Global Infrastructure & Cloud Integrations Flashcards
What is global application?
A Application that is deployed in multiple geographies}
in AWS: Regions and / or Edge Locations
Why is global deployment important?
Decreased Latency - shorter way to the server
Desaster Recovery - failover to a different region
Attack Protection - distributed global infrastructure is harder to attack
How is AWS Global Infrastructure built?
There are regions with multiple AZ and there are edge locations
What is AWS Route 53?
Managed DNS (Domain Name System)
What are common records in AWS?
www.google.com -> 12.34.56.78 == A record (IPv4)
www.google.com -> 2001:0dbb8:…:7334 == AAAA record (IPv6)
search.google.com -> www.google.com == CNAME (hostname to hostname)
What are the Route 53 Routing Policies?
- Simple Routing Policy
- no health checks just hostname to IPv4
- Weighted Routing Policy
- add weights to different EC2 Instances that will correspond to amount of traffic being forwarded to that instance
- Latency Routing Policy
- latency will be used to connect users to the closest server possible
- Failover Routing Policy
- health check on primary -> if failed -> send to the failover server
What is AWS CloudFront?
It’s a content delivery Network.
It allows to store content on more then 215 Points of Presence.
It offers DDoS protection (because worldwide).
First time user requests some content it will be loaded into the regional edge cache from let’s say Australia, and should another user request it again then it wont need to ask australia, instead the local Point of presence will deliver as it has already loaded it.
Where can CloudFront get data from?
- S3 Bucket (can also be used to upload data to S3 [an ingress])
- Custom Origin HTTP (any http backend)
If the result is available in local cache then it will deliver if not it will get it from the origin server and cache it for the future requests
Can cloud front offer access to private S3 Bucket Objects
Yes, but for that we need to create origin access control (OAC) on the S3 Bucket but it gets provided by AWSW.
What is S3 Transfer Acceleration?
It is the usage of Edge Location to enter the AWS private network at an earlier stage and then continue transfer on the private fast network instead of the public one.
What is AWS Global Accelerator?
It is an action of leveraging the AWS private network to optimize and therefore speed up the route to our application for around 60% improvement.
Example: People trying to connect to our app in india but coming from Europe will enter through their closest Edge location and then continue on Private AWS network.
What are AWS Outposts?
Allows business to run hybrid clouds by offering to setup outpost racks on prem and come preloaded with AWS Services.
This however moves the responsibility for physical security to the company.
What services work with AWS Outposts?
Amazon EC2, EBS, S3, EKS, ECS, RDS and EMR
What is AWS Wavelength?
Allows to deploy AWS services (EC2, EBS & VPC) to the edge of 5G Networks of e.g. telecom allowing to reduce Latency for mobile devices to absolute minimum.
The Wavelength zones connections often never reach the cloud itself, however if needed they are interconnected (e.g to access DB).
No extra charges are set for using Wavelength
What are AWS Local Zones?
Are extension of AZs. E.g in the us-west 1 we have the 6 AZs but we also have multiple localzones like Dallas or Houston allowing us to be even closer to the users and minimize the latency.
What global architecture types does AWS offer?
- Single Region, Single AZ
- Single Region, Multi AZ
- Multi Region, Active-Passive - Active (read/writes), Passive (data replication, but only reads). Better global read latency, but writes stay the same
- Multi Region, Active-Active, Both reads and writes have low latency. E.g DynamoDB global tables :)
What 2 types of Communication can happen over cloud?
1) Synchronous - Application to Application
2) Async / event-based - Application to Queue to Application
Why can synchronous communication be bad? What’s the alternative?
When encountering a spike of traffic.
Using SQS queue model or
using SNS pub / sub model
What is AWS SQS?
Simple Queue Service.
Whatever producers put into the queue the consumers will pull it out and after processing delete it from it.
USED TO DECOUPLE APPLICATIONS
How long are the SQS Messages being stored?
Per default 4 days max 14 days.
How to decouple application tiers with AWS SQS?
Put the SQS between Webservers and the Processing instances.
The the SQS can scale the no of EC Instances in Processing with ASG
What is Amazon Kinesis Data Stream?
real-time big data streaming at any scale.
What is AWS SNS?
Simple Notification Service (Observer Pattern)
It has so called topics that are in the end different chats that can be subscribed to.
Publisher
Subscribers
What is Amazon MQ?
Managed message broker service for
1) RabbitMQ
2) ActiveMQ
It doesn’t scale nor integrate as SNS / SQS.
Used as a legacy service for users migrating to cloud but that don’t want to reimplement that parts in their applications.
