Building blocks for highly available systems

Andreas Wittig – 14 Sep 2015

Thanks to the cloud, highly available systems are becoming the new standard. Creating an architecture offering high availability or even fault tolerance is affordable nowadays as you can use ready-to-use building blocks. On top of that, you can operate a highly available infrastructure in the cloud with almost no extra costs.

AWS offers managed services with built-in high availability or even fault tolerance:

  • Load Balancer (ELB) is fault tolerant
  • SQL database (RDS) is highly available on demand
  • No-SQL database (DynamoDB) is fault tolerant
  • Object Store (S3) is fault tolerant

The virtual server (EC2) itself is not high available or fault tolerant as the underlying virtualization or hardware might fail. But tools are helping you to build a high available infrastructure based on virtual servers.

Building blocks for high availability

The following figure shows a typical architecture for web applications offering high availability. The building blocks are:

  • Load Balancer (ELB) fault tolerant by default routing requests to healthy virtual servers.
  • Running virtual servers in different data centers (Availability Zone) in the same region.
  • Running an SQL database which replicates all data from the master database to a standby database in another data center (Multi-AZ deployment).

Building a highly available system on AWS

All parts of this architecture are highly available. There is no single-point-of-failure left.

Andreas Wittig

Andreas Wittig

I’ve been building on AWS since 2012 together with my brother Michael. We are sharing our insights into all things AWS on cloudonaut and have written the book AWS in Action. Besides that, we’re currently working on bucketAV,HyperEnv for GitHub Actions, and marbot.

Here are the contact options for feedback and questions.