Intro to AWS - The Most Important Services To Learn

learning aws can be pretty intimidating

there's hundreds of services and it can

be difficult to understand which one is

the right tool for what you're trying to

accomplish this video is going to help

you navigate the complex world of aws

services we're going to approach this by

examining a standard three-tier

application architecture like you can

see here so let's take a look at this

architecture now before peeling back the

layers and revealing the aws services

behind the scenes so what do we have

here in this application architecture we

have a pretty standard three-tier

architecture here with a web backend

layer we have our application layer here

which is kind of if you're in an

organization with a service-oriented

architecture this will be pretty

familiar this is where all your backend

services will be and then you have your

third tier here which is your database

layer this includes uh can be relational

could be a nosql database you can also

have some caching in there as well and

then we have a bunch of other components

related to this so we have deployment

orchestration to hold on to our source

code and then facilitate deployment we

also have a monitoring component here

for monitoring the state of the system

uh we have a load balancer here with a

dns pointing to that load balancer to

distribute traffic and then we have some

other toolkits here on the right for

event coordination say for instance um

this application did some kind of

something like google where you have

search query submission and like every

time that gets saved to a database you

want to trigger some kind of event in

this layer over here we have storage for

that event we have some analytical

processing a data warehouse and finally

some dashboarding for end users to

examine that content we also have some

toolkits for rapid development to deploy

a little bit faster as well and over

here finally on the left hand side these

two elements so for user pools to define

different users that are going to be

interacting with your application and

then of course to serve cache content so

this is our pretty standard application

here let's take a look now at some of

the aws services behind the scenes that

replace each of these different elements

so the first one is for dns so what is

the service that we want to use for dns

and aws

so for aws that's going to be amazon

route 53 and row 53 is the service where

you define all your dns configuration

including you know where you're going to

route traffic to from the internet also

supports other things like you know

health checks on your endpoints and any

traffic shaping that you want to do

that's going to all be done in your

rel53 service so it's great to be

familiar with rel53 you're probably

going to be using it all the time if you

were defining kind of externally facing

apis or endpoints

now from there your dns will typically

point to an endpoint for a load balancer

so for aws there's a couple options for

load balancers but the kind of top level

category here

is something called the elastic load

balancer service and there's two

variations for elastic load balancer

there's what's called the application

load balancer which operates at the l7

layer and that's more applicable for

those you that want to use content from

your http headers to route your traffic

and for those of you that are looking

for something a little bit more lower

level aws also offers a network load

balancer which operates at the l4 level

which is a little bit

more cost effective and supports higher

throughput limits as well so that's an

elastic load balancer and some of the

different options that you may want to

use now in terms of your web backend

layer over here this is going to be the

first kind of

time that a request from the front end

touches the back end now with compute

there's a couple different options that

you can decide to leverage and those

compute options it doesn't really matter

if it's your web backend layer or your

application layer the compute options

that you select for either of these two

things will be the same so what are the

options that are available to us so the

first option is amazon ec2 ec2 stands

for elastic compute cloud very very old

service and the basic gist of this

service is that you rent

virtual machines that you pay for by the

hour and they're really convenient

because with ec2 machines you can set

them up to do whatever you please you

can use them to host

back-end databases if you want you can

host a wordpress blog on it you can

create and deploy your application for a

rest api so it's a really flexible

service that allows you to do whatever

you want but some of the drawbacks with

ec2 can include just the all the setup

and configuration that you're going to

have to go through to use it so some

folks shy away from ec2 in favor of

something a little bit more hands-off

and for those of you that are interested

in serverless that's where our next

point comes in and that's aws lambda so

aws lambda is a serverless compute

infrastructure option and what that

basically means is that you define what

are called functions and these functions

are just snippets of code they can be

small or large pieces of code and aws

does not require you to have to worry

about any of the infrastructure as you

had to do with ec2 you just basically

write and deploy your functions and aws

worries about deploying your application

onto a container and then scaling that

application whenever the number of

requests to your lambda function

increase

so this is a really really attractive

model because

lambda is paper invocation so it's

really cost effective for application

workloads that have bursty traffic

patterns

or for applications that have traffic

during the day and then it kind of

recedes down to nothing in the evening

so lambda is a really really popular

service it's getting more and more

popular by the day and it's definitely

one to consider if you want to learn

more about aws lambda you should check

out my brand new udemy course in the top

right of this video now if you're not

into kind of deploying your

infrastructure onto machines directly

and you're not into using this

serverless model there is a third option

for those you that are more kind of

docker folks or docker fans

and that third option is amazon ecs so

ecs stands for elastic container service

and there's also a variation of ecs

called eks or elastic kubernetes service

and essentially ecs is just a service to

help you manage your containers helps

you set up servers with integrated load

balancing and auto scaling helps you

facilitate your deployments to those

containers so it's kind of like

something in the middle between ec2 and

lambda and if you're interested in

learning more about like these three

services and different compute options

that are available to you i have a video

where i compare these three things i'll

put that in the comments or the

description section below

so like i was kind of saying uh the

application layer doesn't really change

in terms of the compute options you have

the same kind of fundamental building

blocks that are available to you

regardless of if it's your web backend

layer or your kind of business logic

heavy heavy application layer here now

another service that helps kind of

facilitate the creation of your rest

apis and hosting those apis is one

that's called api gateway over here an

api gateway is a super super powerful

service because it offers additional

functionality on top of just kind of

using a load balancer to distribute your

traffic to different nodes here and the

types of uh kind of features that api

gateway offers are things like api

throttling or authorization on an api

say for example you you're building kind

of a private api that you only want to

be accessible from users in a user pool

which we're going to talk about next

year

you can set up your api gateway to

validate that you know a token is valid

uh by integrating with the user pool

service or you can define your own also

offers other features such as model

validation so you can define what types

of models your api supports and then

have that validation performed before

the request actually gets to your

backend layer here

so you can do like a bunch of different

combinations here you can do your dns

pointing to your api gateway which

points to your load balancer which

points to your infrastructure layer and

that is if you want to take advantage of

some of those features that i just

described so speaking of user pools we

just kind of touched on that we might as

well reveal the service here and that's

one called amazon cognito so amazon

cognito is kind of a very very powerful

but underrated service and what it

allows you to do is to create user pools

and these user pools kind of similar to

what you'd have on any kind of login and

registration website you know you create

a user they provide a login a username a

phone number a recovery option all that

kind of stuff and so with cognito you

can

have users sign up for accounts directly

within cognito using the hosted ui

or you can integrate with other

third-party identity providers such as

google facebook amazon so anytime you've

seen like login with amazon or log in

with google or facebook or any other

identity provider

that could be integrated with cognito as

well so very very useful for

applications that require user

registration and if you combine that

with api gateway you can do things like

ensuring that a user is part of a

certain user group before the request

can be validated and that flows through

to your backend layer here so that's a

little bit about cognito let's kind of

finish this three-tier architecture

discussion and talk a little bit about

the database layers now or the storage

slash persistence layer

so a lot of applications have caching

enabled on them just to increase

performance on some

lookups of items that are quite common

or maybe you just want increased

performance so what's the aws service

that allows that to happen so that

service is called elastic cache and

elastic cache kind of comes in two

different flavors you can either go with

the flavor that

is memcache based or you can choose

redis now redis is probably the most

popular one people when people think

about caches i think redis comes to mind

almost immediately but regardless of

what you choose when using elastic cache

it is a caching service so it's going to

be based on key value lookups and you

are going to have to worry about hosting

that infrastructure essentially you kind

of own a cluster of nodes and

these are memory optimized nodes where

they have plenty of memory to facilitate

your application's needs um but you

still have to worry about the

maintenance of that cluster you know and

node replacement hardware failures

things like that it's a relatively

hands-off service but there are some

nuances that you may need to know about

in terms of maintenance and alarming and

all that now in terms of what should we

store our database in should we store it

in a relational database or a nosql

database so there's a bunch of different

options here that you can choose from so

the one that aws likes to push a lot is

a relatively popular one called amazon

aurora now amazon aurora is a in-house

built amazon database that is compatible

with both my sequel and postgres i

believe postgres is still in preview

mode now so it's a fully managed rds

database that kind of makes your life

easier in terms of worrying about things

like administration monitoring

auto scaling storage auto scaling and

compute auto scaling also offers

something called the data api that you

can use to

call your rds database using a rest api

as opposed to a traditional kind of

database connection so there's a lot of

features that are coming with aurora and

you may really want to consider it if

you're thinking about using a relational

database now a relatively similar

service to that is amazon rds or

relational database service and where

these two are different is that rds is a

database service that allows you to

select which database configuration that

you would like so you get to pick from

common database uh platforms such as

mysql postgres microsoft sql server

oracle cassandra and probably a couple

other ones that i'm missing as well so

rds is probably what i would think most

people are familiar with but a lot of

customers are deciding to go with aurora

just because it's more of a hands-off

option and just makes life easier one

other thing about aurora is that it does

offer a serverless compute model as well

where you don't need to worry about

provisioning any type of hardware behind

the scenes as you would with an rds

database you can use the serverless

model and it's kind of like an auto

scaling type of database thing similar

to what lambda did for ec2 aurora server

list is doing for rds so it's it

provisions your infrastructure whenever

the request rate requires it so if you

have a bursty workload it'll add more

nodes and scale you up so that your

database can handle more volume

so that's a little bit for rds here

let's talk about nosql database options

now now definitely the most popular

nosql database that you're going to hear

about a lot is dynamodb now dynamodb is

a nosql database that is optimized for

key value lookups it is a fully managed

database service so that means that you

don't worry about anything with regards

to infrastructure or hardware all you

really worry about is your scaling

configuration and dynamodb handles the

auto scaling for you behind the scenes

and so it's a really really powerful

nosql database and it's kind of used as

the building block for much of the

internet really if you take a look at

one of the aws white papers where

dynamodb unfortunately went down one day

it brought down with it a large part of

the internet including common

services that we all love like netflix

and other websites as well

so dynamodb is a super super powerful

and popular nosql database on aws and

really this is this service is kind of

at the heart of many other aws services

as well like behind the scenes

so yeah that's it for dynamodb now if

you're a mongodb type of person and

you're coming to aws you don't want to

learn dynamodb

there is a service option for you so aws

also offers what's called documentdb

like you see here and documentdb similar

to dynamo is a fully managed service but

this time it is compatible with mongodb

so that you can you know use mongodb as

you normally would in a fully managed

way so that should satisfy any mongodb

lovers and if you're looking for

something that supports more flexible

querying at scale

you may want to consider a service

called open search open search is the

new name for the traditional elastic

search service i believe aws and

elasticsearch had a falling out so aws

kind of came up with their own service

here but really this is elasticsearch

behind the scenes now what open search

allows you to do over some of these

other nosql databases is perform queries

that are more kind of fuzzy in nature so

give me all the records with with value

equals x y and z and you know value two

equaling something else and value three

equaling something else also allows you

to do some really powerful grouping

features dynamic grouping as well

comes with cabana as well which is an

open source dashboarding technology to

take a look at your data inside your

open search database

this has been used quite effectively as

a replacement to rds in some cases but

it's a really neat service that you

should consider as well now i wanted to

pause here to talk about some other

services that don't really fit the molds

here but kind of play a role in terms of

packaged infrastructure because so far

what i've talked to you about here these

are all kind of lower level building

block services but aws does offer these

kind of packaged infrastructure services

that make your life easier as a

developer you can tell that if you want

to build this three-tier application

architecture here there's a lot of

moving pieces here there's a different

compute option there's load balancing

there's api gateways there's databases

there's a lot of stuff going on so aws

does have some services that kind of

offer a combination of these different

elements as a single product and so

instead of having to worry about you

know each of the building blocks and

deciding on your own which one you want

to use you can use these pre-packaged

infrastructure services that bundle this

functionality together

and often they abstract some of the

complexity away from you at the

sacrifice of kind of configuration and

control so some of those different

services are

well the first one that i have for you

here is elastic bean stock now elastic

beanstalk is a pretty old service it

allows you to set up any kind of web

application could be a containerized web

application as well and it just makes it

easier for you to set up your app with

all these different components so a

backend layer load balancing also lets

you set up auto scaling and monitoring

so it comes with a lot of the components

here but you manage it in one spot

which is the elastic bean stock console

so you're still controlling the

infrastructure with elastic bean stock

but it's kind of like an orchestrator

service it'll go out to all these

different services here and provision

what it needs for the the type of

application that you're trying to deploy

onto it

now another service that just came out

pretty recently that does something

similar is one that's called app runner

and app runner behind the scenes uses

ecs and something called fargate which

is a kind of a serverless mode for

running containers where it uses

provision containers that you can

specify but you don't need to worry

about the infrastructure

so that's what app runner relies on so

with elastic bean stock you know it's

orchestrating the deployment of your ec2

machines your load balancers and any

other stuff that you may need however

you still have visibility or insight

into that infrastructure you still need

to worry about maintaining it with app

runner it's a little bit different all

of the lower level components are

abstracted away from you you just kind

of worry about your application

configuration and deployment an app

runner will worry about deploying that

onto your infrastructure and scaling it

if it requires it so that's another

important service to know as well and

also there's another option here which

is amazon light sale and light sale is

what i actually use for my personal aws

blog you can check it out at be a better

dev.com

and it's another one of these

pre-packaged services that just makes

your life easier so it's similar to

other uh kind of cloud vendors like i

want to say godaddy or digitalocean

where you kind of select the type of

stack that you want to set up for your

application and there's a bunch of

different pricing models kind of

pre-packaged pricing models where you

don't have to worry about the details of

kind of which node type is right for you

as you would have to do with your ec2

machines over here you just pick

different pre-packaged options for

compute and the costs are reasonable as

well and in terms of what you can deploy

on lightsail i use it to deploy my

wordpress blog but you can also use it

to deploy a lamp stack a mean stack a

ruby application you can also use it to

deploy your own containers

and you can also add other components as

well such as load balancing and auto

scaling uh so there's a lot of different

features that are built into light sale

but it's a much more simplified and

paired down version so you can add all

this extra stuff but you're doing it

within this kind of i want to say safe

safe zone or safe version of the aws

console so things are much more

streamlined in light sale you have very

few options of types of things that you

can do and it's a very beginner friendly

option for those of you that are just

getting started not recommended for kind

of production grade applications but

something great for smaller applications

or even a wordpress blog such as in my

case

now another honorable mention that i did

want to talk about briefly for those of

you that love graphql maybe you're a

front-end guy you love graphql aws does

offer a fully managed graphql as a

service

service called appsync and appsync just

makes it easier for you to develop your

graphql applications by providing you

with that graphql functionality

so you can use it to integrate with

other backend aws services such as

dynamodb you can use it to integrate

with lambda functions if you want to

have some custom resolvers and it can

also scale really really well too

completely transparently to you as an

administrator depending on the level of

traffic that's hitting your application

so appsync is another popular one that

you can think about if you're a graphql

user

now one other thing to mention as well

is in terms of cached content so for

many of these web applications you're

going to be serving different types of

cached content whether that be image

files your javascript your html your css

anything that you may want to cache and

basically put close to the end user so

you can get better performance

and for that you're going to be using a

service called amazon cloudfront

so cloudfront allows you to deploy a

cloudfront distribution

so you can have your application source

deployed in for example north america

but what if you have customers that are

located in europe or asia or australia

if you don't use cloudfront then any

customer is going to have to hit that

north america server which you know does

take some time so there's going to be

some performance degradation with

cloudfront you can set up and deploy

distributions that replicate some of

your content from your general object

storage which we're going to talk about

a little bit later and then replicate

that content to regional nodes that are

located all across the world close to

your end users and what that allows you

to do is get some better performance for

much of this static content so great for

applications that want to optimize the

experience for the user

okay so so far we talked about quite a

few concepts talked about like routing

apis user pools load balancing compute

databases

packaged infrastructure caching

now i want to talk about some other

components in terms of like how do we

actually deploy and monitor these

applications so in terms of deployment

there's actually four different

smaller services i want to say that aws

provides and these kind of work hand in

hand for deployment pipeline so let's

peel these back one by one so the first

service at hand is code commits now what

code commit is basically for is for

storing your source code so you can

either store your source code directly

inside code commit as a service or you

can integrate code commit with third

party providers such as

github if you have a private github

repository really the option is up to

you now code commit on its own isn't

too impressive or too powerful but it's

the integration with some of these other

services that do make it powerful and

the next one is code build now code

build allows you to take your source