Select Region and the appropriate deployment model when using AWS

Tram Ho

1. AWS Region and Availability Zone

AWS Region is a separate geographic area that spreads all over the world with the idea that when there is a natural or man-made disaster in a certain region, there are still other areas that are guaranteed to work. interruption.

Having multiple regions also benefits speed in data retrieval or using services close to customers.

Availability Zone (AZ) is an independent resource set in each Region.

Each Region needs to have at least 2 AZs to ensure safety, eg in the case of a fire or explosion at 1 AZ should not affect the remaining AZs.

The AZs are linked together via a high speed transmission line. This helps you in designing a system that runs as only a centralized data center while still ensuring the safety of an AZ failure.

Refer to the details of AWS Region and Availability Zone at:

https://aws.amazon.com/about-aws/global-infrastructure/

There will be some information about:

  • Available Region and AZs and coming soon
  • AZ number specific in each Region
  • Time for each specific AZ to be officially put into operation

When selecting Region for your system, there are some points to note:

  • In most cases, the region closest to you should be chosen
  • Not all regions have full AWS services, some newly launched services will only be piloted first in a limited number of areas and will be widely developed in the future (Amazon Comprehend, Amazon Connect, etc.). ..)
  • The cost of the same service may be different if different regions

Responsibilities of both parties when providing and using AWS services

In general, factors such as data centers, hardware, software, connectivity will be guaranteed by AWS. Users will be responsible for customer data, configuration, application deployment.

For example, a car manufacturer, they will ensure that when you press the gas, the car will go faster, brake will slow down. How to use them is up to you.

2. Some deployment models

2.1 One Cloud Server

This is the simplest model, usually used in the development or minimal cost when starting a project.

Clearly, the drawback of this model is that it can only withstand one fault (Single Point of Failure).

Monolithic applications can also use this model, to ensure reliability when deployed on the cloud instead of using their own physical server.

2.2 Many cloud servers

This model is more secure because there are 2 cloud servers in 2 AZs. But there is still no guarantee in case the whole Region is in trouble.

In addition, this model also increases the flexibility of the system, allowing you to perform live updates by separating a server to perform updates and upgrades while ensuring uninterrupted service because there is still one Other cloud server works

2.3 Multi-Region model

This is the most complex model and of course also requires the highest cost. Amazon Route 53 will act as a DNS service.

In addition to ensuring the safety of the system compared to the two models above, this model also has performance benefits in case customers are distributed in many regions of the world, we can choose the appropriate Region for target customers to optimize the transmission speed.

Reference source:

https://app.pluralsight.com/library/courses/aws-network-design-getting-started

Share the news now

Source : Viblo