind

1595.549 -> and on demand instance, is as follows. You purchase it, and you don't exactly know how

1604.51 -> much you're going to use. So you can on demand. So on demand is very useful. And even if you

1609.21 -> do know what your capacity is, you're probably still going to use on demand. I could also

1613.95 -> say specify five servers and use on demand for additional capacity. So on demand, you're

1620.94 -> not 100% sure what your needs are, where you might scale. You pay more but for the for

1626.799 -> the youth, but it's not always there. The next purchase or renting option, if you want

1633.33 -> to call it is something called a reserve.

1637.269 -> What is that? Well, it's a follow. I know I need 10 servers with 28 cores, and 128 gigs

1648.14 -> of RAM.

1649.14 -> I tell AWS I'm going to use 10 servers with 128 cores, and or 24 cores and 128 gigs of

1658.47 -> RAM. And guess what? I tell AWS I'm gonna buy it for one year, or for three years. And

1665.35 -> the longer I commit to it, the cheaper the price will give me why I just committed to

1670.93 -> it. Then again, if you're going to commit to something, it's really no different than

1674.039 -> calling Dell and buying it for your own data center. But there's that. But when you tell

1680 -> them, I'm gonna buy your stuff for a long period of time, it enables AWS to know how

1684.19 -> much capacity they're going to need for their systems. And they'll either know to buy new

1688.08 -> servers to support your needs. And that's the way that work. So on demand, pay by the

1693.49 -> second, you pay the highest price, reserve, and you can reserve it consistently full time,

1699.51 -> which is what we're talking about. You pick a duration from one to three years. And the

1703.299 -> longer you commit to buying somebody else's stuff, the cheaper the rate you go. And the

1711.549 -> next thing that we'll talk about is scheduled reserved instances. So let's say for example,

1720.89 -> I know that I'm going to be running a big batch computing job, and it's going to be

1724.49 -> every Friday, every Saturday and Sunday, and it's going to be for 48 hours straight. I

1731.96 -> can tell AWS and schedule this capacity, and I'm gonna prepay and commit to or at least

1738.69 -> commit to paying unnecessary prepay for say three years. And I'll get a cheaper rate than

1743.88 -> if I used on demand. So on demand, pay about the second highest price. Next low, the lowest

1749.94 -> we're gonna get here for Well, the next one of the lower prices is going to be to reserve

1753.42 -> it for one to three years scheduled reserve, we're going to pay more than if we reserve

1757.159 -> it constantly for three years, but we're still going to get a discount on our rental prices.

1762.72 -> Now the cheapest option is something called the spot instance. Google calls it a preemptable

1767.83 -> instance what's a spot instance? AWS usually has extra capacity. And you can bid in an

1775.399 -> auction like manner on this x extra capacity. If your bid gets one, you get cheap access

1784.01 -> to compute power. Sounds great, right? There's always a caveat. Always a caveat. If you're

1790.85 -> using a spot instance, and somebody outbid you on that instance, your system gets shut

1795.24 -> down, whatever you're working with stops and you're out of luck. So, so Spot Instances,

1803.71 -> which are basically virtual machines that you bet on are great. If you've got something

1807.929 -> that's not super important, and you've got systems that are tolerant and being turned

1811.62 -> on and shut down, but don't put anything that matters on it on demand, highest price, standard

1819.031 -> reserved instances where you're scheduled for one to three years, low or one of the

1823.26 -> lower pricing options for long term consistent use scheduled reserved, I'm going to reserve

1829.711 -> it every Saturday on Sunday or something like that. You get a discount for it. Spot Instances

1833.87 -> the cheapest but do you really want to put your systems on something that wasn't get

1837.899 -> shut down? Because somebody else tells you when you're out of luck and you're offline?

1841.28 -> Probably not. So maybe good for experimentations? No, we'll talk about tendency options. Typically

1854.47 -> speaking, when you're dealing with a cloud computing environment, there's something called

1857.929 -> oversubscribed and your stuff is based on what's oversubscription. Just like your internet

1863.43 -> service provider, your internet service provider does not have enough capacity for everybody

1868.72 -> to use their stuff completely 100% at a time, AWS won't really talk about this, but their

1873.639 -> service provider and all service providers do this, generally speaking. So if I've got

1879.96 -> 120 cores on a server, I might sell 156 cores, because we assume most people aren't using

1886.66 -> them all at the same time. Now, if everybody uses their stuff, on the same time, there's

1892.57 -> performance constraints. I'm sure all of you have experienced oversubscription in your

1896.63 -> life, when we all got stuck working home from from home from COVID. I mean, I've been working

1901.33 -> home from decades. But when we all got forced working from home from COVID, and all of a

1905.299 -> sudden everybody's in their house, and the kids are watching YouTube videos, and somebody

1908.76 -> else is playing games and somebody else is watching Netflix and you're trying to work

1911.679 -> on your internet came to a screeching halt, is because your internet service provider

1915.809 -> did not have the capacity for all the people that would be home at these times. So anytime

1920.48 -> you're dealing with cloud computing, it's a service provider know they're going to be

1923.929 -> oversubscribed on their networking on their computer. But let's talk a little bit more

1928.34 -> about shadow tenant. This is standard, you rent some compute space from from your cloud

1933.779 -> provider, and your stuff is on there, your competitor stuff is on there or somebody else's

1939.029 -> systems are all on the same server that shared tenancy. This is standard unless you do something

1945.679 -> else about it. Now the next thing is something called a dedicated instance. Now this is basically

1956.37 -> where you say, Go cloud careers is reserving this entire server, an all of our virtual

1964.039 -> machines will be on that we can guarantee that we're not over subscribed this way. And

1970.179 -> we can guarantee for security purposes that nobody else is on our server is there usually

1974.909 -> speaking a security risk of having other people on the same server, not really, the hypervisors

1979.029 -> are pretty darn secure. But if we want to know that whole servers ours, we can do so.

1986.2 -> Now, what's the next option, the next option, if you really need performance is to get a

1991.309 -> bare metal server. This is going to give you the same performance that you had in the data

1996.03 -> center. Hey, wait, I can conceive data center like performance of the cloud. That's how

1999.49 -> you do it with a bare metal server. And on your bare metal server, you can do whatever

2003.759 -> you need. Maybe you're running some an application that needs access to the actual MAC address

2009.72 -> on the Ethernet cord or a serial number. Or you need to stick a security key and not have

2013.83 -> those kinds of things. That's the bare metal server when a secure windows when you need

2017.74 -> access to the physical hardware. And guess what, you got a staff that's fully trained

2023.8 -> on Nutanix. And they don't want them to learn any of this AWS stuff, guess what, or VMware.

2030.77 -> You can purchase bare metal servers and run your stuff directly on it. And then you don't

2036.21 -> even have to train your people in AWS and AWS is basically transparent to them. You're

2040.49 -> just using their stuff. And it feels like Nutanix or it feels like VMware. For the most

2047.7 -> part, it minimizes the learning curve for you to do this. So now we know a little bit

2053.13 -> about the tenancy options. And we know a lot about the purchasing options. Now there's

2061.53 -> more to cover on these virtual machines. Chris, how long have I been speaking because we may

2065.379 -> need to take a few questions.

2068.27 -> Yeah, it's time to take some questions.

2071.27 -> Let's take some questions. All right. And I'll quickly talk about the DevOps vs architect

2076.111 -> for about one minute even though it's not related to the concept at some point.

2080.34 -> All right, so before we take some questions, I want to ask everybody if you like what you're

2085.849 -> hearing and seeing, hit that like button, hit that subscribe button, hit that notification

2091.53 -> bell button so you don't miss the next two sessions that we've got. A see we've got 344

2099.86 -> People really watching but only 104 of you like it apparently. So if you're enjoying

2104.59 -> it, make sure to hit that like button. So let's get to some of these questions. And

2111.569 -> then we'll get back to the content after

2113.67 -> Mike has finished with these questions. Why don't you go ahead and start with that DevOps

2121.96 -> versus,

2122.96 -> I'm gonna deal with the DevOps one, as well as the Linux and Python question at the same

2126.7 -> time. That came out of butser. DevOps is a career for software developers, because you

2134.05 -> must be a great programmer first, who automate software release cycles. They get involved

2139.59 -> in tools like Jenkins and Spinnaker. And they are 100% related to automation. And they all

2144.089 -> they are as a techie, that's all day long. Automating software releases, cloud architects

2149.829 -> design present until technology, we don't touch Linux, we don't touch Python, and we

2154.52 -> don't do DevOps. It has zero related to our career. They are completely somebody else's

2160.15 -> career. Kind of the difference between an airplane pilot and an airplane mechanic. An

2165.74 -> architect is a business executive who designs presents and sells a technology solution.

2170.76 -> And there's a list of everything you need to know to be an architect, and we're going

2173.64 -> to talk about all of it tomorrow, on the completely free how to get your first cloud job. DevOps

2179.03 -> is a completely different career. Linux administration is completely different architects are not

2184.92 -> allowed to touch the technology. If they work for AWS, Azure, Google. Oracle, if they work

2192.319 -> for Accenture, Capgemini, Deloitte, architects design it and sell it present it. Cloud engineers

2198.569 -> go and build it DevOps engineers go and build and automate things. Once it's fixed, finished,

2203.93 -> it goes on to a maintenance team called sis ops people. And once it's done that, if anything

2208.47 -> breaks, they call the Technical Support Center, which is a different team. So now let's get

2212.08 -> back to the content. What's the difference between bare metal physical and virtual server?

2219.71 -> A physical server is a bare metal server, meaning you call Dell you call IBM to call

2224.46 -> HP and they ship you a server that's bare metal. A virtual server is after you take

2230.99 -> that physical server, and you install hypervisor, the AWS nitro hypervisor, the VMware hypervisor.

2238.64 -> Nutanix has a beautiful hypervisor KVM, which is the hypervisor I'm pretty sure that's used

2244.15 -> in the OpenStack cloud. And then you chop that server into multiple little virtual machines,

2249.97 -> which is a logical image. And that's there. In fact, let's kind of Oh, actually, I thought

2255.95 -> I had it. I thought I had it in here. But for some reason, maybe I don't have a picture.

2259.839 -> And that's such a good question. I'm actually going to draw it out for you give me a minute.

2270.56 -> Okay, so let me do this. I'll actually go back to my, my whiteboard, because I love

2274.24 -> this question. And I would go once you guys to understand it. So let's go to season two,

2278.079 -> this is not okay, this is tenancy options are here and create my slide, which is really

2284.25 -> what I wanted to do. Okay, so here we go. Because this is a great question. We got a

2288.94 -> physical server. So let's say this is our server hardware. This is our server. Next

2300.851 -> thing we're going to do is we're going to install a thin layer of software, it's going

2304.93 -> to be called the hypervisor. And the hypervisor is what's going to be involved in chopping

2315.63 -> up the server, two little mini servers. And here, what we're going to be doing is we're

2320.31 -> going to be creating virtual machines all on the same server. And what happens is, the

2325.18 -> virtual machine is going to have its operating system, it's then going to have to do this.

2332.39 -> It's then going to have its application something dependencies. And that's what it's going to

2338.75 -> be like so this could be a way that so we can have one of these on the server and we're

2342.94 -> going to typically speaking of multiple, we can have an another one. We could have another

2347.94 -> one. We can have another server, and we can have another server. So that's typically what

2358.66 -> we're speaking with. We've got our server. On top of that, we've got a hypervisor. And

2363.849 -> then we got all of our virtual machines and one virtual machine could be Windows, one

2367.43 -> virtual machine could be Red Hat, one virtual machine could be Ubuntu. So what we're really

2371.28 -> dealing with is taking one system, chopping it down to another system. So virtual server

2376.359 -> is any one of these things that we're actually dealing with any one of these things is a

2380.24 -> virtual server. The physical server is the thing that you actually buy. Bare Metal is

2386.35 -> the physical server that has nothing on it. Great, great, great question. How often does

2394.66 -> Spot Instances get shut down? It's based upon utilization bidding and capacity and it changes

2399.77 -> over single day isn't worth the risk. It depends on what you're doing. If I'm doing a test,

2404.24 -> it might be worth the risk. If I want to test the 1000 routers running on virtual software

2408.26 -> and stretch them out for a period of an hour, yeah might be worth the risk. Am I going to

2412.43 -> put something important on it? Of course I'm not. So everything is everything with architectures

2417.619 -> based upon in business is based upon business requirements. 100% There's no best tech ever.

2425.01 -> There's what works. And yes, there's entry level Cloud Architect roles. I get people

2430.81 -> hired every day with zero background whatsoever. Or any use good use case to Spot Instances?

2438.18 -> Yeah, if it doesn't matter, if you're looking for cheap compute capacity, and it doesn't

2443.46 -> matter. Yeah. But for anything real, I wouldn't be using it. test environments are beautiful.

2456.78 -> So a dedicated instance is different than a dedicated host. A dedicated host is a bare

2463.44 -> metal system that you can do anything you want with. install VMware ESXi, install Ubuntu

2473.589 -> Linux directly on it have physical access to the hardware, and be able to use all the

2479.02 -> cores at maximum capacity, including physical and virtual, a dedicated instance is basically

2486.13 -> running the AWS hypervisor, and you can create all your virtual machines that are specifically

2492.14 -> AWS virtual machines, all there. So kind of keep that in the back of your mind, there

2501.19 -> is a difference.

2504.89 -> can use VMware, AWS VMM is another problem with Google Cloud, well, what you can do is

2509.35 -> you can create your virtual machine, which is what I would, and then I would have that

2512.819 -> converted into an AWS virtual machine image. And you could have that one converted into

2517.44 -> a Azure virtual machine image and another one Google and then you can run three different

2520.91 -> clouds on it. So yes, you can convert one to the other, but I would create my own virtual

2525.109 -> machine, the hard way, the original way, like to VMware and then move it is the MO hybrid

2533.839 -> cloud, running VMware and connecting to a public cloud as a hybrid cloud just like running

2537.98 -> Nutanix and connecting to a public cloud as a hybrid cloud, just like running OpenStack

2542.25 -> and connecting it to a cloud as a hybrid cloud. What are the use cases for a dedicated host,

2552.16 -> dedicated host, I want 100%. Total Control, I want to put VMware on my system, so I don't

2558.75 -> have to deal with any of these AWS Management consoles, and I want to do it my original

2562.27 -> way dedicated hosts. Or there are certain critical business applications that require

2568.27 -> access to the physical hardware, and they look for the physical hardware to boot up

2572.72 -> for the licensing purposes that must go into dedicated hosts dedicated instance would probably

2577.19 -> be a good case, as I'm the US government. And I don't want the Chinese government systems

2583.599 -> on the same system, or the Russian systems or the Greek systems or the Israeli systems

2589.75 -> on the same system server physical server as mine. That's when you use a dedicated Gnosis.

2599 -> Good question.

2604.29 -> We will discuss NAT gateways and Nat instances. And we will cover that when we get to that

2610.63 -> content. What are some fundamental skills needed to have better understanding of this

2614.28 -> course, network and data center, if you don't understand the network and data center, you

2618.67 -> will never understand the Cloud. Please join us tomorrow on the completely free how to

2622.45 -> get your first job webinar. And we will give you 100% list of every single skill you need

2627.53 -> to know to be employable as a cloud architect.

2635.849 -> So let's get back to the content if you can give me a hashtag. And the hashtag could be

2641.76 -> AWS Certified Solution Architect Associate.

2650.94 -> Okay, so let's get back to the content. Now in AWS world. They tell you, you secure your

2663.76 -> virtual machine access with the security group, I'm gonna tell you right now, that's great

2668.19 -> if you want to get hacked, but if it really matters, you're gonna need to do much more,

2672.569 -> you're gonna need host based firewalls Host Based ideas, you're gonna have to be removing

2676.75 -> unnecessary packages from the operating system and closing unnecessary ports. But in AWS

2682.42 -> certification world which is very artificial in nature, and contains about 5% of the knowledge

2686.619 -> of what you need to know to build any good career. They say you can secure your AC till

2691.099 -> instance by six using a security group, which is basically like a host based firewall, but

2695.76 -> that doesn't mean you shouldn't use your own host based firewall to an all high security

2699.02 -> environments do So what's it look like? We're going to talk about the concept of a security

2706.81 -> group. We'll talk about it much, much more when we get to the security section. But I

2711.309 -> just want to briefly touch on that. Now, when you set up your virtual machines, and if you

2715.19 -> download our free Labs, which is we gave me the link earlier in the video, you'll be,

2721.619 -> you'll be setting up security groups. And basically what it is, is it's a pre firewall

2725.89 -> before your traffic gets inside of the of your virtual machines. Keep that in the back

2734.55 -> of your mind. And we'll we'll discuss much, much, much more and much, much, much more

2740.28 -> depth are with me when we actually get to the security section. So how do you give an

2751.559 -> IP address to assist them? Well, when you set up your VPC and we'll talk about that

2759.819 -> much more later, and you're setting up your virtual private data center, because that's

2764.88 -> what your VPC is, they call it a virtual private cloud, but it's a virtual private data center,

2769.29 -> not a virtual private network, what you're dealing with is you have to create your own

2773.64 -> IP address, base and cider range. And inside of that every single virtual machine that

2779.42 -> you have, is going to be given an IP address on the subnet that you create. And it's going

2785.21 -> to receive that address via the Dynamic Host Configuration Protocol. Now, if it's inside

2790.819 -> of your VPC and it's internal, you're going to be using private IP address spaces. And

2795.16 -> if you need to connect it to the internet, you're going to need a global or a public

2800.16 -> IP address, which of course, the world calls it a public IP address and AWS marketing people

2804.82 -> name that name that an elastic IP address, I don't know where they come up with these

2814.589 -> marketing names. So keep that in the back of your mind. Now you can also get a ipv6

2821.079 -> address. Now all ipv6 addresses are public, we don't have any kind of RFC 1918, private

2826.349 -> IP addressing spaces like we normally would keep that in the back of your mind. They're

2832.78 -> automatically assigned an ipv6 address if you don't need ipv6, shut it off. The more

2839.29 -> addresses you have, the more things that are open, the more the world can hack you. So

2843.109 -> only use what you need. Now, how do you manage these systems? Well, how would you manage

2850.77 -> any virtual machine, you can either secure shell into them or SSH, just like you would

2859.079 -> any other Linux machine, or router or switch or viral. You could use the AWS Management

2867.901 -> Console, which is a web based browser based way to do things, it is super easy. It's basically

2874.16 -> self explanatory, click, click, click, click and it's done. And if you don't know, you

2879.97 -> can just Google the instructions. Having said that, it is slow. Versus SSH, which is much

2885.78 -> faster if you know the commands. And great question there from George, do I have anything

2892.88 -> he gets the marketing team? No. But there's millions and millions of people that are 100%

2897.99 -> Confused by whether something is because the marketing people made up funny names. If it's

2903.48 -> a Windows system, you can manage it via the Remote Desktop Protocol, RDP. And you could

2909.78 -> do a lot of the management via the software development kit. And here's how you can really

2913.4 -> set these things up. The whole world uses TerraForm, they're gonna have one DevOps engineer

2917.291 -> deploy 1000s of these at the same time, in many cases, and they can all do it via infrastructures

2922.41 -> code. So that's how you can set up these things. Now we're also going to talk about outpost

2929.78 -> and what is outpost, it's fairly new. It's a fully managed service that uses a virtual

2934.99 -> machine or an EC two instance. And something that's an AWS supplied instance, it's a physical

2941.02 -> appliance that shipped to the customer and plugged into the customer data center. Why?

2945.3 -> Because the latency of going to the cloud is high versus running a virtual machine in

2949.531 -> your data center is there. So you can order a server directly from AWS, stick it in your

2954.44 -> data center, and that's called an outpost, I'm gonna tell you right now, you could buy

2958.15 -> that server from Dell for far, far, far, far, far less than it would ever cost you to rent

2963.83 -> it out from AWS, at least, at least 400% less, but it's up to you and how you want to do

2968.76 -> it three 400% cheaper to buy your own server. But if you buy your own server, you're gonna

2972.98 -> have to have some knowledge on how to set up the virtualization. If you buy this AWS

2977.38 -> outpost, the cool thing is, you can just click a few Buck buttons on the AWS Management Console

2982.32 -> and set up your servers and it's fully managed by AWS. You don't have to think about patching

2986.42 -> your hypervisor and things like that. So it's the convenience factor. And in fact, when

2994.099 -> you go to a convenience store like a 711 in the US or a Walmart in the US, you pay a lot

2998.7 -> more than you would in the supermarket. Right, for a convenience. That's what we're talking

3004.26 -> about here. They do all the work for you, they manage it. And because of that, you pay

3008.93 -> more but you got great performance just like close to buying your own server. Okay, now

3016.15 -> we're gonna get into databases. So to make sure I know you're here, give me a hashtag

3020.52 -> that says Databases.

3025.42 -> I see a hashtag that says databases give me a hashtag that says Databases, we don't use

3033.869 -> acronyms. As architects spell it out databases, your CEO doesn't know what you need. And an

3040.98 -> acronym, your hiring manager doesn't know what you mean. And the reason we never use

3045.47 -> acronyms in technology, and some people do, but you should never is they all can mean

3050.619 -> 10 different things. Somebody says VM, I don't know if it's a voicemail vulnerability management

3054.64 -> or a virtual machine. And that's the point, there's millions of things that all need the

3060.079 -> same thing. So communication, lack of clarity causes errors. And I want you to have a great

3065.91 -> career and avoid acronyms whenever possible. So we're going to talk about data bases. Even

3076.95 -> dB, I think of it means Dunnville. So keep that in the back of your mind. That's why

3081.28 -> we never use acronyms, because they all mean different things to different people. So what

3086.5 -> is a database? A database is this is an application that enables us to store large amounts of

3093.86 -> information, large amounts. It facilitates the sorting, calculating reporting and information

3100.24 -> sharing. And it is a critical component to modern applications. Now, when we talk about

3105.67 -> these databases, I'm going to cover all of them, I'm going to tell you which ones you

3109.57 -> should probably never use in real life. Why? Because they will make multicloud impossible.

3116 -> And remember, 98% of organizations use multi cloud. So I will cover all so you can pass

3121.79 -> the AWS exam, and then I'm going to tell you don't use this, when it comes to things like

3126.26 -> Dynamo DB, and Amazon, Aurora, and a few other ones, but we're going to cover them all. And

3133.79 -> then I'll tell you why you should or shouldn't use certain ones. Now, when we're dealing

3139.339 -> with databases, for the most part, we're dealing with the same ones that exist everywhere.

3144.19 -> We've got relational databases like Oracle, we've got no SQL databases, like Mongo DB,

3152.119 -> or Apache Cassandra, we've got data warehouses, which are things like Postgres. We'll talk

3158.45 -> a little bit about creating data lakes, I'll show you the AWS way. And if we have time,

3163.41 -> I'll show you the real way. But we're gonna first talk about relational databases. And

3171.18 -> this is the most common database that we deal with in business. And it provides information

3177.609 -> and data that's related to each other. send you the cat, Science Diet chicken. So Cindy,

3185.14 -> the cat goes and buy Science Diet chicken, and got that in the database. Mike buys 500

3192.13 -> pounds of Sandy the cat of, of Science Diet chicken, because it's 30% off. Okay, great.

3202.53 -> So that's what we're really talking about. For my team. Lots of people are having a hard

3206.609 -> time finding the webinar tomorrow, please pop that in there. So people can reach it.

3211.96 -> And we have a great cut Cindy, by the way, oh, hopefully she'll pop it. So the reason

3217.13 -> organizations use relational databases is to find information that's related to each

3222.809 -> other. 20% off yield this much in sales 30% off yields this much in sales and enables

3228.91 -> the business to make better business decisions by finding information that is related to

3233.39 -> each other. Because remember, no business buys technology because it's cool. They buy

3237.66 -> it to improve their business. So with a relational database, we've got a lot of structure. Basically,

3246.05 -> we have a row. And we have columns, kind of like an Excel spreadsheet exactly that way.

3258.819 -> Now, when you're dealing with relational databases, they're what's called a Tomic. And what does

3266.681 -> that mean? They use this thing called an acid model. You could find a certification question

3270.54 -> on this, but you should know those transactions are all or nothing meaning sending the cat

3276.549 -> order got placed or getting got placed. That's it. They are consistent. The second I purchase

3284.4 -> Cindy's cat food, the database knows it and all the read replicas etc. all know about

3290.59 -> it at the same time. It's consistent. It's isolated. I buy fresh shrimp for Sunday. That's

3299.349 -> on a different entry. Then when I bought her Science Diet, and Cindy gets her shrimp every

3305.08 -> single day, believe me, she's not getting that she's getting turned in scarves every

3310.099 -> day. She's a very Happy Cat. And adorable, meaning, once an order goes into that database

3317.539 -> or transaction goes to the database, it is not lost. So they follow the atomic, consistent,

3325.26 -> isolated and durable model otherwise known as acid.

3332.24 -> So when you look at a database, the key is relational database, show the relationships

3336.96 -> between variables. That's why businesses are using if you've got your order ID, your customer

3343.079 -> ID or amount, your email address, the person's name and everything that's related to each

3349.78 -> other. So transactions, things like that, you purchase something, and it's pretty great.

3355.799 -> So that now now you know what we're talking about information that is related to each

3359.829 -> other. Now, when you deal with Amazon, you got two options for your databases, you really

3369.47 -> want full control, don't use any other database, anything, set up a virtual machine and install

3373.839 -> the database yourself, just like you did everywhere. If you don't feel like setting it up, you

3378.869 -> can take the easy way out. And you can set these things up. Now the relational databases

3386.619 -> available on Amazon or Amazon, Aurora, Maria dB, Microsoft SQL Server, my SQL, Oracle DB

3393.44 -> and Postgres, which is typically used as a data warehouse. We're going to begin with

3400.69 -> Amazon Aurora, and I'm going to tell you right now, you probably should never touch this,

3405.23 -> but it's going to be on your certification exam. If you use a proprietary database, when

3409.77 -> you use three clouds, you're going to have trouble. And 98% of customers use more than

3414.03 -> one cloud for good reasons, single clouds, a single point of failure and amount of how

3417.64 -> many regions and availability zones you use. Because a control plane failure, network failure,

3423.799 -> or hacking event will take down an entire cloud, as we've seen AWS go down first, globally

3427.519 -> for seven hours, which they call a power failure. Although I've never seen a data center power

3431.97 -> failure in 25 years. So Amazon, Aurora is a proprietary, meaning it's going to cause

3439.38 -> problems if you deal with anybody else. WARNING WARNING, WARNING WARNING WARNING is a fully

3445.2 -> managed relational database. They say it's MySQL and Postgres compatible, which means

3451.46 -> it is but try and get your information in and out, you're going to be using tools. Now,

3457.16 -> what is good about this Amazon Aurora database is it gives you some of the enterprise grade

3461.579 -> features that you would get an Oracle type database or a paid database. And it's relatively

3467.859 -> inexpensive. So there is design. And if you only had a small business that you never thought

3473.99 -> was going to grow, and it could tolerate an hour or a day of downtime, this might be fine.

3478.29 -> But I don't know any business like that. Least the cons I work with, and Amazon will tell

3484.64 -> you, it's five times faster than standard MySQL. Yeah, MySQL is kind of a really slow

3489 -> database, and three times faster than Postgres. Now, it's kind of a SaaS application or software

3497.369 -> as a service application, meaning a server list and there's no servers for you to manage.

3501.66 -> And when you don't have any servers to manage, it's like going to McDonald's and getting

3505.23 -> a hamburger, you've got no control over it, because it's managed by somebody else. Versus

3509.83 -> having your grandmother who's a chef. Make you the perfect hamburger. Kind of keep it

3518.27 -> that way. So now we're gonna get into MySQL. It is an extremely common relational database.

3525.67 -> It's open source, it's been around for decades. Oracle owns it now. Even though it's free.

3532.38 -> And it's used in a wide variety of applications, you've probably heard of the LAMP stack, Linux,

3537.329 -> Apache, MySQL, PHP. That's what we're really talking about. Now we can talk about Postgres.

3544.351 -> Postgres is an exceptional data warehouse. It's also considered a relational database,

3550.34 -> very enhanced features that mean huge functionality improvements over MySQL. And that's one you

3556.119 -> can either use their setup, which is basically going to be on an easy to instance, otherwise

3559.78 -> known as a virtual machine. And guess what it's going to be using block stores. Or you

3562.83 -> can just set it up yourself, whatever you want. Maria dB. Now, this is another really

3568 -> exceptional relational database. It's open source, which means it's free. And it works

3574.04 -> everywhere. It's created by the people that created my SQL, but it's got a lot more additional

3578.27 -> features and functionality. Now we're gonna get into a paid database, Microsoft SQL, a

3586.14 -> lot of this stuff out there is on Server supports my SQL Server 2008 2012 2014. And this basically

3597.68 -> allows organizations that have Windows one workloads that are dependent upon Microsoft

3602.15 -> SQL SQL server to be used. Now, Microsoft has very different clustering and failover

3609.349 -> options than most databases. And there are four versions of it express web enterprise,

3614.789 -> and standard. Look, you can use any of these. And if you needed something different, you

3620.22 -> can always build the virtual machine and install it. Now we're gonna go to the king of all

3628.46 -> relational databases, the Oracle database, it matters in business. For the most part,

3634.63 -> they're using the Oracle database, and is one of the most popular relational databases

3639.329 -> in the world. It has, for the most part, one of the most extensive feature sets and functionalities.

3646.75 -> And it's developed, licensed and managed by Oracle, AWS, relational database service support

3651.579 -> standard one, enterprise, and standard. Each of these versions has different performance,

3658.34 -> flexibility and scalability options. There's two versions of licensing supported by AWS

3663.819 -> for the Oracle database license included. And basically, relatively speaking, in this

3668.41 -> version, the database is licensed to AWS, and you're using their license. And you can

3673.631 -> use standard edition one or standard edition two, or you can bring your own license to

3678.84 -> Microsoft. And this is assuming you have a license you bought it, I mean, bring it on

3683.579 -> licensed AWS, you brought it and then you're going to host your database. And the best

3686.67 -> now you got lots more license and flexibility, standard enterprise Standard Edition one,

3693.06 -> Standard Edition two. Now in a bit, we're going to talk about tuning the performance

3699.37 -> of these things with read replicas, caching and killing, we're gonna have lots of fun

3703 -> with that. But before we do, we talked about the relational databases that are supported.

3709.3 -> Now we're going to talk about no SQL databases. Now, a no SQL database is not new there from

3717.069 -> the 1970s, if I remember correctly, from IBM, and what a no SQL database means not only

3724.319 -> SQL. So we talked about a relational database having very strict columns and rows with data

3730.66 -> that's related to each other. Great for transactions. But what if you need a little more flexibility?

3738.74 -> See, anytime you've got tight constraints, this must be this, this this? You start limiting

3746.15 -> scalability. So when no SQL database was designed to give you a lot of flexibility in the way

3752.96 -> you store your information, you can store structured data, like transactions, unstructured

3759.47 -> data, like where you stopped in a Netflix movie. Oh, by the way, there's this new show

3763.579 -> called Night agent, I was watching it the other day on Netflix. And you know, it's pretty

3767.7 -> interesting. Every time you push pause, and you come back to it hours later, it takes

3771.42 -> you to the same spot. Because they're using a no SQL database. I believe it's Apache Cassandra

3776.44 -> these days. Meaning I'm 99% sure they're using Apache Cassandra to store your place. And

3783.279 -> that way, it stores your information. And you can pull it can't do that kind of thing

3786.7 -> with a relational database, video game people that actually play a game, right? And they

3791.32 -> go back to the same game, they stop it that they're using a no SQL database, because it's

3796.569 -> very flexible. And basically, what happens is we've got these pairs, and we've got IDs

3802.309 -> and values, and that's how the information is retrieved.

3810.89 -> So AWS, has their own managed relational database called Dynamo debate, something that I will

3818.329 -> never use ever, ever, ever not because it's a bad database. It's an exceptionally good

3823.099 -> database, but use Dynamo DB. And guess what? Now I got a problem with Google, Oracle, Azure,

3828.51 -> Nutanix. And OpenStack, I can't have that for my business. I can't architect single

3832.74 -> points of failure. So this is something that's right for the trash can AWS invented it, now

3838.779 -> you use this, you're stuck, you're handcuffed to AWS. And when they raise your rates, you're

3842.22 -> out of trouble when they have an outage or you're done and there's nothing you can do

3845.81 -> about it. So Dynamo DB, trash can go back to something like MongoDB or Apache Cassandra,

3852.619 -> but you got to do this. It's not that it's a bad database. It's that I don't believe

3855.589 -> in architecting single points of failure into environments. So we will talk about Dynamo

3862.88 -> DB, because they're gonna have lots of questions of it on your exam, because they want to put

3867.039 -> you in an environment all cloud providers are all vendors want to put you in an environment

3870.98 -> where you're exclusive to their stuff. And it gets really, really hard to leave, like

3877.94 -> the Hotel California, so you can check out anytime you want. You can check in anytime

3881.49 -> you want, but you can never leave. I'm not gonna have a career as a rock star, which

3885.48 -> that's another story. But so let's talk about it. AWS has a fully managed serverless database

3890.73 -> called Dynamo DB. They say it's highly available as long as AWS works, it's going to work for

3896.72 -> you. It's serverless which means there's no management Have the servers, they're still

3901.31 -> servers, still servers, but they manage the operating system, your storage, the security

3907.71 -> of it, and hopefully they do a good job of it. And it stores your information on an SSD

3912.71 -> storage for better performance. Now, the good news is, it's got low millisecond latency

3924.779 -> and encrypts all your data by default, and it can be backed up with little or no performance,

3928.94 -> and it can be set up for global cross region replication. Now when I say that it's proprietary,

3934.15 -> that doesn't mean you couldn't convert your information and move to a second cloud, the

3938.309 -> problem is going to be when you want to synchronize your data between the Azure cloud and AWS

3942.859 -> cloud, you can't do it with Dynamo DB. You also can't do with Google Cloud, big table.

3949.349 -> And you can't do it with any of the proprietary databases. So you really need to understand

3953.529 -> that. So it can be set up for cross region replication again, great, but if the cloud

3960.77 -> goes down, doesn't matter how many regions you're in. Now, because we're dealing with

3966.059 -> a no SQL database, we're dealing with name value pairs. And we've got what's called the

3971.25 -> primary index, which is basically your primary set up. But we can also set up secondary indexes,

3976.839 -> which allows applications have access to different query products. DynamoDB secondary indexes

3984.47 -> can be something called global or local local indexes will have the same partition key as

3988.75 -> the base table. Global indexes can span across all database partitions. Now, there's going

3994.799 -> to be some limitations on saying a key value can't go above 10 gigabyte, but that's pretty

3998.799 -> big. And to increase scalability, DynamoDB is eventually consistent. What does that mean?

4007.609 -> It means that if I write information to the database, other parts of it for a second or

4015.22 -> two may not have access to the most up to date information. So does it matter? It depends,

4020.349 -> if it's a bank, and I purchased a million dollars a Cisco stock, and then I'm going

4025.339 -> to sell a million dollars for an FIR for an eighth of $1. More a second later, well, yeah,

4034.65 -> better be better be immediately consistent, and this isn't going to work. But if there's

4040.119 -> something where you stopped in a you were you stopped in a in a Netflix video, doesn't

4044.92 -> matter. If it's inconsistent for a period of five seconds, of course it doesn't. So

4050.72 -> as you increase the scalability, you become eventually consistent versus instantly consistent.

4056.35 -> Now, of course, you could configure DynamoDB, should you be using this thing to be strongly

4062.19 -> consistent, meaning instantly consistent, but that'll knock down your scalability. So

4067.03 -> everything with architecture is going to be a choice. Everything you do if one thing affects

4070.39 -> something else, it's like throwing a pebble in a river, or a lake. And you notice that

4075.089 -> reverberates and you see these things. That's why architecture and engineering are different

4079.77 -> engineering is focused on the tech architecture is focused on the big picture, because you

4083.98 -> got to be able to see everything was going good. Keep that in the back of your mind.

4089.52 -> Now with Dynamo DB need to understand that you provision the capacity, you have to tell

4095.29 -> it how much you're going to need before you use it. You provision your read capacity,

4102.12 -> and your right capacity a ahead of time. And that way there's sufficient capacity for your

4107.77 -> needs. Now, this is really scary, you could set up Dynamo DB for auto scaling. Now normally,

4116.1 -> auto scaling adds capacity when you need it and removes capacity. Pretty cool, exciting.

4119.48 -> And that's the whole reason we're on the call. If it wasn't for auto scaling, for the most

4123.35 -> part, the call would be more expensive and lower performance data center. But auto scaling

4127.33 -> is really exciting. Now DynamoDB auto scaling is about the worst I've ever seen in my entire

4132.72 -> life. It scales up and doesn't scale back down. Why is this so bad? Let's say you had

4138.95 -> a period of five minutes, where you needed to scale up maybe 30 or 40 minutes, and then

4143.6 -> your capacity was reduced for the rest of the year. That means what's gonna happen is

4147.56 -> you're gonna pay for the rest of the year for the peak performance that you had five

4151.76 -> minutes, which is something don't ever want to do. So with DynamoDB, which I don't recommend

4156.64 -> you use because it's proprietary, and it's going to lock you into a single cloud, which

4160.779 -> no customer wants. If you've got to use it, provision it ahead of time and don't allow

4167.28 -> auto scaling to work because it scales up and not know. Now, in order to make this look

4173.76 -> more attractive to people, they offered a new option because because people are not

4176.42 -> going to use something that doesn't work and multi cloud and 98% of organizations are using

4181.71 -> multi cloud, AWS had to do something, they created an ability to create an infrequent

4186.299 -> Access table, which gives you lower costs for infrequent data you can save up to but

4191.6 -> there's a 25% fee to save and retrieve your data. And it's DynamoDB as always priced upon

4197.17 -> throughput on demand capacity as we're talking about is going to be available for a higher

4201.409 -> cost than fixed capacity. So what are some use cases that AWS will tell you for Dynamo

4207.35 -> DB, or the same as any other no SQL database when near unlimited scalability is required

4212.889 -> when lower latency is required, because these things scale very well on latencies are low.

4217.55 -> All no SQL databases, when you got to store a large amount of stuff, Internet of Things,

4222.81 -> devices that are all over the world. And I think even doohickeys here, she wrote a beautiful

4229.86 -> article. And she wrote it on edge computing, where you got all these Internet of Things

4233.219 -> devices coming in. That's what we're talking about. Dynamo DB, or any no SQL database would

4238.909 -> be great for that game player state where somebody's in a video game leaderboard, that

4244.53 -> kind of stuff. Netflix movies, when Netflix uses Apache Cassandra these days, but you

4248.1 -> know, keep that in the back of a huge number of financial transactions, ecommerce shopping

4253.35 -> cart inventory and tracking. You know, that's what we're going to use this. Now we're gonna

4258.15 -> get into a data warehousing now for everybody in the entire world, that means Postgres.

4267.08 -> Keep that in the back of your mind. AWS will also have their own one, which again, I'm

4273.86 -> going to strongly recommend you don't use because it's proprietary. So we'll talk about

4278.56 -> data warehousing. Data Warehousing, is where you store large amounts of data. Why do you

4286.94 -> store humongous amounts of data? Same reason we do any piece of technology to improve business

4294.32 -> performance. Keep that in the back of your mind. So we take all this information. And

4305.23 -> for example, we stick it in a relational database, and then we can run a business intelligence

4310.53 -> tool. And the business intelligence tool can help us visualize the data, we can then take

4316.27 -> this data out and we can prep it loaded, create our data, lakes, things like that. Now, you

4322.461 -> know there is that. So again, I'm going to show you the AWS propriety proprietary way

4328.98 -> to do this, which is not what I'm recommending. Here, you take data, you can store it, for

4334.11 -> example, in an object storage bucket, which again, we're going to use on all clouds, we

4338.461 -> love this so far, then you're going to have to map and reduce and normalize your data.

4343.32 -> The rest of the world uses a Python spark script, I'm gonna recommend you create a Python

4347.57 -> spark script, or at least your database people do. Why? Because it can take the date of the

4353.02 -> Python spark script and use the same script on the Azure cloud on the Google Cloud at

4357.22 -> the same time. And then you're going to stick your information into your data warehouse,

4362.07 -> which I recommend Postgres, not this Amazon Redshift. And then from there, you can look

4366.65 -> at your data with a visualization tool like Microsoft Power BI, Amazon Quickstart. So,

4374.57 -> now, we're going to talk about redshift, which again, is something I don't recommend you

4378.731 -> ever use. I recommend you use Postgres or another data warehouse, but I would never

4384.09 -> use a proprietary anything, because it's going to help it's going to hurt you when it comes

4388.61 -> to multicol. So Amazon Redshift is an AWS proprietary managed data warehouse solution.

4395.17 -> It helps you just like any other data warehouse would work. And it's going to help you find

4401.75 -> actionable information, you can use it for business analytics. And you can use redshift

4406.67 -> spectrum to provide real time insights into your business. We'll talk a little bit more

4411.4 -> about this proprietary database,

4414.63 -> data warehousing database, AWS will tell you it's fast, powerful, and fully managed. It

4421.27 -> can do petabyte scale warehousing, as can any Postgres. It's based upon Postgres, which

4425.8 -> is good because you can do SQL queries, and it works with applications, but just use Postgres.

4429.83 -> And don't even deal with this. And then you can use the same thing on multiple clouds.

4434.219 -> know, when you're dealing with Amazon Redshift, the primary architecture is built upon clusters

4439.77 -> of computing nodes, you're gonna have a primary node that's going to be considered a leader

4443.48 -> node, and the compute nodes are going to support the leader node. And your queries are always

4447.61 -> going to be directed to the leader node. What we're going to do because I know I've been

4453.73 -> going long as I'm going to briefly mention data lakes in the context of what AWS would

4458.29 -> consider for your exam, which is typically different than when we do in real life. And

4462.699 -> then we're going to open up some questions before we get into the storage and things

4465.57 -> like that. What is data lake? Data Lake is a repository where you store structured and

4474.429 -> unstructured data, it's typically an object storage. The reason organizations create data

4480.09 -> lakes, is because we want to have create a location whole process a large amount of data,

4485.51 -> and it doesn't require you to structure the data as you would in a database. For one of

4490.1 -> the people with a blue wrenches. Can you find the data lake presentation where we had Praveen

4496.199 -> a really wonderful Big Data Architect on a cloud architect speak for about two hours

4501.52 -> on how to create a data lake and what a data lake is and pop them into the chat box for

4505.95 -> everybody that's here, because we don't have the multiple hours to cover that. And I wish

4509.989 -> I was, but I've not been for being those an incredible amount of Big Data architectures.

4513.25 -> I am an enterprise architect, a cloud architect, a network architect. And I'd rather you hear

4518.69 -> from someone with 20 years of Big Data experience on this, because you get better information.

4525.42 -> So that's why we created data lake, we store large amounts of information, it's cataloged

4533.46 -> information. And that way you can query data and you can look for it. This is what the

4539.651 -> AWS, once you see as a data lake, you're gonna have your data sources, which is going to

4545.139 -> typically be a no SQL database, your data warehouse and your relational databases, you're

4550.26 -> going to typically have someone write a Python spark script for your data transformation.

4555.57 -> And then you'll create your data lake. And there's several steps in the process of creating

4559.3 -> the data lake of normalizing data, analyzing data, but that's where listed what we're talking

4562.96 -> about. Okay, let's go to some questions. And then, because I know I've been speaking for

4569.909 -> approximately 30 minutes, and then we'll get back to the content, we'll have all kinds

4572.929 -> of fun. At least I'm having fun. Uh, you guys have been fun. How to make a relational database

4577.909 -> highly available? Well, Abby, Chuck, one is what none, two is one and three is greater

4583.739 -> than two. So we're going to talk about that when we talk about database performance tuning,

4588.62 -> but never have one. And never have a single database. And realistically speaking, if it

4593.481 -> needs really needs to be highly available, put it in multiple clouds, not a single cloud.

4597.53 -> Great question. And we'll get to more of that later. What DB options you have to save MongoDB

4604.34 -> data to AWS? Well, if I was you, I would stick a Mongo DB servers inside of AWS. And that's

4610.17 -> all you'd have to do. Read Lavaca does it like taking backup but read replicas are for

4618.79 -> something else. But it can partially help you with backup with back but it's really

4624.739 -> designed for something else? We'll talk about that more when we talk about high availability.

4630 -> How do you ensure high availability and data durability for databases in AWS don't just

4638.61 -> stick them in AWS Look, we've seen global outage with AWS recent global outage with

4643.37 -> all the cloud providers. If you use a single cloud provider, I promise you, you will see

4648.29 -> an outage. And it doesn't matter how many regions you use, all tech fails, all service

4654.65 -> provider fails. In network architecture and Enterprise Architecture we've been taught

4658.07 -> for the last 25 years, never put all your eggs in one basket. And finance they say diversify

4663.13 -> your portfolio. We always add redundancy. No matter how many availability zones and

4668.83 -> regions to use in a single cloud, you are architecting a single point of failure. If

4672.62 -> you stick it in one cloud, take your database and stick it in multiple clouds. Now you've

4677.12 -> got a truly high nobody database. Can I explain the advantage of redshift? None? Don't use

4683.88 -> it. Instead use Postgres. Anytime you use proprietary for everyone benefit you gain,

4690.48 -> you gain a whole lot of problems long term. I wouldn't recommend you use it. Can you convert

4698.38 -> to DynamoDB to another type? Yes, you can. But that means you can't use multiple calls

4704.63 -> at the same time. So you could be with AWS, AWS as another seminar or global outage and

4710.48 -> they make up and they talk about something like a power failure. You don't believe it's

4714.44 -> a power failure. Now you want to go to Azure, or they raise your rates and you want to go

4717.98 -> to Azure. And now you got you got a problem. But Sarah, if you want to run Azure and AWS,

4724.52 -> at the same time for high availability purposes, you can't be converting data back and forth

4729.09 -> to each other, you need to use the same database, trashed Dynamo DB, and use Mongo DB or Apache

4735.52 -> Cassandra across your clouds and you'll never have a problem.

4742.81 -> So you have to match storage sizing with databases. Yes, absolutely. That's the way we've done

4747.67 -> it for the last 20 years. Does AWS support Cassandra? Yes, they do. Do they do without

4756.69 -> having to provision a minimum underlying infrastructures? Yes, they do. They have a managed keyspaces

4761.02 -> Mark. I don't recommend using serverless. I recommend your architects design something

4767.14 -> in your engineers build it whenever possible. You have no control when somebody else provisions

4773.11 -> and managers have no control whatsoever. You go to McDonald's. And they say Would you like

4779.11 -> fries with that and your hamburger comes out consistently identically in a good enough

4783.75 -> manner? Every single time? I haven't been to McDonald's in 10 years. But you know there

4788.48 -> is that if I go to Morton's and ask them to make me a hamburger, I can select prime beef

4793.5 -> prime rib, filet mignon, New York Strip to grind that hamburger into and they can basically

4800.4 -> make it medium where medium well, or whatever I want. When you use a managed services, you

4806.46 -> don't get any of that. It's the do you want fries with that? Now let's manage services

4812.07 -> Good. Well, they're often a little easier to manage. But there's a trade off, you're

4817.03 -> going to trade off something, performance tune ability. So the customers that I work

4824.031 -> with don't use as many managed services, they use their own. Remember, you know, if you

4829.19 -> use managed services, it's a little cheaper, and it's a little easier to manage. You don't

4833.05 -> need as many expensive people working on it, but you give up control. And is that a good

4837.719 -> thing? Sometimes, yes, sometimes no. If you have a team of inept people, and you don't

4841.17 -> want to train them, great. Use a managed service. If you have good people, and you're not worried

4847.949 -> about their training, you can do many more things. So what are the business requirements

4852.52 -> will always determine what you should use?

4859.84 -> Can I explain the redshift primary node? Absolutely.

4869.87 -> So basically, speaking, redshift is monitored around clusters of computing nodes, you've

4875.91 -> got the primary node, which is called the leader node, and every other node is called

4881.94 -> a compute node and supports the leader node. But all the SQL queries go straight to the

4888.4 -> leader node. Good question. How do you normalize data in the Data Warehouse? Well, once it's

4894.36 -> in the data warehouse, it's normalized. Typically speaking, when you go from one database to

4898.9 -> another database, you're going to have to map and reduce it. Now, AWS, of course, has

4903.84 -> its own proprietary Elastic MapReduce, but and that's really, we'll talk about what that's

4910.07 -> based on what organizations really do is their way to Python sparks growth, pi Spark is really

4914.86 -> created for normalizing data. And you can set that same PI sparks threat up for at least

4922.92 -> your data, people will do it. And we architects on toucheth, for example. And they'll set

4926.73 -> up that same script. And you can run that same script on Azure and Google, or Nutanix,

4932.25 -> or OpenStack. And that's why you know, for us, you know, we like to use standards, and

4936.79 -> we try to avoid proprietary things whenever possible. How do you maintain a basic between

4942.86 -> multi cloud the same way you synchronize between multifaith and you've got IP connectivity?

4946.76 -> That's it. As long as the network works, all this stuff works. Is it possible to run multiple

4953.33 -> databases in the database instance? I don't know what you mean by that. You can partition

4957.429 -> the database, which is sort of like up and not exactly. But I'm not completely sure I

4963.82 -> understand. What you mean. Can you virtualize the entire data center? Yeah, that's what

4971.9 -> a cloud is. Every Data Center has been virtualized for the last 20 years for the most part. If

4978.96 -> you and the second half of that was, if you've virtualized data center, and it goes down,

4983.69 -> you're done. Which is, and if you, you're done. And if you've got 10 data centers connected

4989.54 -> by the same network, or 100 data centers connected by the same network and the network goes down,

4997.07 -> guess what?

5000.8 -> What you're dealing with is all that goes down, which is why you should never use a

5007.42 -> single call. Same problem. DynamoDB is not scaled down your retinas the fact that his

5014.96 -> maximum number increases in the data 27. I don't know what you're referring to, according

5017.672 -> to the AWS documentation that scales up in terms of the throughput and capacity, but

5024.239 -> it doesn't scale back though. Is there a difference between managed and fully managed? fully managed

5032.29 -> means you got no control whatsoever? Manage generally means you have limited control.

5045.69 -> hypervisors not only split the physical machine into multiple logical machines, they do that.

5049.65 -> But they also combine multiple physical machines into one logical machine. Is this correct?

5054.35 -> Generally speaking? No. They combine multiple physical machines into a compute pool that

5060.36 -> you can pull from just like a cloud provider with good question. Good question. So before

5073.27 -> we go back to the content, let's talk about let's me give you give you a hashtag because

5079.21 -> I want to know your awakened learning. Why don't we chain that the one that that we give

5083.3 -> the next hashtag to hashtag free AWS course. That way, I know you're awake, alert and oriented

5092.639 -> and while I'm waiting for that, you know If you get a VMware vSphere, one of these environments

5102.65 -> that takes all your servers and adds it to a pool, tech electric, if you add getting

5107.69 -> Nutanix called or an OpenStack cloud, it's going to do the same thing and add it to a

5111.83 -> physical pool. That's what cloud software does. It's different than a hypervisor. It's

5116.36 -> the control plane that manages it. AWS has the control plane returning. So it's the control

5120.84 -> plane, every cloud has a control plane. And if that control plane goes down, you lose

5126.33 -> everything.

5128.45 -> And yes, a shout out to all the people behind the scenes, whether it be my Chief Operating

5143.27 -> Officer Christopher Johnson, my chief marketing officer, Alonzo, whether it be Leo, who's

5148.38 -> back there, whether it be child who's back there, whether it be Tyrone or Eddie or Anson,

5153.34 -> there's lots of people here, there's some volunteers like even Doyle, who always thrilled

5157.03 -> to have here. Thank you all for all your participation. And I know I'm missing people. AJ is one of

5167.03 -> them really great guy who's also helping.

5174.86 -> So let's talk a little bit about where you're going to store your data. And database, right.

5182.53 -> Realistically speaking, you're going to have three options. Option one is going to be you

5187.61 -> put it on provision IPs, which is AWS fastest, which is still very slow storage, as I showed

5194.8 -> you compared to other storages yesterday. Now your next option is general purpose, SSD,

5202.909 -> which is again, much slower than that. And your last option is magnetic storage. So realistically

5209.4 -> speaking, if it matters, you're going to be provisioned IO PS, you may get away with general

5213.65 -> purpose SSD on really small environments, but you're probably not going to use magnetic

5218.42 -> storage for any kind of database. The latency, okay. There's a cup, there's one more database

5226.23 -> I want to talk about. Before we get on. We get onto this, and it's the quantum ledger

5233.3 -> database. Amazon has a quantum ledger database that is fully managed and serverless. That

5240.28 -> automatically scales with application being serverless that eliminates the need or worry

5243.61 -> to process server capacity uses tables and indexes to query stored historical data. And

5250.3 -> unlike traditional databases, it's with not immutable record, keeping audit logs like

5255.8 -> relational databases. AWS quantum letter does not permit an update or delete operation.

5261.409 -> So if you've got something that's a database, in a highly regulated industry, you can use

5266.69 -> the quantum ledger database. But again, there's other industry databases that do this, that

5270.659 -> are not proprietary, so we recommend those. Now, let's talk about database optimization.

5284.92 -> And we're gonna have Amazon database optimizations, which is what we're going to talk about, these

5288.47 -> are the same things you would do with any database, but either way, we'll talk about

5292.32 -> backups, automated backups, database snapshots and encryption. You know, anytime we're gonna

5301.27 -> be dealing with optimizing our databases, we're going to be talking about scalability.

5305.61 -> We'll talk about read replicas, we'll talk about caching queuing, multi AZ, and realistically

5311.1 -> speaking, think multi cloud, which is not part of any AWS certification, because it's

5316.469 -> part of reality backup. If it matters, you kind of back it up right. Now, if you use

5324.69 -> one of the AWS managed database services, like MySQL, or Oracle, they do some really

5331.63 -> good things for you. I like full control, but you know, I have no problem using the

5335.63 -> Oracle managed database service on AWS. What they do is they backup your data in a very

5341.94 -> great way. The entire database is backed up on to an image. And you can retain this backup

5349.24 -> from one day to 35 days. And the backups happen at the same time each day, for the most part

5355.88 -> to find window, which is really great. And during the backup, you kind of got to know

5362 -> because it's pulling data off of the drive, the database may be temporary unavailable.

5368.02 -> And when you're pulling the data off of the drive to back it up the performance of the

5372.15 -> database, if it's not available, maybe severely degraded, kind of keep this in the back. So

5378.3 -> when it gets backed up, it's going to be in the form of the DB snapshot. Now you can also

5382.17 -> make your own snapshots. You know me I love control. And DB snapshots are a point in time

5388.48 -> copy of the entire storage file, like that old fashioned ghost image. That's 100%. It's

5395.92 -> got the operating system, all the patches you put in there all the dependencies in there

5399.48 -> or your applications, and of course, all your data so you can backup the whole thing, which

5404.65 -> is really cool. And you can relaunch that thing instantly, should you have a problem,

5411.19 -> another region, another availability zone, etc. And when you make a DB snapshot, it's

5422.01 -> available until you delete it.

5429.989 -> So what does it really look like? You've got the relational database, and you can create

5436.51 -> a snapshot of it.

5444.139 -> Now, when you restore a database from a snapshot, you're gonna get an identical new virtual

5456.27 -> machine with one exception, it's going to come up with a new IP address, which means

5462.39 -> the old DNS name that you have is no longer going to work. It's going to have a new DNS

5468.47 -> address too. So you may have to update your DNS records if your systems use DNS to point

5473.19 -> to the new system. And the people typically use DNS. And if you use the IP address of

5477.869 -> this system, by comparison, we don't recommend for lots of reasons, then you're going to

5481.8 -> have to update the IP address mapping and your application servers, which are going

5485.76 -> to the database. And when you restore it, you take your snapshot image, and poof, you've

5495.88 -> got a new database with a new IP address, just like I described. Add a new DNS drop.

5504.54 -> know, if you're going to store your data on a database, or any hard drive in the hard

5511.03 -> drive lost on your identity is compromised. Because what people can read your information.

5516.449 -> Do you want that? No. So Amazon supports encryption at rest for all your database. So what does

5526.57 -> this really mean in all practical terms, it means all the data stored on your server is

5532.54 -> encrypted. Effectively, what's going on is the EBS volume of the block store to the virtual

5537.699 -> hard drive is encrypted. This is enabled by enabling the Key Management Service, which

5545.35 -> makes it really easy to control the keys, we talked about the key management service

5548.69 -> and when these kind of a lot of these things yesterday. AWS also supports transparent data

5554.42 -> encryption. And Transparent Data Encryption is typically used to be with Oracle and Microsoft

5566.35 -> SQL databases by default. And you can set up transparent data encryption with the cloud

5571.42 -> HSM module hardware. It's like a hardware key encryption kind of manual. We'll talk

5575.84 -> more about that later. And it's transparent data encryption is really kind of cool. In

5581.59 -> encrypts the data on demand and decrypt the data on demand. So when you store the data

5585.63 -> with transparent and encryption, it is encrypted. When you pull the data is decrypted. And the

5593.34 -> cloud HSM is a hardware device for storage and management of your encryption keys.

5601.4 -> AWS also supports encryption and transit. What does this mean? It means that your data

5610.13 -> is sent as an encrypted encrypted on the way to the database to be stored. And how does

5616.94 -> this work? Well, basically, it uses the TLS protocol and SSL certificates. And you use

5624.06 -> a certificate to basically assist with the authentication of the endpoints and your data.

5629.08 -> Basically, in the same way, when you go to a website, I use it a little lock. And it's

5633.48 -> using SSL based encryption. That's really what we're talking about. Now, databases have

5639.88 -> become really mission critical, mission critical applications. So how do we improve the scalability

5647.05 -> of these things? Well, the simplest method is to scale up, meaning we're in a server

5653.52 -> that's got eight cores and 32 gigs of RAM. And we bump it up to a 64 cursor server with

5659.54 -> a terabyte of RAM, or 192 course server with six terabytes around and we pick but at some

5665.86 -> point, I promise you, you're gonna run out of capacity, no matter what you do, will run

5672.639 -> out. So when you run out of capacity, you're going to have to add capacity. Now with some

5683.61 -> databases, like Apache Cassandra, for example, you can write all databases as the same time

5689.77 -> I'm 90% sure Oracle database allows that as well. But most do not. So we'll talk about

5699.59 -> how we're going To deal with this because it's not like you can just auto scale a database

5703.36 -> the same way you can auto scale a web server. Lots of trickiness here. And at some point,

5710.199 -> you just, you're gonna have to start getting creative. No with no SQL databases. This is

5716.38 -> simple. Apache Cassandra, you just add servers and it writes to them all at the same time.

5722.86 -> With Dynamo DB which sock and we're not recommending you can basically partition the database and

5727.69 -> it chops or shards the database into partitions, and the application will have the intelligence

5732.219 -> to route your traffic to the correct shard. No, now with relational databases, it gets

5739.9 -> a little more complicated. What we do is we add read replicas. Okay, what's a read replica?

5747.429 -> A read replica is a read only copy of the data. Except for Maria DB as well. Now read

5758.13 -> only so what does this really mean? Right now I'll give you an example. There's all

5763.61 -> the blue wrenches in this YouTube chat box. For today, I am the primary master database,

5771.35 -> meaning I'm out here providing the information. And many people are asking questions or things

5777.25 -> that I covered and they need a little clarification. And they're usually the little blue wrenches,

5782.61 -> whether it be my chief operating officer Chris, who's over in Tampa, Florida, or for Forca,

5788.27 -> who's over there in Cameroon, who's answering questions or child who's over in Dallas, Texas,

5793.219 -> little supertall over there. And she's answering questions. Or Alonso, who's over there in

5798.989 -> Katy, Texas answering questions or Edie over in them in the Cameroon that's answering questions.

5803.57 -> They're kind of like read replicas, they basically enable me to focus on what I'm teaching. And

5810.09 -> they help answer things. Kind of keep that in the back of your mind. So the way we use

5815.96 -> read replicas is the database has something called right capacity and read capacity in

5820.63 -> a primary database does it both. A read replica is a as a read only copy of the instance.

5832.23 -> And what happens if we take the read load off of the primary server and we push it on

5836.88 -> to some other servers, the primary server can focus only on writing. And the read replicas

5842.65 -> can handle the read traffic. And remember, if you've got a server that's going to read

5846.38 -> and write, and you can remove all the reading and only has to write it can scale further.

5853.37 -> And why are we doing this because as I mentioned previously, there's only going to be a certain

5858.21 -> amount of server cores and D RAM and disk performance you're going to be able to get,

5862.73 -> so we're gonna have to we're gonna have to get past us. And that's how we're going to

5872.06 -> do it. It's going to reduce the load. So what does it look like architecturally, here's

5877.461 -> what it looks like. Basic three tier application, we've got our web servers, which can auto

5882.23 -> scale. We've got our app servers with who can auto scale, we've got our main or master

5888.19 -> database. And we've got the read databases. And what happens is you'll point your questions

5892.96 -> to the read database, and that'll free up the right database to do more.

5904.67 -> So when you use Read, read read replicas when there's read activity, if it's all right,

5908.929 -> activity, read replicas aren't going to do anything. When query traffic, meaning people

5914.65 -> are trying to read read read, read read is slowing things down. You need a read replica.

5921.8 -> If you've got four times the read capacity, and in the right capacity, add four read replicas.

5927.09 -> Because adding extra capacity. Now while we're at it, I want to make a clear read replicas

5934.61 -> are used for performance. They do not aid in availability or disaster recovery for the

5940.969 -> most part. Now the next thing we're going to talk about is database caching. What is

5951.32 -> caching? Caching is a service to take frequently accessed information and put it in memory.

5959.46 -> Caching works by taking a request and temporarily storing the results of the request. Now, why

5966.57 -> would we use recursion? Let's say I'm going to we're going to use a no SQL database with

5974.5 -> caching right now. And the reason we're going to use a no SQL database is Taylor Swift gets

5978.8 -> a brand new cat. And she posts to Instagram, Facebook, Twitter, Tik Tok, and LinkedIn.

5990.3 -> All the new photo of her cat and somebody's pulling the information out. Pretend the cat

5998.13 -> is stored in the database the cat photo

6001.84 -> Now the read replica can answer the cuts question. And then the cache can keep re answering it

6007.199 -> offloading the read replicas. So the read replicas offload the right, or the primary

6012.85 -> database, and the caching can offload the read replica. Now of all the information if

6017.07 -> one thing is for teller, so let's count into listings for Katy Perry's cat. Another thing

6020.69 -> is for Christmas, Cool Cat Sunny. And the next request is for my super awesome Princess

6025.08 -> cat, Cindy. By the way, I bought her from my wife and realized two cats the second level

6029.52 -> my level of this little cat. But you know, that's neither here nor there, you know, then

6033.42 -> the caching is not going to help. In fact, if you add caching in an environment, while

6036.96 -> the requests are different, it will slow things down. Caching is used from frequently accessed

6042.46 -> information. Now, typically, speaking, organizations have been using caches forever. There's two

6053.53 -> caches which we that are typically used in business. For the most part, businesses use

6059.219 -> Redis caches. And they also can use Memcache D. Now, of course, in the cloud, you could

6069.96 -> set up your own servers with your own DRAM, and set up your own Rama Redis Cache, or your

6075.35 -> own bucket, rhoncus D cache, or you can use the fully managed AWS cache, there's no reason

6081.69 -> not to use these things. It's pretty simple to set up. And you can basically use a premade

6089.91 -> Redis cache. Now people use Redis, because it's got the most robust feature set of caches

6095.659 -> to typically use. And you can manage, you know, Redis workloads to Elastic cache, or

6100.86 -> you could set up your own cache, which in certain cases makes sense, it's based on the

6103.59 -> business requirements. Now, if you need something simple, simple, simple, simple, simple, you

6109.03 -> can use elastic cache for Memcache. D, its simplicity. And elastic cache is compatible

6115.239 -> with for Memcache. D is also compatible with memcache D. So there's, you know, these caches

6120.44 -> aren't that different from each other. But if it matters, you're going to be using Redis.

6125.23 -> I told you how caching sort of helps, I'll show it to you visually. If the requests keep

6133.96 -> coming in for information that's on the database, the cache can store that in memory, and provide

6139.01 -> the answers so that you don't have to do so of course. Now, let's talk about database

6148.46 -> killing. And if you want, I'll actually architect these things together for you to try and put

6153.34 -> it into context for you. What is killing for the Americans is a complicated concept for

6161.09 -> the English or anybody that follows, you know, colonial English language. It's simple because

6165.86 -> caching means put everything in a line. You got to a plane in the UK, and they say, Please

6172.14 -> form a queue. So killing is really a means to schedule the delivery of your data. It's

6180.24 -> used in lots of applications and why we're using caching test question here for the AWS

6185.13 -> Solution Architect Associate. Why do you use the cache is to decouple the traffic destined

6190.489 -> from the database. And your application services used for application decoupling, you may see

6195.96 -> a test question on that most likely will. Come caching is used to decouple it. Here's

6202.381 -> the way caching really works. Caching is used for you got a sender who's sending a message,

6213.23 -> they stick the message in the cache. When the receiver is ready to receive the message,

6220.369 -> guess what happens? It gets pulled from the cache, I'm sorry, the cue, keep going to cash

6226 -> that's a cue. And then it's removed from the queue. So I want you to think about this.

6231.3 -> If I'm sending messages as fast as I can, and the receiver is not ready, they're going

6235.03 -> to be lost and dropped. But in this particular environment, in this particular environment,

6242.78 -> I can dump all those messages into the queue. And by dumping it into the queue, it gets

6252.86 -> a holding pattern. So I got a million messages, I stick them in the queue. And when the system

6258.239 -> is ready, it will drain them. So caching promotes scalability. It enables you to root CPU sizing

6265.72 -> and I'll show you why. And generally speaking, the killing lowers your cost I'm sorry, keep

6270.489 -> calling, capturing and referring to queuing. Now, most businesses use Apache Kafka as a

6276.71 -> cue. And you can set up a virtual machine and use Apache Kafka on all your clouds at

6281.71 -> the same time and use your same beautiful cue Apache Kafka as a queueing system. Or

6288.81 -> you could use the proprietary which I don't recommend AWS SQS or simple queue service.

6295.52 -> And by doing so, it's a pre managed queue for you. But if you're going to use three

6299.889 -> cloud Have, you're gonna have three different proprietary queueing systems might not be

6304.44 -> the simplest, most elegant thing. But if you use Apache Kafka, which works on all clouds

6310.449 -> and all data centers, at the same time, we're gonna simplify your thing. So again, SQS is

6314.449 -> another service that, you know, I wouldn't be architecting a day anything, unless it

6318.7 -> didn't matter. Because everything I do is multicolor, just like 98%. So when it comes

6326.159 -> down to it, there's two options, there's a simple step, there's a standard version, which

6330.02 -> is a simple queue. Basically, messages come in and out as fast as they can, there's no

6335.449 -> guarantee of the order of the messages. And if you need to guarantee the message delivery,

6343.73 -> what you could choose to do is you could set up a FIFO queue or a first in first out queue,

6349 -> message, one goes, then message two goes in the history, but it's going to slow it down,

6352.73 -> slow it down and slow it down. Why? Because what if message one was 1500 bytes. Message

6357.56 -> two, two was, I'm sorry, was 1500 bytes. But message three, four, and five were 64 bytes.

6364.219 -> The three, four and five is going to get there most likely before message two. And when you

6368.06 -> set up first in first out, you're going to slow the system down, but it's up to you.

6371.69 -> Some require some application and business requirements required to do so. So the reason

6378.48 -> we're using these killing systems are as follows. This, let's pretend we're looking at the CPU

6385.44 -> performance of a database. In your typical 333 tier environment, I use proprietary technologies

6391.94 -> here, just because it's an AWS class, when I do these pictures, you got your web server

6396 -> coming in. You got your app server coming in, its uses this proprietary queue, and then

6403.449 -> it gets stuck into the database, proprietary DynamoDB. Now, if this was a regular database,

6410.48 -> which we're going to see as the CPU is going to go up here is nothing, it's going to spike,

6415.3 -> it's going to go down CPU is going to spike and it's going to go to nothing. But by using

6421.239 -> a killing system, we can smooth out. And by using killing effectively, what we're doing

6431.409 -> is is realistically speaking, we're smoothing it out. So that way we pop stuff in the queue,

6439.11 -> and we take it on a consistent basis, it's kind of like driving a 40 miles an hour in

6442.92 -> your car, versus two miles an hour, 100 miles an hour, two miles an hour, 100 miles an hour,

6448.28 -> which do you think is going to be more efficient on fuel your cars longevity. That's why we

6456.81 -> use the cueing. So killing helps remove read content, and I'm sorry, write contention that

6464.28 -> keeps messages from being lost. And what else I'll also call is you can use the depth of

6468.67 -> the messages in the queue to trigger auto scaling of different servers, which we love.

6476.65 -> So when should you use a queue when you want to increase the scalability because there's

6480.5 -> a lot of write requests. So caching, reduces read load, read replicas, take the read load

6486.58 -> off of the primary database server. Killing reduces the right road, the right load. And

6494.54 -> it keeps you from losing critical messages. So let me whiteboard this out. So you can

6500.3 -> kind of see how he would scale the database. So wake up, everybody. Wake up, wake up, wake

6507.62 -> up. This is this is bonus content. This is actual architecture information, which goes

6511.889 -> way above anything that's covered in the course. So it's bonus bonus bonus. So make sure you're

6519.909 -> awake. So what does this really look like? If you're, let's say here, is a web server

6529.92 -> or group of servers. I go, I think in this direction. Some people think in the other

6537.25 -> direction both work depends on which country you're from. Let's say you got your web servers,

6541.98 -> and you got your app servers. On the way into the database,

6548.25 -> we put a kill. And Mike, we don't see your thing.

6552.969 -> Oh, thank you for that. So we set up a web server and app server and then we put a queue

6559.599 -> here. This could be SQS. It could be Apache Kafka, it really doesn't matter. We're going

6566.23 -> to use queue

6572.679 -> by doing using the queue, we can increase the scalability on the way into the primary

6579.21 -> database.

6591.42 -> And I'm going to show this for it's to make life simple. And this is going to reduce the

6597.57 -> right load Because web messages coming from the web server, the app server, instead of

6605.57 -> all being thrown to the database and potentially lost, we're going to add a queue. Now if we

6610.25 -> want to increase the capacity of these other right, the primary database reads, writes,

6617.3 -> reads writes. So if we want to offload all the read work from the database over here,

6624.87 -> we're going to we're going to add some read replicas.

6631.21 -> And now we're going to point all the traffic to the read replicas. Now the read replicas

6645.03 -> are doing the answers. But what if we want to reduce the load on the read replicas? Well,

6653.969 -> realistically speaking, we could stick a cache here. And the cache where I'm putting it here

6659.38 -> architecturally is debatable, I'm using it for simplicity and elegance to make it clear,

6664.17 -> what we're using it for the cache can reduce the load on the read replicas that are doing

6673.489 -> all the rework. So you know, that's the why we're using this. The cue reduces the right

6683.55 -> load on the cache and the read replicas reduce the read load on the system.

6698.51 -> think now's a good natural time to pause and answer some questions because I want to make

6702.09 -> sure everybody has a great learning.

6711.3 -> All right, give me a few seconds to find Sure, the questions.

6766.15 -> Can you speak how this integrates to an instance high levels for like instance, must because

6771.63 -> I don't know what that means it's missing. Cloud, I'd love to answer your question. But

6776.88 -> it's missing. What's missing the question you say how this integrates how that integrates?

6782.52 -> If you can help me, I'll be thrilled to answer. You may have asked it at the time during the

6789.51 -> presentation. But you know, I don't know what I was saying at the exact time. Chris, let's

6793.71 -> go to the next question. And I'm thrilled to answer that if cloud provides the context.

6804.53 -> Any options to my to mitigate the greater performance during a backup? Nope. It's just

6815.11 -> part of the system the way AWS has designed it.

6821.17 -> Can you pull out your data of a proprietary system? Lady Godiva? Absolutely. How difficult

6828.579 -> is it? Well, anytime you. Anytime you put stuff into a proprietary system, you pull

6837.61 -> it out, there's going to be challenges. There's going to be cost there's going to be development

6841.44 -> cost to kind of do these things. But yes, you can do it. Will information be lost in

6847.89 -> the process? Yes. Is it going to be a perfect migration? The answer is no. So it's best

6854.98 -> just to avoid proprietary in the first place. What's the difference between a web server

6861.57 -> and an app server? You go to www.go Cloud careers.com. You got a web server. Now when

6871 -> you sign in, as a student of our cloud architect career development, which teaches you how

6876.78 -> to be a real architect and gets you hired as an architect and changes your life forever,

6880.469 -> and gives you a great salary, our content is housed in the application server that defines

6887.21 -> the logic Are you allowed to get in? What are your permissions? What are your rights?

6892.219 -> That's the application where's the content delivering store? That's an application server.

6896.38 -> And then of course, your information would be in a database so you I like to, I like

6901.52 -> to call it into the this layer, the web server presents it to you. Your business logic runs

6908.23 -> on the application server, and the database stores your information. I hope that helps,

6915.8 -> because it's a great question.

6920.21 -> So I'm guessing that this comment that Clyde just put in is clarification.

6928.27 -> How does the up

6932.07 -> and up because you asked about clarification for what is this, I

6935.23 -> still don't know what that means code. Does anybody on my team could know what that means.

6945.09 -> So how this integrates to an instance, they're asking about how an app integrates with an

6954.98 -> instance,

6955.98 -> cloud, I really don't know what you mean. But if you're asking about how an app works

6960.6 -> with a virtual machine, then what we're talking about is somebody who writes code and the

6965.9 -> code sits on the operating system. Just like anything that you would on your computer.

6971.4 -> How does the app talk to a database over an application programming interface?

6975.77 -> Gotta hope Okay, that's how you that's that's the kind of API's Okay, okay.

6989.579 -> Does Apache Kafka have the same level of compatibility integrated with other services such as SQS

6996.17 -> Apache Kafka is the industry standard, and it's used everywhere. It's nothing to set

7002.79 -> up and use. Everybody uses it on all clouds and all data center.

7015.05 -> Difference between standard queue and first in first out? Yes, standard queue, stuff goes

7020.56 -> in the queue and as fast as it can be drained, it gets drained with no regards to message

7024.46 -> delivery. First In First Out, number of men message goes in and number one message goes

7029.92 -> out number two, message goes in number two, message go out. Number three, message goes

7033.77 -> in number three message goes on.

7061.199 -> Can the apps be accessed across to any apps can be accessed anywhere in the world, across

7065.56 -> any cloud that you want, as long as you've got IP connectivity, and you set it up right?

7083.11 -> I can't take too many unrelated I can't really take unrelated questions, but I will do do

7088.8 -> again the difference between a DevOps engineer and a cloud architect and please join us tomorrow.

7093.5 -> For the how to get your first cloud job on our Cloud Architect is a business executive.

7097.67 -> I'm gonna say it again, it is a business executive who is at least 5050 to 80% business and 20

7104.84 -> to 50%. Tech, that designs presents and sells the solution. We never touch the data. We

7109.75 -> are not a lot of code, we are not a lot of Configure. We're not a lot of types of systems.

7113.739 -> Ever. A DevOps engineer is a software engineer first, who then automate software release

7121.159 -> cycles. In order to make software development more agile DevOps and architecture have nothing

7126.71 -> to do with each other. They're kind of like a parrot and alliances are that different

7131.8 -> from each other. kind of fight FIFO close lose data. Any queue can lose data if the

7140.36 -> queue crashes. But as a rule, it doesn't matter whether it's FIFO. Or, or first in first out

7148.68 -> unless the application or the database for for some reason requires ordered delivery

7152.429 -> of messages. There's a load balanced and needed for an app server sometimes. If you need more

7162.82 -> than one server for redundancy, or performance purposes, which is generally Yes, you definitely

7168.02 -> might want to use a load balancer before we wrap this over.

7179.69 -> Okay, going back to the content, um, we're going to talk about extraction, translation

7188 -> and loading tools. I'm not a database professional. I'm an architect and have been for decades

7199.96 -> but it, there's times where you need to get data out of one database and put it into another

7205.46 -> database. And what we use for this is extraction, translation and loading tools. AWS has their

7212.25 -> own, it's called Glue, which is okay, but it's proprietary. And there are other industrial

7219.69 -> grade ETL tools that you can use across all your columns. And if you've got a lot of databases,

7230.88 -> because each database has its own strengths and weaknesses, and you want to pull information

7235.29 -> from one database to another database, you will use an extraction, translation and loading

7244.51 -> tool. And Amazon has their own proprietary branded one called Amazon glue. I don't really

7249.239 -> use proprietary anything when I don't have to, but they have one. It's fully managed

7254.34 -> server list tool. And you know, according to Amazon, it's real, real, real simple. You

7261.389 -> just point, your data, the glue to your data, it'll automatically take care of everything

7268.42 -> you needed to worry about. According to AWS, that's what they're going to tell you. It

7272.949 -> discovers the data and stores the metadata in the catalog. And after the data is cataloged.

7277.98 -> It's searchable, and variable. The data can be queried or stuck into another database.

7282.52 -> You know, what are we really talking about from an architecture perspective to see what

7287.21 -> it looks like? There we go. We've got data, for example, an object storage. And then it's

7295.15 -> going to stick it anywhere we need in which database we want. Athena redshift, EMR again,

7301.13 -> map reduction is basically to take stuff in and out of one, again, Python Spark, not Amazon

7305.69 -> EMR in most cases. And then you could visualize the data with quick side Power BI tabular

7310.71 -> or some other data visualization tool. Now, if, again, if you wanted to take your data

7323.02 -> from your own database, and then use one of these proprietary AWS databases, you could

7328.631 -> use something called the schema conversion tool. And what that does, is it takes your

7333.5 -> data from your database and massages it so it sort of fits into an AWS proprietary database.

7342.29 -> Now, no schema conversion is going to be perfect. So if you use this, it's going to take some

7349.8 -> work and development work from your database team. But you can use those. And basically,

7356.53 -> it helps you migrate your database to a format that's compatible with your target database.

7364.28 -> And if you started out with a really bad database choice, and you outgrow it, what you're talking

7370.489 -> about is, this is a great way to get your information into a better database, like Oracle,

7377.96 -> for example, or Maria dB, or some non proprietary database, where you could use this to move

7384.4 -> it into a proprietary database, such as Amazon, Aurora. But now you're on a stuck on a vendor

7389.88 -> proprietary system. But if you're already on an Oracle database, there's no reason to

7394.691 -> use this tool in the first place. It helps you get where you're looking for. So a schema

7403.63 -> conversion tool is really used to migrate between a heterogenous database to convert

7408.599 -> the schema. And it basically helps with warehouse application codes and SQL procedures, it's

7416.52 -> a nice tool to help you actually get to your goals. And use this type of a tool to move

7424.46 -> from one database to another. Now let's talk a little bit about high availability database

7432.02 -> design, at least as it pertains to your certification. In our cloud architect career development

7438.69 -> program, we get much much much deeper because we have the time we spend between five and

7442.36 -> 700 hours training a cloud architect, because that's really what it takes to get to get

7446.28 -> your first job as an architect, we've got 15 hours here, or less. So we've got to focus

7450.84 -> on what we can actually teach in this period of time, which is the exam. And I'm throwing

7455.46 -> in as many bonus nuggets of wisdom as I possibly can for you because I want you to have the

7461.489 -> best. So assuming a single cloud is highly available, which we don't believe in, but

7466.571 -> assuming you're gonna put it in a single cloud per your exam. You know, we're going to talk

7473.38 -> about how they recommend you design a high availability database on AWS. As a reminder

7481.119 -> from yesterday, AWS designs are things in the regions and availability zones, or regions,

7486.179 -> a large geographic area, such as parts of Europe, or half of the US continent by comparison.

7499.07 -> What we're talking about here You guys all still hear me? Chris, can you still see me?

7504.28 -> Because my YouTube that I'm using for monitor just went Yes. Okay. Okay. So according to

7513.09 -> AWS, if you want high availability, you can put your databases into two availability zones

7520.52 -> in two different data centers. And if one data center fails, you got a backup data center.

7528.69 -> Of course, if the cloud providers network fails, if they get hacked, or the control

7531.88 -> plane goes down, you lose everything. But this is an AWS certification exam. So you

7537 -> got to know the AWS principles. So I'm going to tell you that a high availability database

7541.76 -> design uses a multi AZ environment. So what happens is the database copies itself into

7549.699 -> another availability zone or data center. Now, by keeping a copy of your database in

7558.75 -> another availability zone, you do not get increased performance. It just has the information

7565.1 -> and it synchronizes it all the time. And should your primary database go down the one or your

7570.6 -> data center go down the backup in the other data center, otherwise known as availability

7577.01 -> zone will take over. So you copy a message goes to the database send you the cat, but

7582.219 -> some new cat food transaction stored in the database that gets copied to the database

7586.13 -> in the next availability zone. Kind of keep that in the back of your mind, and you're

7590.77 -> good to go. And it's copied synchronously. What does this look like architecturally,

7599.5 -> you've got the same thing in two different data centers, your same web servers, app servers

7605.44 -> and database servers. And what happens as your data is copied to the master, and it

7611.619 -> gets copied in availability, Zone A is copied to the standby and availability zone date.

7617.55 -> So that's really what we're talking about with regards to creating a high availability,

7620.91 -> if you want to call it that database architecture in a single cloud. So let's say you've got

7633.17 -> your database and availability zone one and your database and availability zone two. What's

7639.63 -> going to cause your database to failover from datacenter, one to datacenter. Two, well,

7645.489 -> if the primary database fails, meaning the server, guess what it's gonna shift to availability

7652.67 -> zone two, if the entire data center goes out, like a power failure, because both power companies,

7660.29 -> and both generators and backup generators and batteries backup style, then poof, your

7667.25 -> information is gonna go to the next one, where likely you have a network outage in the data

7673.02 -> center. If you change the database, enter the database instance, service type, poof,

7678.8 -> it's going to fell over. If you want to do maintenance, like patch, or upgrade the database.

7686.52 -> It'll fail over to the backup while you're doing the maintenance. Or if you issue a manual

7691.75 -> failover I'm going to reboot this reboot with failover you reboot it and the backup one

7696.829 -> takes over and that way you don't lose anything. Okay, we are now going to get to my favorite

7706.23 -> content, which is networking. But I'm going to pause for five minutes now to make sure

7711.68 -> that we don't have any other questions related to the database. Now we're gonna get to the

7719.38 -> fun stuff, which is networking, which we're gonna get a little geeky. I'm sure. That's

7723.76 -> okay. It's my favorite.

7724.929 -> All right, give me just a second here.

7732.469 -> caught me off guard with the shorts short slot there. Yeah. All right, here we go.

7764.679 -> Can the same load balancer be used for the app server and web server in the cloud there?

7769.01 -> Your app server and web servers are going to be on different subnets you're going to

7771.909 -> use a different load balancer.

7778.46 -> The database queue you mentioned is it something that comes with as part of the AWS database

7782.56 -> feature or it's an architectural concept? No, it's a both an architectural concept and

7787.45 -> it's a server concept. So either either you know you you use a sword over like a Kafka

7801.32 -> queue, or use the SQS service, but it is an actual functionality that is an architectural

7806.53 -> concept and a compute concept at the same time. So, one would use the schema conversion

7815.52 -> tool, if I want to go from my datacenters, Oracle server to Azure, Amazon Aurora, which

7823.801 -> I would never do, then I would use the schema conversion tool, whatever use glue, no. But

7829.88 -> for your exam, if I want to pull data from one of my databases like redshift, and pop

7835.32 -> it into Aurora, I might use glue because it kind of normalize your data. And that's a

7840.46 -> great question, sir.

7845.51 -> So you're saying that ensuring high availability multiple easy's I don't call that high availability,

7851.68 -> but it is port carding exam, high availability in multiple regions? Guess what? Still not

7857.56 -> high availability. Because you've got a single cloud, one hacker comes in and knocks down

7861.17 -> the cloud. And guess what, you lost everything. There's a big BGP problem on the AWS cloud

7867.059 -> like or like the Google Cloud was taken down from a BGP problem or Facebook was taken down

7871.27 -> like a whole day from a BGP problem where at least half a day or eight hours, something

7874.9 -> crazy like that, poof, whole cloud goes down. So I don't consider a single cloud ever a

7880.05 -> high availability system? I'm just telling you, what do multiple reasons do. It puts

7884.599 -> your stuff halfway in the US and halfway and Europe. But if the cloud goes down, and guess

7889.22 -> what you're still done. So that's not really a high availability, performance environment.

7897.239 -> Is a web server the same thing as a client server know, a client server could be a web

7901.59 -> server or an application server, a file server. A web server is a type of server that serves

7907.5 -> web pages.

7917.78 -> Okay, we'll get back to this. So Chris, or Alonzo, when are you guys in the chatbox?

7933.07 -> If you can help me, or when there was a song when I was younger, and it says y'all having

7936.23 -> a good time, then they go t t, t, t, t, t, t, I don't remember the song. But if you're

7940.53 -> all having a good time, you know, let's get back to this, please give me a hashtag that

7944.15 -> says AWS Solutions Architect Associate. So I know you're awake, alert and oriented, and

7953.48 -> blocked out to high availability is never in a single cloud. It's like putting all your

7957.88 -> money in a single basket, and then hoping that you don't get stolen from or the basket

7963.82 -> doesn't catch fire or you don't drop it or you don't lose it or nobody breaks into you.

7967.32 -> So no, central cloud is never high availability. So we're gonna now get into some basic, basic,

7978.88 -> basic networking. And again, if you guys didn't hear the hashtag, if you guys can put hashtag

7984.9 -> AWS solution, Architect Associate, we're not fans of any kind of acronyms around here.

7991.45 -> Kind of keep that in the back of your mind. So let's do a basic networking review. No

8000.73 -> basic, anything would start without the OSI model. The OSI or the open systems interconnect

8007.86 -> model is a model that network engineers, network architects, cloud engineers, cloud architects,

8014.25 -> and anybody who works in tech needs to know why. Because, you know, you've heard me be

8022.34 -> very unhappy with AWS marketing terms or Azure marketing firms or Google Marketing teams.

8027.13 -> Why? Because they complexity. Imagine a doctor trying to work in an environment where the

8034.36 -> same pill has 50 different names. How many people will die when Doctor one and Doctor

8041.46 -> two and Doctor three wouldn't even know they're on the same patients in the same medications

8044.87 -> people just die. In it's anything that matters depends upon clear language. And we must speak

8053.809 -> the same terminology. If we want to be serious. If I walk into a Chief Information Officers

8060.09 -> office and talk about s3 and EC two, I will be fired my replaceable thank me for the nice

8065.17 -> job and I will be escorted out of the room by security because the Chief Information

8068.73 -> Officer is not going to understand the gibberish that I'm talking about. Now if I talk to the

8073.341 -> chief information officer about his virtual machines and his object storage, now we're

8078.4 -> on the same page. So the OSI model is standard language that everyone uses in the networking

8086.92 -> world. It's a means to get rid of garbage marketing term, so we always always can communicate.

8094.25 -> If I speak to somebody in Tel Aviv, guess what? And I talked about a layer one problem

8099.21 -> in networking He or she is going to No, no, if I speak to Tyrone in South Africa, guess

8105.251 -> what, and I say we've got a layer two problem Tyrone's going to know what it is. And we

8110.321 -> must have precision language. And we're dealing with precision anything. The network is the

8115.07 -> heart of the cloud. It's not software is the network, network goes down, cloud goes down.

8121.159 -> And that's why organizations typically don't think of the network and those that don't

8126.74 -> think of the network pay a big price in terms of outages. And anybody that knows, networking

8132.61 -> knows you can't use a single service provider. Because we've never been allowed to for decades.

8138.01 -> So let's talk about it. And I'm going to whiteboard it out for you because I want you to see it.

8141.89 -> I'll share my screen here. layer one, the physical cable between you and your and your

8152.38 -> switch, or your router wire, whether it be fiber optic. If so, you know on a wire, we're

8159.949 -> sending electrons, technically, on fiber, we're sending light. So layer one that we're

8166.75 -> talking about is the physical layer, cable cable. Layer two is the data link layer. That's

8177.191 -> the actual hardware we're using. It's your Wi Fi card. It's your Ethernet card, hardware,

8185.929 -> hardware, whether it was a serial interface for a wine interface or an ISDN interface,

8191.96 -> or an Ethernet card. So layer one, why are layer two physical card hard coded address,

8200.59 -> hardware address. If you go to your computer it has a MAC address, layer two address physical

8206.63 -> wire, layer two data link hardware. Now next, we'll move up to the network layer where we've

8214.55 -> got a logical address your IP address 192168 1.3 That is an IP address. layer one wire,

8225.61 -> layer two hardware card layer three logical addressing. You can't really change the MAC

8230.5 -> address on your computer, but you can change its IP address now let's say we're talking

8238.229 -> about transport. The next layer four is transport. Do I send my data in a reliable fashion meaning

8244.58 -> TCP? IP? Am I sending UDP traffic for a real time traffic such as voice or video? Or am

8253.48 -> I sending a test message like a ping to a Windows computer? Ping Alonso's computer like

8261.54 -> an ICMP echo when he sends me an ICMP reply that you've got. Now realistically, when it

8269.469 -> comes to network engineers and network architects, that's it. layer one wire, layer two card,

8276.069 -> layer three logical address or IP address layer four TCP, UDP or ICMP. We're going to

8282.46 -> cover the rest. Now at layer five or the session layer, we're dealing with something called

8287.109 -> the socket which really controls the connection. At layer six, we're talking about presentation

8293.71 -> of data. But there is some networking that occurs here. Encryption, for example, occurs

8299.38 -> at layer six. And the applications are what you use you go to your web browser that is

8304.13 -> a layer seven application, think HTTP DNS sec. So layer one wire, layer two card, layer

8310.51 -> three, logical address layer four protocol, TCP, UDP, ICMP, layer five, session, layer

8316.501 -> six, presentation and encryption. layer seven is the application itself. Now when it comes

8326.5 -> to networking, whether it's in the data center, whether it's in the cloud, it's completely

8331.71 -> irrelevant. Everything needs an address. Why do you need to address you need to be able

8336.51 -> to communicate with the system? Let's say for example, I wanted to send Eva do IKEA

8342.09 -> a letter and thank her for some of the really great blogs that she collaborated with me

8347.33 -> on. By the way, on our website, there's some really great cloud architect interview question

8352.059 -> blogs, you could do I could work very closely with me on that. And she wrote a beautiful

8355.389 -> archinaut article on on edge computing, I recommend everybody read. So how are you going

8365.569 -> to know how to find it? Well, you got to know to go to www.co co careers.com. Otherwise,

8370.16 -> you're not gonna see these great articles. If I want you to send a letter to my mother,

8378.219 -> I need to know her address. How else would the post office know how to send my letter

8382.63 -> there? Well, when it comes to message delivering and computers, we need an IP address. It's

8387.06 -> basically a no different than the address in your house. And every address on your network

8393.53 -> must be unique. How will the mail system to work 123 Main Street I have a 123 Main Street

8402.39 -> in Philadelphia, New Town Bucks County, Ben Salem Bucks County Levittown Bucks County.

8409.13 -> Every city in the world has like a 123 Main Street, okay. But what's different, that makes

8415.85 -> that unique? The postal code or the zip thing, same thing with IP addressing every device

8421.75 -> needs a system that's going to talk to each other, which must be unique. Now, inside of

8427.97 -> your system, you can use private addresses, but they must be unique inside of your organization.

8432.69 -> And anything in the external internet also needs to be unique. Now when we deal with

8437.72 -> IP addressing, we're going to deal with two versions ipv4 and ipv6. ipv4 is the 32 bit

8445.59 -> address that we've been using for as long as I can remember, I've been using since the

8452.95 -> late 80s. And ipv6 was actually invented a long time ago, but we're starting to use ipv6

8460.391 -> addresses, the world still hasn't even adopted 3040 years later.

8469.47 -> So when you pop addresses for your VPC, inside of your VPC are going to use private addresses.

8479.64 -> If you want to know Private Addressing, you truly want to understand it, I recommend you

8483.571 -> read and you should all read this, the Internet Engineering Task Force RFC Request for Comments

8489.37 -> 1918 that as the specification for IP addresses, that all network Architects like me. And inside

8498.939 -> of that they specified it internally. Because we don't have enough IP addresses. Organizations

8504.841 -> should use the 10 dot zero slash eight address base, the 172 16 dot 0.0, all the way to 172

8514.95 -> 31 dot 0.0 slash 16, which can also be summarized or aggregated into 172 16 dot 0.0, slash 12.

8523.99 -> And the 192 168 dot 0.0 Slash 16 address base. These are private IP addresses to be used

8530.24 -> inside your organization. And they are not globally relevant. I'm going to mention this

8538.05 -> right now just because we're going to talk about classless inter domain routing for a

8541.88 -> little bit. We used to have these things 3040 years ago called IP classes. And that was

8548.64 -> basically meaning where every single network used a specified subnet mask. So let's say

8557.61 -> we had the one dot 0.0 slash eight network. We had 16 million addresses on that address.

8563.18 -> But here's the problem. Every card on our router needs to be on a different subnet.

8569.69 -> So if we use four different slash eights, or classic addresses, we'd be burning through

8575.69 -> 64 million IP addresses. Now the Class B address base had a slash 16 and was from the 128 dot

8589.48 -> 0.0 all the way to 190 1.255255255. Now guess what? If you do that each subnet uses 65,534

8603.31 -> addresses with 535. I think it's 534. And that would basically if so if you had four

8610.649 -> different subnets, or routers on a card, you would actually burn through 102 260,000 addresses

8616.601 -> again, it would be ridiculous. Now a Class C address, which had a slash 24, which was

8623.58 -> in 190 2.0 dot zeros all the way to the 223 to 255255255255 Slash 24 would be 254 addresses.

8631.45 -> And of course, there's the Class D address spaces were used for IP multicast. And that's

8637.45 -> the TT 4.0 dot 0.0 and the 239 dot 255255 2.5. Nobody uses IP classes. We're using classless

8644.609 -> inter domain routing since then that really means just subnetting. Modern Times classes

8651.17 -> addresses and routers are going to build a map of the network. And what happens is matter.

8657.52 -> Routers are going to have a table in them. And they're going to say to reach the 192

8661.65 -> 168 dot 1.0 slash 32. Take interface X to reach the 192 168 1.4 subnet takes interface

8673.39 -> y to reach the 192 168 1.8 slash 32. So slash 30 subnet reach zero slash 30 subnets. I'm

8683.58 -> giving you right now take interface Zed. And that's what we're talking about. And that's

8690.35 -> what we're talking about is subnetting because we have to optimize our IP addresses space

8694.71 -> and we can't waste it. Every interface needs to be there. So let's Let's say for example,

8701.3 -> we used a single class C IP address 192 168 1.0 slash 24, which gives us 254 hosts. And

8714.21 -> we had one subnet, which was the 192 168 1.0, slash 28. And then the next subnet would be

8719.89 -> 192 168 dot 1.16, slash 28. And then the next subnet would be 192 161 68.1 dot 32 slash

8728.61 -> 28. And I did a free subnetting webinar, somebody from my team posts the free subnetting webinar

8734.58 -> inside of this chat box to help people get to their goals. Because I can cover the four

8740.77 -> hours of that webinar over here. But I'll also have a couple of examples. Here, I decided

8752.43 -> to create slash 28. So that same IP address that I showed you, and you can see the different

8759.5 -> subnets. Actually, do you want me to do a subnetting webinar, if you want me to do a

8763.84 -> subnetting webinar, do two things type hashtag subnetting webinar? And also Chris will do

8770.95 -> a poll to see if we got enough of you, we'll probably do something anyway.

8774.729 -> Yeah, I'll put a poll in the chat box for everybody if you want to subnetting web.

8784.29 -> So let's see how many people pull it and how many hashtags subnetting webinars we got because

8788.23 -> you want it I'll do it

8789.23 -> don't know when will fit in the schedule, but I'll find a way to do it.

8802.27 -> And as you can see, I submitted the slash 28 into multiple smaller subnets.

8817.68 -> Now if subnetting is taking a big network and chopping it down to little networks, of

8823.939 -> course, we got to do the opposite, right? We've got to be able to take multiple small

8829.63 -> networks and bring it into a single big network. Why would we do that? Let's go back to what

8840.66 -> I just showed you see how we have all these slash 28 subnets. Now if you've got a direct

8849.53 -> connection to AWS and to your VPC, you can only give them 100 routes. Could you imagine,

8854.96 -> you know, we've already using all these routes over here. But we can summarize that into

8860.979 -> a single route. Because when we only have 100 routes, we gotta get real, real crap creative.

8867.24 -> So super netting, which is done for route summarization to reduce the memory load on

8871.38 -> the routers, and the CPU load on the routers is the exact opposite of subnetting. And it

8877.64 -> is absolutely critical. Anytime we're dealing with those, you're only going to hear this

8882.569 -> here. You won't realistically, you'll barely see it on AWS advanced networking, because

8887.14 -> it's so basic, it's not even worth your time. But you need to know how to do this for an

8892.31 -> architect need to know this because if they get this wrong, the whole system falls apart.

8897.8 -> So Super NES take many small subnets and combine it into a giant subnet. And it's really done

8904.85 -> for the router. So here's an example. Here we've got 192 168 dot 0.01 92 168 dot 1.0

8916.45 -> slash 2004 192 168 dot 2.0 Slash 24 and 192 168 dot 3.0 slash 24. And then we summarize

8927.63 -> that into 192 168 dot 0.0 slash 20. This is classless inter domain routing, subnet down

8934.109 -> supranet up, and it's all related to your traffic engineering that your network architects

8938.561 -> and your cloud network architects and your cloud network engineers need to be careful

8942.25 -> of and believe me, without the cloud networking people, everything will fall apart. Next thing

8953.62 -> we're going to deal with is ipv6 addresses. Again, it is a new form of IP addressing new

8959.42 -> is in the last 20 to 30 years old, and people are starting to use it. Like everything in

8965.11 -> tech. It moves super slow. We all think it moves fast because the networking marketing

8970.42 -> vendors keep changing the name of the same old things. But the things that I worked on

8975.04 -> in 1996 are the same things I worked on today. Of course now it's better, faster, cheaper

8980.03 -> and more reliable, but there is not. So ipv6 addresses are just a newer form. And every

8989.16 -> interface, as I mentioned previously is assigned an ipv6 Global Address. Where do we typically

8995 -> use ipv6 addresses and mobile phones? Now have IP Six addresses with a 32 bit binary

9002.58 -> meaning 01 for 32 bits. ipv6 uses 128 bit hexadecimal address, binary 01 hexadecimal

9018.22 -> 0123456789 Alpha Bravo, Charlie, Delta Echo Foxtrot. So we've got 16 vs two, so 16 to

9031.649 -> the 120/8 power versus two to the 32nd power. See the difference here are talking scalability,

9039.62 -> hugest. Chris, the next thing that we want to cover is the virtual private cloud otherwise

9045.729 -> known as the virtual private data center. Where are my time? Was

9056.79 -> it gone for about 20 minutes.

9058.13 -> Okay, let me take a few minutes of questions before we get to VPC?

9085.65 -> Good question, what's the difference between a high availability failover site and a disaster

9089.37 -> recovery site? A failover site is you have a data center over here. And you've got a

9095.88 -> data center here. And if this fails, everything goes to this one. That's high availability

9100.99 -> failover. And disaster recovery site is a complete and total or partial backup of your

9108.59 -> systems either ready to go for failover or in an in a manner where it's just stored and

9114.84 -> not ready to go. So for example, if I'm going to use a physical data center, and I've got

9121.08 -> two data centers, one in New York, and one in New Jersey, and one in physical Philadelphia,

9126.09 -> I've got three data centers in a close proximity of a nuclear bomb, were to attack or a massive

9134.22 -> earthquake that covers that small area, or a massive hurricane were to come. And all

9143.3 -> those environments go down. I'm done. Now, by comparison to a high availability failover

9151.12 -> site, a disaster recovery site is typically like 1000 plus miles away. And it's owned

9156.96 -> by different people under different people's control. So AWS would say, Hey, you can have

9163.05 -> your stuff in US east, west, east and US West, and you can back it up to Europe. And then

9167.94 -> when the AWS code goes down, you got nothing, something my grandmother would call book us.

9173.12 -> For those of you that know what I mean. By comparison, if I was just using, if I was

9179.75 -> using the AWS cloud on the Azure Cloud for high availability, I might store all my data

9185.05 -> into Google. And that way, and all my virtual machines into Google on that with AWS health

9190.229 -> and Archerfield, I got a backup cloud. It will be on your exam, but you should never

9195.67 -> ever, ever, ever, ever, ever, ever do your disaster recovery in the same cloud as your

9200.78 -> systems. Because that's like putting all your eggs in one basket all over again. So I hope

9204.95 -> I answered your question

9208.72 -> What's the main difference between TCP and UDP? Huge difference. TCP is reliable. So

9218.93 -> Chad, I send you a message and you say got it, like send me a message to I sent a message

9224.33 -> to and Chad you say got it. And then said if you got to let me speed it up, we're gonna

9229.189 -> send you messages three and four. And you say got him and then I'm gonna send you some

9232.24 -> more. I'm gonna let you send you messages 5678 You're gonna say got it Mike. And then

9237.191 -> I'm going to send you 910 11 and 12 and you don't respond. So I'm going to resend you

9242.96 -> 910 11 and 12 until you respond. Of course I'm going to slow it down and listen you want

9248.151 -> message after your last one by comparison UDP as me sending you data as fast as I can

9253.55 -> and I don't care if you receive it or not. So if I'm going to send you something really

9259.1 -> critical that I need acknowledgement on it's going to be sent via TCP. So TCP is typically

9266.399 -> sent for files for example now What if his voice What if I said you my cat Cindy is beautiful.

9275.43 -> And I and you lost the word my cuts and you got my cut Cindy is you lost is and how beautiful

9282.38 -> my cuts end beautiful. That's voice that's like your cell phone for example or video

9287.75 -> on Netflix. It pixels up for a second and it goes back to normal that's supposed to

9293.109 -> be TCP retransmits so now what if I said to you, beautiful is my cat Cindy, because we

9298.361 -> lost them in the messages and we say em I'd still be able to interpret it but you so for

9302.25 -> voice and video, it's always UDP. For reliable transport of anything mission critical like

9309.38 -> a file, it's always done via TCP quick question.

9320.5 -> Good failover site Florence runs at the same time as the original site, and the data should

9325.17 -> be synchronized and identical. Yes. Good thinking. Kira and Charles, what are the limitations

9335.55 -> of ipv4, realistically speaking, the number of IP addresses two to the 32nd power, major

9341.33 -> problem. Otherwise, it's perfect. What are some challenges associated with transitioning

9347.21 -> from ipv4 and ipv6 Bri addressing your systems, setting up the routing protocols and rebuilding

9355.04 -> the routing table? That doesn't sound like much, but I'm going to tell you, cci is like

9360.13 -> me had been cleaning up people's IP addressing scheme massive mistakes for decades, what

9365.171 -> happens is typically, they have somebody that's like a sysadmin, or a programmer that thinks

9370.46 -> they know IP addressing. And it's not that these people aren't smart programmers, no

9375.109 -> programming sysadmins no systems administration. And the person that sets up the IP addressing

9380.7 -> plan needs to be the best network architect in the entire building. Because it takes so

9385.48 -> much networking capabilities. In order to be able to do these things. They need to be

9389.399 -> the most senior person because no IP addresses are kind of like the roads in a city. If the

9394.78 -> roads are designed poorly traffic is traffic jams all the time. So the readdressing as

9400.33 -> the main plan and resetting up the interior gateway protocols such as OSPF or many systems

9404.979 -> to intermediate systems to deal with it. And the exterior gateway protocols such as BGP

9409.13 -> to deal with a new ipv6 address family, not that it's nothing to somebody like me, because

9414.05 -> I've been dealing with this forever. But it's something that's going to take the help of

9418.16 -> a strong network engineer, at minimum or on a great network architect. Good question.

9428.779 -> How do you create an IP address? I don't know what you mean by that. How do you assign an

9433.83 -> IP address is going to typically done via DHCP or Dynamic Host Configuration Protocol.

9443.66 -> Okay, I'm gonna get back to the content. But if you want a subnetting webinar, please vote

9450.95 -> for the subnetting webinar. So vote for the subnetting webinar, let me know I'm hashtag

9460.01 -> AWS Solution Architect Associate as well, as you know, click on that vote for the subnetting

9466.25 -> webinar, we know your way. So we're gonna you know, the AWS VPC section is pretty deep.

9475.4 -> And because we're now dealing with your virtual private data center, which was all that a

9479.54 -> cloud is, what I'm going to do is, is cover as much as I can today, we may go over a little

9484.34 -> bit, because I'm trying to avoid Saturday for everybody. And I want to give you guys

9488.14 -> even if we go a little longer. So let's talk about the components of the VPC, its routing.

9497.37 -> It's routers that connect to the Internet called Internet gateways. We'll talk about

9501.8 -> egress only internet gateways, not instances and not gateways. We'll talk about elastic

9506.46 -> IP addresses, VPC endpoints, VPC, peering access control lists, specifically network

9511.97 -> access control lists, and security group.

9520.36 -> To begin, first, let's talk about routing tables. How do you get your traffic to its

9525.67 -> destination? The routers doing? So how do the routers do it? Typically speaking, they

9531.41 -> run a little protocol, they all talk to each other. And they tell each other the routes,

9535.76 -> they build a map of the network. And when it comes to routing, there's going to be two

9540.59 -> ways where you can build your routing tables. Option one, you manually tell it. So for any

9546.21 -> of you that are like 50, or I'm not quite 50, but any of you guys that are between 40

9550.7 -> and 65. Perhaps you remember you wanted to go to your friend, Billy Bob's house or Julie's

9557.08 -> house or Sarah's house and you didn't know where they lived and they gave you their address.

9562.64 -> And you picked up a physical piece of paper map and you looked at it. And then you wrote

9567.399 -> down some paper that said I 95 for 22 miles north. Get off at the i 95 Exit take route

9575.689 -> to 22 for three miles east. Then make a right onto 123 Main Street at her house is 22 123

9583.97 -> Main Street, wrote it down via paper. Guess what? That was great. And then poof, we're

9591.479 -> trying to drive to our friend Julie's house and the road is blocked by police officers.

9596.05 -> Now we don't know how to get there. And they'll point you to some detour and you get lost

9599.97 -> Four hours later, you get there, you're all frustrated, you're just came the invention

9605.86 -> of the GPS, recalculating, and they got to the destination. So with routers, we've got

9612.069 -> two options, we can use a dynamic routing protocol called BGP, make sure you follow

9616.62 -> me on LinkedIn, pop this thing out, I'm gonna get you some unbelievable blidi pre training

9621.649 -> completely free. I'm gonna release it real soon. Follow me on LinkedIn if you want to

9627 -> get it. And so realistically speaking, anything that matters is going to use dynamic routing

9639.17 -> protocols. But the routers are going to build a map. And here's what the map is going to

9643.78 -> look like. It's going to say, hey, to reach the 172 16 dot 1.0. subnet, it's right here.

9648.819 -> I'm already on that subnet. So it's local. To reach the 192 168 dot 0.0, subnet, use

9655.399 -> this interface. Papa Charlie X Ray 123456. To reach the one to 168 1.0 slash 24, which

9665.61 -> is more specific than the previous one, reach the Papa Charlie X ray 654321 interface. I

9673.34 -> want to go to the internet. Look, notice we have what's called the default route. So all

9676.511 -> zeros, which says if you don't know where to go, go here. Reach out Internet Gateway,

9682.899 -> India golf whiskey 123456. So kind of keep that in the back of your mind. I use phonetic

9689.149 -> alphabets constantly. Because, again, I'm all about precision language, a language that's

9692.98 -> used in every country in the world to make things simple.

9699.569 -> Now, what is it looking like in the enterprise, I'm wanting you to guys have some more knowledge.

9707.3 -> Typically speaking, we're going to have what's called an Interior Gateway Protocol that's

9710.979 -> optimized for speed. And you know, there were lots of internet Interior Gateway routing

9717.25 -> protocols. Over the years there was rip there was roughly two there was IGRP, EIGRP, OSPF

9723.859 -> and immediate systems intermediate systems. In today's world, it's either OSPF which is

9730.74 -> what most enterprises uses, or the global Internet service providers, we use OSPF or

9736.37 -> intermediate systems, intermediate systems. So what you'll see is, these are dynamic routing

9740.87 -> protocols, the organizations themselves internally will run their own Interior Gateway Protocol

9746.069 -> that's optimized for speed. And we'll run what's called an exterior gateway protocol,

9752.91 -> which is what's used to connect to internal external entities. So Interior Gateway Protocol

9760.641 -> locally, exterior gateway protocol between things. When an organization connects their

9766.01 -> data center to the cloud. Inside of their data center, they're running an Interior Gateway

9770.74 -> Protocol, like OSPF. And when they connect to the cloud provider to exchange routing

9776.32 -> information, they're using BGP, because that's really the only exterior gateway protocol

9781.25 -> we use in modern times. So that's the way that kind of looks. And I'm going to briefly

9789.8 -> graze over BGP, because it'll take me at least four hours to do a BGP workshop. And I've

9796.609 -> got a really beautiful document coming for you please follow me on LinkedIn, you don't

9800.05 -> want to miss this document. So when you're connecting to AWS, overthinker, direct connection

9808.25 -> or potentially even a VPN, you've got to find a way to exchange routing information, why

9813.38 -> your data center won't be able to reach them if you don't have the routes. And the cloud

9816.99 -> won't be able to reach your users if it doesn't have their route. So you need to get the routing

9820.84 -> work. And it's going to be via BGP, and AWS, as well as all cloud providers support connecting

9827.21 -> to them via BGP, because it's, it's your organization's and the cloud providers exterior to you. We

9834.63 -> use BGP because it's incredibly tunable and highly scalable. We can do all kinds of traffic

9839.68 -> engineering, a, take this link to go here, take this and load share back and forth BGP

9845.06 -> is amazing. It's beautiful. I've got over 10,000 hours of experience with it, and I

9849.27 -> love it. And I use it for everything as the old network architect. test question for you

9854.44 -> might see it, you'll see you'll definitely see it on the AWS advanced networking, which

9857.72 -> I don't recommend you take because it's too basic. You'll likely see it on one of either

9861.59 -> the Certified Solution Architect, associate or professional, probably the professional

9865.33 -> but you may see it on either one. BGP uses TCP port 179. I'm gonna say it again, BGP

9872.27 -> uses TCP port 179. Why do I say this to you? Well, chances are, you're gonna have a firewall,

9880.36 -> right? Or an access list somewhere for security. If you need to connect and you've got a firewall

9885.529 -> between you and AWS, and you don't allow TCP ports 179 Guess what? BGP connection will

9894.21 -> close. No traffic will get anywhere. Also with BGP, you will require it Autumn autonomous

9901.149 -> system to identify your organization. When you connect to BGP, you must use it with direct

9911.91 -> connections. And there's a tremendous amount of tuning options, but document that I'm releasing

9916.779 -> will explain every last one on. AWS also supports the community, no expert, which is great if

9924.75 -> you don't want to become transit, which is way beyond the concept here. It's also far

9928.42 -> beyond the AWS advanced networking, we're getting into CCIE concepts, which I can't

9932.319 -> do in the short period of time. But what it means is this. If I tell Chris information

9940.74 -> about my routes, and Chris tells Alonso information about my routes, Alonso can reach all of my

9947.11 -> routes through Chris, that's called transit. If I tell Chris, Chris Do not tell Alonzo

9954.35 -> about any of these routes, Alonso can't reach me through Chris. So the note export community

9962.06 -> means don't tell your people downstream or upstream any routes that you've learned from

9967.66 -> this provider, and AWS supports. AWS implementation supports very basic BGP implementation, but

9976.12 -> they do support weight local preference as path specific as your routing information

9979.64 -> etc. And they only only will let you use 100 routes. So 100 route is nothing. I worked

9990.23 -> on networks years ago that had 20 and 30,000 routes and that's before right when Robert

9994.18 -> for small with like a couple 100 megahertz CPUs in them, kind of put that into context,

10001.96 -> megahertz. So it's all about using the right IP addresses.

10012.62 -> Let's talk about an Internet Gateway, what's a gateway everybody, it's just a router. A

10017.47 -> gateway is a router. So if you want to connect to the internet, you need a router that connects

10023.25 -> to the internet, right? So AWS calls the routers that connect to the internet, internet gateways.

10029.37 -> Good logical term truthfully told the really owes. An AWS will tell you there's no bandwidth

10037.12 -> constraints or performance limitations and your Internet Gateway. There's always performance

10041.29 -> limitations, but for the most part, it's about the speed you need. And it's just a router

10045 -> that connects to the internet. Here's how you create one it's very simple. You basically

10052.84 -> go to the management console CLI and you attach an internet gateway to your VPC, which is

10057.13 -> a virtual private data center, you create a default route, which was that zero dot 0.0

10061.41 -> route and you send all unknown traffic to that Internet Gateway, you need a public IP

10068.55 -> address on the Internet Gateway. As well as any systems that need to be reachable from

10074.399 -> the internet now, an Internet Gateway means the system that are behind it with a public

10078.55 -> address are reachable from the internet. Internet Gateway reachable inbound and outbound from

10086.02 -> the internet both ways. Now, what do you guys think this means? Means hackers can find you.

10093.311 -> hackers can find you. So Internet Gateway means you're available to be hacked. So he

10100.28 -> has the right next generation firewalls, intrusion detection, intrusion prevention systems, etc.

10106.25 -> Set up your demilitarized zones intelligently none of that's covered in the AWS Certified

10110.681 -> Solution Architect, associate or professional. But if you're going to be an architect, you

10114.109 -> need to know these things. We teach it obviously in our cloud architect career development

10117.46 -> program. So just understand, Internet Gateway means reasonable in and out. What's it look

10125.88 -> like over here architecturally, according to AWS looks very simple.

10132.79 -> You've got your virtual private data center, which you can see as you've got some virtual

10139.21 -> machines.

10140.81 -> Let's say the they're all behind a load balancer behind the router, the public address on the

10147.16 -> load balancer would have this IP address of 3.3 dot 3.3. It's a public address, which

10152 -> is routable to the internet, all systems behind this which all fall in the cider range of

10157.34 -> 172 16 dot 0.0 slash 16. If they don't know what to go, they look at the routing table.

10163.84 -> So just look at this. You can see two subnets over here. 172 16 dot zero slash 20 for which

10170.59 -> the routing table shows us local in the upper right hand corner, the 172 16 dot 0.0 dot

10176.39 -> 3.0 Slash 24 is also local because it's inside of our environment. And there you go. That's

10182.801 -> your default route 0.0 dot 0.0 slash zero which says if you don't know where to go,

10187.29 -> go to the Internet Gateway. That's all we're talking about. Same thing we've done routers

10196.32 -> internet, own internet gateways all that ingress which means coming in an egress which comes

10204.14 -> out, traffic comes in, you can be hacked so secure accordingly. Now let's talk about something

10213.05 -> called an egress only Internet Gateway egress means it allows your traffic go out, but it

10219.979 -> doesn't allow any traffic to come in. So an egress only Internet Gateway is designed for

10226 -> your systems that use ipv6. And it allows your systems to go out to the internet, maybe

10233.27 -> update their operating system, download patches, etc. But it doesn't allow external traffic

10238.819 -> in. So it's much more secure than using Internet Gateway. It's still be using firewalls and

10243.93 -> things behind it anyway. But keep that in mind. Internet gateways are stateful. What

10253.05 -> does this mean? So here's the thing. I'm sitting behind my phone, my phone just became a firewall.

10261.279 -> And I want to go to the cindy.com website to see photos of the beautiful scenery of

10268.79 -> the cat doing cat things, jumping up, sleeping, wherever it's needed to cut, though she has

10274.6 -> all kinds of fun per day. She looks like this, by the way. But my phone is the firewall and

10282.34 -> I want to go to the cindy.com website. I type www.sedar.com on my website. DNS tells me

10289.77 -> the ipv6 ipv4, ipv6 of the website and my traffic hits my default gateway goes through

10295.5 -> my firewall and out to the internet. Now when I stuck my traffic through the firewall, the

10303.13 -> firewall paid attention to me and it says Mike Gibbs is going to www dot send you the

10309.97 -> cat.com. So my my request goes through the firewall, it pierces the firewall, it goes

10316.279 -> out to the internet to the sydney.com website. sydney.com says, here's the photo of me your

10321.59 -> requested Mike, or daddy, whatever you want to use, it goes through the firewall, and

10326.29 -> it comes back to me. Why is her traffic allowed through the firewall, because the firewall

10332.22 -> has a table that says Mike Gibbs went to www.sedar.com sydney.com was answering Mike Gibbs so allow

10339.42 -> the traffic. Now, conversely, we now have a hacker on the internet wants to get the

10344.989 -> mic get his computer that's behind the firewall, he or she sends their traffic to the firewall

10350.5 -> to me to the firewall and the firewall says denied. Next hacker comes in denied next hacker

10355.8 -> tries to come in denied. And why is that because the firewall doesn't know about it, it has

10361.47 -> no state. So the state is merely tracking that. What happens if the firewall saw me

10369.88 -> go through the connection, remember that it's stateful. That's what stateful means. So egress

10379.24 -> only internet gateways allow your traffic out to the internet and returning back. But

10382.99 -> it does not allow me to connectivity. And that's typically used for your host to get

10387.979 -> back to get patches. Not instances. Well, realistically speaking, not instance, is something

10397.75 -> that in the old days, you would have stuck behind your Internet Gateway, which would

10402.84 -> translate your private addresses into public addresses. Now, the reality is AWS has a better

10411.04 -> service called the NAT gateway, we'll get to this but you might still need to create

10414.439 -> your own version of a NAT instance, even if it's not an AWS NAT instance, Nat translates

10419.25 -> one address to another. Why might you need to do this company a address their systems,

10427.16 -> you then 10 dot 0.0 address base. Company A is a big, strong, powerful company and it

10432.359 -> just bought Company B and get switched up IP addresses, company B is using 10 dot 0.0

10437.72 -> slash eight. Now I told you your address base needed to be unique between them. So if Company

10444.92 -> A wants to talk to Company B, and they use the same IP addresses, they got a problem

10448.25 -> doesn't work. So we might use to use NAT or network address translation to translate these

10454.25 -> addresses into different addresses so the systems can talk to each other. And there's

10458.13 -> multiple forms of Nat, one to one NAT one of many NAT static NAT dynamic NAT, Pat, I

10461.74 -> mean, most of those aren't even covered in the AWS advanced networking, but you may need

10466.229 -> to do them. So critical information for the cloud architect or the cloud network architect.

10471.17 -> And that's why we have training this way beyond this, but we're focused on certifications

10474.24 -> in order to pass the exam. So in that instance, is available as an AMI, meaning an image that

10483.1 -> you can use from AWS or you could create your own NAT instance if you needed to do not.

10487.67 -> But if you're going to use not connecting to analysis to connect to the internet, you

10491.07 -> have to put it in a public subnet. And basically it's gonna have a default route to the gateway.

10494.899 -> And all your hosts will have a default route to the will basically have a default route

10499.93 -> to The NAT instance, which will then have a default route to the NAT gateway that's

10504.83 -> going to look like this gotcha uses in a private subnet. They want to go to the internet there

10509.35 -> have a route that says go to the NAT instance, the NAT instance send your traffic to the

10514.84 -> NAT gateway, which sends your traffic out there to the internet. Now that was getting

10524.08 -> pretty complicated. So AWS decided to come up with a simpler solution. And they came

10529.75 -> up with a NAT gateway, which is a fully managed service. The NAT gateway connects you to the

10535.96 -> internet and translate your private addresses into a public address. Now NAT gateways do

10542.57 -> something that's called Pat, or port address translation is also known as NAT overload,

10548.41 -> where they translate a tremendous number of addresses into a single address. And they

10553.21 -> do that by using an IP address plus a port number to separate each IP address. If you

10558.42 -> want to know more about that, we've got a free CCNA course that you can actually see

10563.14 -> on our website, my team and pop a link to them. So what happens is a NAT as a NAT gateway

10570.42 -> is fully redundant inside of availability zones, if you've got two availability zones,

10574.22 -> guess what you need two NAT gateways, you put it in a public subnet, and it's got a

10578.05 -> public IP automatically assigned to it for the life of the gateway. And basically, you

10583.83 -> just give all your systems a route to the NAT gateway, it's kind of like a NAT instance,

10589.54 -> Internet Gateway all in one without you having to think about the security concerns to the

10593.729 -> same degree as you would with a not a not an Internet Gateway, which basically provides

10598.12 -> full internet access behind the NAT gateway. So your systems are here, you attach to the

10602.23 -> NAT gateway, and it provides Internet access and not services network address translation

10607.24 -> at the same time.

10613.42 -> Now, on your computer, if you want to connect it to the internet, you got to plug it into

10617.83 -> the network, right, the Ethernet card on your computer. Or you could use Wi Fi. Tip secret

10625.24 -> here. If it matters, you don't use wireless ever, you're not going to see a data center

10629.53 -> built on wireless. They wire things out. If you want to go see a concert, chances are

10636.52 -> most of the stuff behind the scenes is going to be wired certain things are going to be

10640.33 -> wireless that they have no choice but most everything's going to be wired. Why because

10644.5 -> wires are more responsible than wireless are more reliable than wireless. So every computer

10653.2 -> needs to be plugged into the network. And what is it used is called a network interface.

10658 -> So when we deal with AWS, their marketing folks call it an elastic network interface,

10663.5 -> an elastic network interface, there's just an Ethernet port, or a virtual ethernet port.

10668.8 -> And by default, you basically would set up, you turn on your system and it comes with

10675.92 -> one network interface. Now there are times where you might want to put a system on two

10683.58 -> different subnets at the same time. And you can basically put two coasts. And if you did

10689.399 -> that two subnets for us to interest fail, you'll see a lot of people teaching you to

10694.1 -> do this with a bastion host, which is one of the worst security things you could potentially

10697.72 -> do. Except for not the Azure bastion host. But the way most people do it, I'm not even

10702.05 -> going to cover that just don't make a bastion host. We've got a video on why you shouldn't

10706 -> create Bastion hosts at least for most parts like they teach in the certification. But

10712.35 -> done it intelligently it can be used. Actually, there's no intelligent way to do it. What

10717.85 -> happens with a bastion host is you stick a host on the internet on the public internet

10721.739 -> with two network cards, and it's got a backdoor into your private systems. You can SSH to

10726.63 -> this thing on the internet, wide open on the internet. And then you can backdoor into your

10731 -> systems but so can any hacker. But there are business reasons. Maybe you've got users on

10736.39 -> two different subnets and maximum performance. You don't want to route between subnets in

10741.61 -> your router that there are reasons you multihomed thing. Or you can create a private management

10745.62 -> network and manage things over the management network. There's there's lots of reasons you

10751.029 -> need a multihomed service. Now, if you need a public address, like on a web server, or

10758.71 -> on a load balancer, if there's multiple web servers behind the load balancer, we're going

10764.25 -> to public IP address you're going to need one. So what do you think a public IP address

10768.34 -> is called? An elastic IP address because the marketing terms of the word elastic. So what

10775.41 -> what is an elastic IP address? It's a public address that you borrow from AWS and you keep

10781.73 -> it as long as you need it. And when you're finished with it, you just have to return

10787.75 -> it. AWS is global addressable, and it gets given to another user when they're ready.

10793.56 -> An elastic IP address can be a single public address. It can be it can make could have

10798.491 -> a public address that's mapped to Many private addresses, as with Nat overload, otherwise

10802.359 -> known as port address translation with a NAT gateway. And we can can set it up in multiple

10808.91 -> ways. Here's what it actually looks like, architecturally speaking.

10818.42 -> You can see you've got your systems. And the systems basically have a public IP address.

10826.649 -> That's

10829.21 -> Chris, how long have I been talking?

10839.85 -> Not sure, actually, I just assumed you're going to be finishing out but it's been about

10844.16 -> 20, maybe 25 minutes, maybe.

10849.71 -> What I'm going to do is I'm going to cover endpoints. And then after endpoints, I'm going

10858.2 -> to stop it, I'll take some questions. And I think we probably should finish from there,

10861.66 -> unless people want a little more. So let's discuss endpoints. And planes are a way to

10869.54 -> connect things to each other. That's why it's called then. And we're going to be dealing

10875.88 -> with two kinds of VPC endpoints. And endpoints are used to allow your V PC to connect to

10881.96 -> another AWS service, or an another network. We use endpoints because their performance

10888.2 -> is better, the latency is lower, and the security and cost is better than going to the end and

10893.42 -> going to the internet. Here's an example of an endpoint and its action. Let's say you've

10898.77 -> got your VPC, your servers there want to communicate with object storage, otherwise known as AWS

10905.88 -> s3, there's two ways you could do it, you could send your traffic out to the internet,

10911.88 -> and back into s3. Now, the internet's not secure. So you'd have to encrypt your traffic.

10918.67 -> And here's the scary part, you have to pay to send your traffic to the internet, the

10922.172 -> internet performance is slow, and not guaranteed. And then it would come back to AWS. Or you

10929.79 -> could do the other option, which is you could just send your traffic across the AWS network.

10934.3 -> And that's the point of the implant communication. You have no control over the AWS network.

10940.83 -> But you, you have no control over the internet, and you have no control over the AWS network.

10947.561 -> But AWS can control the performance of their network, but they can't control the internet.

10952.03 -> So endpoints are gonna have lower latency but our communications etc.

10960.17 -> So endpoints are really virtual devices that are because they're virtual, they don't go

10963.59 -> down there. High Availability is something the cloud doesn't go down. We're going to

10967.12 -> talk about two kinds of endpoints gateway endpoints and interface endpoints. Gateway

10973.96 -> endpoints provide high speed access to AWS services, like s3. And the way it works is

10983.08 -> it prints a route to the service and puts the route on the routing table and allows

10988.59 -> private access from, say object storage to your VPC and vice versa. When you create an

10995.189 -> endpoint for s3, what happens the prefix list and a VPC endpoint or created the prefix list

11001.569 -> will adhere to the naming convention of a PRI, and it's going to look like polyoma P

11007.42 -> O dash, and then you're gonna have a bunch of whatever it needs to be things coming after

11012.109 -> that like, extra, extra, extra, extra, extra, extra, etc placed in the routing table. And

11017.19 -> that way your routers and the VPC, the virtual routers will know if you want to reach us

11020.5 -> three, go that way.

11027.78 -> So let's talk a little bit about securing an endpoint, you're going to set up an endpoint

11033.01 -> policy that's going to limit resources that are available to the endpoint. Remember, if

11038.35 -> you don't have a route to it, you can't reach it. So only limit limit your routing information

11043.2 -> to the subnets that need it. Because if you don't have a route, you can't reach it. So

11046.83 -> that's a great way to start with some security right then in there quickly, cheaply and easily.

11053.569 -> Next, we'll discuss interface endpoints. Now interface endpoints are a way to connect to

11062.08 -> different AWS services or other organizations. So let's say EC to Systems Manager, Kinesis

11068.77 -> load balancers. Maybe you're a car manufacturer and you make a car but you got a tire manufacturer

11075.109 -> and a batting mat, battery manufacturer, and a steel manufacturer all on AWS, you want

11081.48 -> to connect to them directly across the AWS cloud. That's why you're using an interface

11087.75 -> endpoint. Interface endpoints work a little bit differently than the gateway endpoints.

11091.899 -> We describe what happens when you create an interface endpoint and effectively creates

11096.97 -> a network interface on your VPC that's local to your VPC and use that interface to connect

11105.689 -> to the third provider. AWS will automatically generated DNS names. You don't have to remember

11111.27 -> the IP address and you can connect to it via the name. Interface endpoints actually use

11117.72 -> the AWS private link service, which is a one way like pseudo wire or virtual wire to connect

11123.041 -> to things across the AWS network. It's like a virtual private line. And the private line

11128.859 -> creates these network interfaces that you use, and it restricts all traffic going across

11134.55 -> the endpoint between your VPC and that service with a customer partner. What does it look

11142.729 -> like architecturally speaking, let's say you've got your your VPC and you want to reach a

11150.53 -> service provider VPC and the service provider VPC is VPC to what happens you create a VPC

11158.55 -> endpoint and you can reach the system inside of EPC Joe

11163.92 -> I'm going to stop there. Because VPC peering is another concept and I think I've been speaking

11175.62 -> a long time and I don't want to confuse people. Chris, are there any questions for me? I'm

11182.931 -> sure there are.

11185.75 -> All right, let's see if there are questions

11195.59 -> I'm worried we covered so much and people are getting tired

11207 -> Oh, yeah, comments on cat.

11215.26 -> We love my cat Cindy. She came into our house like a storm. I bought her from my wife, because

11224.59 -> my wife loves cats. And the next thing I realize the cat sleeping with me following me from

11231.17 -> room to room special. How do you recognize an ipv4 versus ipv6 address? Leo for my team

11238.51 -> did a great thing. And ipv4 address is going to look something like what 10.0 dot 1.1 28.

11247.97 -> Whereas an ipv6 address is going to be in hexadecimal. And it's going to look like 2002.

11257.71 -> And it's going to look like zero delta echo seven, for example. And you'll see another

11261.88 -> call in it'll be Alpha Bravo, Charlie, Delta, colon. And then you'll see like a 001 thing,

11267.03 -> and it'll go on for 120 minutes. Great way to great example.

11281.88 -> Does endpoints use patch cables? Now these are virtual things. So between every server

11292.18 -> they're going to be plugged into switches, typically two that are going to be plugged

11295.13 -> into the routers. The routers are gonna have cables in between each router. But the endpoints

11302.88 -> are virtual, they're gonna use the IP network that's already been established. Good question

11306.529 -> that's.

11315.12 -> For the endpoint, they must have IP connectivity, everything is plugged into the network, which

11319.38 -> means you may need a network

11320.38 -> just to concern, you can access your s3 bucket from new Seatoun instance. And a private some

11333.83 -> was via a gateway endpoint. That is correct. You could also go out and reach it through

11337.33 -> the internet if you wanted to. But that wouldn't be efficient.

11343.36 -> Are there any implications using any IP? No, not at all. That's your only option. When

11352.439 -> it's recycled back? Well, the implications are, if you wanted to use a different one,

11358.01 -> you'd have to update your DNS mappings and things like that. Kind of like when you go

11361.96 -> from one service provider or another, but everything else would be okay. You just have

11364.45 -> to change your IP addresses and your routing. Well, you don't really have to do so much

11368.46 -> with routing on the cloud. It's much easier and your DNS Good question.

11374.6 -> Are there cases when you would prefer an interface? Gateway or an interface? Yeah, you have to

11382.84 -> use it based upon when you're using. If you're connecting external clients, it's an interface

11386.979 -> endpoint. If you're connecting to s3 or DynamoDB, I think it's s3 Definitely think DynamoDB

11393.109 -> also uses a gateway on point everything else uses an interface on point

11409.649 -> Okay, so I do want you to at least I want to make sure you read the book as well. You

11420.3 -> know, in the book, we have the ability in the time to provide even more content. And

11425.59 -> there's a relationship between the content that we're discussing here and in the book.

11429.1 -> In the book, we're giving you much more focused AWS content for the exams. And I'm giving

11435.09 -> you as much non AWS content as a candidate time permitted. But I'm still making this

11441.21 -> bootcamp focus. So I want you to read the I want you to read the book, I want you to

11445.3 -> practice the loves. And I also want you to watch all these because I want the best for

11452.04 -> you. To end points of playing ipv6, it's kind of a weird route that they're putting in the

11457.17 -> routing table. Remember that as a prefix less common xx xx, so it's something slightly different.

11464.729 -> Does AWS generate the DNS? Yes, absolutely. Free on points? Do you secure BGP routing

11476.359 -> information and prevent unauthorized changes to the routing table? Well, you can do certain

11484.81 -> things like you can set up MD five message authentication, which is not the greatest

11489.55 -> crisis in the world, you manually assign your BGP peers, which is something but what you

11496.051 -> typically want to do is you typically want to set up your BGP policy with either a distributed

11502.68 -> list, for example, or a route map, that you're only accessing routes that the come from the

11507.47 -> subnets that you're supposed to. And that way, you're generally there. There's some

11511.51 -> very good guidance on acceptable ways to use BGP. But there and I actually have an article

11517.34 -> that I wrote on for hacker noon on how to secure your BGP as well, and BGP hacking and

11522.75 -> hijacking. But kind of keep that in the back of your mind. Yes, those are the kind of main

11526.84 -> things that you can do, it's not a lot. But the message dot the message authentic the

11531.47 -> BGP peer authentication, the fact that you manually define the peers. And if I know that,

11537.21 -> I'm going to receive the 172, once one, six, dot 0.0, slash 19. I'm only going to take

11550.75 -> one route from that and not allow any other routes to be injected into my systems. And

11554.96 -> the cloud provider would do the same question. What is the AWS backbone, it's their high

11563.561 -> performance, high speed network.

11572.63 -> Strong questionnaire.

11581.02 -> Please download those. And most importantly, tomorrow. There's a lot of questions on careers.

11587.38 -> And there is a massive difference between certifications and getting hired in today's

11593.25 -> world. I want you to get hired, I want you to earn a lot more than you dreamed possible.

11600.05 -> And that's very easy to do if you know exactly what to learn. And most of those things are

11605.29 -> not in certifications. And all of you can do it, regardless of your background. So please

11609.439 -> join us tomorrow on the become the ultimate Cloud Architect webinar, not only will be

11615.5 -> present for about 30 minutes, but we will spend an additional 90 minutes answering your

11619.75 -> questions live live. to kind of keep that in the back of your mind. And we can do it

11629.82 -> face to face. And if you had a good time and you're learning please hit the like button,

11635.51 -> please subscribe to our Youtube channel and hit the notification bell so you'll be notified

11639.63 -> when we do these things. And, you know, I it costs as much as buying a new car to put

11645.84 -> on one of these productions. And we do it to help those that can afford training. Because

11650.979 -> I really want to help the entire world build their best career. Please share this, tell

11657.96 -> your friends to take this course we're keeping it live on YouTube completely free. So please

11662.88 -> send an email a tweet, make a LinkedIn post. We put a lot of time, effort and money on

11669.649 -> this. There's about seven members of my team working on this right now. Let alone the hundreds

11675.74 -> and hundreds of hours behind the scenes to put something like this together. Please share

11679.96 -> this so we can help as many people as possible. BGP will be have to be configured on both

11688.449 -> sides. BGP is only configured on routers, not switches unless it is a layer three switch,

11695.819 -> which is really a switch router combined. So thrilled that you're there. Sara, thank

11700.621 -> you so much

11711.17 -> I don't know what you mean by that. But anybody can follow this course I don't care whether

11714.6 -> you take the certification or not, I care that you get the knowledge. I've got people

11721.62 -> getting hardest caught architects every single day. And some of them never even took a certification

11727 -> exam, but they're trained for the job. And they don't need any fancy education. They

11732.68 -> don't need any experience. They needed to read the book and watch this course again

11737.12 -> and again, each time I promise you will pick up other things. CMS, thank you so much. I

11742.41 -> really owe me thank you so much. Lonzo another awesome way to live and go code bootcamp,

11750.359 -> I think so Alonzo, I hope so. I want to thank you so much.

11760.04 -> Chris, thank you. The Go Cloud Architect with him is thrilled to help David. We're thrilled

11766.77 -> to be here for you. Karen, thank you and team Mike and team. Thank you so much. Thank you

11772.479 -> all having a great time. Allows you to remember that song y'all having a good time. And so

11776.851 -> put in the chat box because it's driving me nuts. Not the pitbull version. But the original

11781.07 -> version. I like the purple version. Thanks, Jason. Thanks. Which webinars that for tomorrow,

11788.319 -> that is the how to get your first cloud architect job. And not only will tell you how to get

11793.26 -> your first cloud architect job, but how to be great at it so you can have a great career.

11798.07 -> Thanks, Lady Godiva. Thanks, Jim. Thank you. Great. Thank you. And we appreciate it so

11808 -> much. We love the cloud community. And I'm thrilled you're here. AJ, thanks so much.

11812.56 -> Please hit the Like button. Come back. We have to thank AJ for a service in the Marine

11816.41 -> Corps. He is a great guy. We're thrilled that you're here. blocked out learned a ton. I'm

11822.96 -> so happy that makes me happy you. And Tom. Well, thanks so much. We're thrilled you're

11828.46 -> here. Thanks, Kristen. We're happy to help. Thank you, Collins. Peter, thank you so much.

11839.109 -> And Emanuel. We're thrilled you're here. And some from my team. Thanks so much. Thank you,

11848.38 -> Samira, and thank you, Lady Godiva. Ajay, we're thrilled to see her. You feel certified

11854.279 -> already. Wonderful. Here. We that's our point. And our goal. You're more than welcome, Victor.

11864.92 -> Sure. You're so welcome.

11865.92 -> Omar, we're so happy to keep providing content. We love doing this. Great job. And back to

11873.171 -> you and your guests. Thank you. I really do love my cats. Actually, Chris has a beautiful

11876.689 -> cat to another educating still learning things with ease. I'm so thrilled to hear that. And

11881.82 -> Igor, we're so happy to help.

11893.74 -> As a current student like has helped me land a career and I'm I'm so thrilled that and

11897.81 -> your continual learning more, I'm thrilled to know that Dino? Thrilled That's our whole

11905.84 -> point. Change lives and get people higher. Would you suggest taking a Solution Architect

11912.75 -> first and the CCNA? I don't know what your goals are. Please join us on the how to get

11918.46 -> your first cloud job webinar tomorrow night. Because your goals determine what you should

11923.96 -> learn and how to do it. Sandeep, thank you. So Sanjay, thank you so much. Please join

11933.42 -> us in that webinar tomorrow. It will be a life changing event.

11939.75 -> So thank you all so much. I'll see you all tomorrow in class. And please make sure you

11948.61 -> join that webinar. It is really valuable information that I want to order. No. Have a wonderful,

11953.52 -> wonderful night.

Source: https://www.youtube.com/watch?v=vSpuS2HDgww