July 28th, 2021

Machine translation is used partially for this article. See the Japanese version for the original article.

Introduction

There are many types of clouds, but this time we will explain "Why are redundant configurations necessary" on public clouds using Amazon Web Services(hereinafter called “AWS”) and Microsoft Azure(hereinafter called “Azure”) as examples.

Cloud Services and SLAs

An SLA stands for Service Level Agreement and is an agreement on service levels between service providers (cloud vendors) and service users (users). Virtual machine cloud services (Amazon EC2 and Azure Virtual Machines) provided by AWS and Azure have well-defined SLAs and detailed virtual machine uptime.
Now, despite the clear definition of uptime percentage, we will explain "Why are redundant configurations necessary on cloud" using AWS and Azure as examples.

Necessity of Redundancy Seen from SLA for Amazon EC2

The SLA for Amazon EC2 can be found at:
https://aws.amazon.com/compute/sla/

The General Service Commitment section says, "AWS will use commercially reasonable efforts to make the Included Services each available for each AWS region with a Monthly Uptime Percentage of at least 99.99%, in each case during any monthly billing cycle.". The reason why Amazon EC2 is said to have an uptime percentage of 99.99% is from this expression.

However, what is noteworthy here is the definition of the monthly uptime percentage. "Monthly Uptime Percentage” is calculated by subtracting from 100% the percentage of minutes during the month in which any of the Included Services, as applicable, was in the state of Unavailability. Also, “Unavailability” mean "when all of your running instances (other than Single EC2 Instances) deployed in two or more Availability Zones in the same AWS region (or, if there is only one Availability Zone in the AWS region, that Availability Zone and an Availability Zone in another AWS region) concurrently have no external connectivity.". To put it plainly, Monthly Uptime Percentage is calculated only when multiple Availability Zones running instances become "Unavailability".

An Availability Zone is one or more discrete data centers in an AWS Region(*) and there are six Availability Zones in the N. Virginia region as of July 2021.
When using Amazon EC2 not for Single EC2 Instances, it is assumed that Amazon EC2 should be placed in these two Availability Zones, i.e., "Redundant configurations are necessary".
* https://aws.amazon.com/about-aws/global-infrastructure/regions_az/#Availability_Zones

Necessity of Redundancy Seen from SLA for Azure Virtual Machines

The SLA for Azure Virtual Machines can be found at:
https://azure.microsoft.com/en-us/support/legal/sla/virtual-machines/v1_9/

At the beginning, it says:
“For all Virtual Machines that have two or more instances deployed across two or more Availability Zones in the same Azure region, we guarantee you will have Virtual Machine Connectivity to at least one instance at least 99.99% of the time.”.
For Azure, the point is the concept of "Availability Zones" in if you have two or more instances deployed across two or more Availability Zones in the same Azure region.
An Availability Zone refers to a fault-isolated area within an Azure region, providing redundant power, cooling, and networking. By deploying Virtual Machines across two or more Availability Zones in the same region, Service Levels of Virtual Machines are guaranteed.

This means that "Redundant configurations are necessary" when using Azure Virtual Machines.

Conclusion

We explained "why are redundant configurations necessary" on clouds using AWS and Azure as examples mainly from the viewpoint of SLAs.
Both require "redundant configurations" as a prerequisite for uptime percentage of 99.99%.
In addition, the SLA explained this time is only a SLA for a cloud service. It should also be noted that the SLA of the entire system is the responsibility of the cloud users.
SLAs may be updated, so check the description of each site regularly.