What do those Availability 9s mean anyways?

Anup MarwadiAWS, Cloud Architecture

Have you ever been responsible for designing highly scalable web applications or SaaS Platforms that cannot have any downtime? Or have you, like us have been involved in DR (Disaster Recovery Planning) for Cloud Platforms? Well, we help Enterprises both large and small with their Cloud Architecture Strategies and it is absolutely important to know a few things when it comes to availability.

For starters, you know that you don’t get to throw the availability 9s liberally. What do I mean by that?

Well, all SaaS platforms have SLAs (Service-Level Agreements) that define the operational constraints. An important constraint is the uptime/downtime aka Availability SLA.

Most organizations throw around 99.999% uptime like it is a game. SPOILER ALERT – You’ll need millions of dollars and some really insane developers, devops architects, noc operations to get there.

Even the addition of one extra 9 changes strategies by hundreds of thousands of dollars, and in some cases to the tune of millions of dollars.

So be wise my friend. Be sure you’re aware of what you promise.

Here’s a quick breakdown of how crazy the addition of a single 9 can be.

AvailabilityMaximum Disruption (per year)Max Disruption (per month)Max Disruption (per day)
99%3 days 15 hours7.31 hours14.4 minutes
99.9%8 hours 45 minutes43.83 minutes1.44 minutes
99.95%4 hours 22 minutes21.92 minutes43.2 seconds
99.99%52 minutes4.38 minutes8.64 seconds
99.999%5 minutes26.3 seconds864 milliseconds