The Cloud Contract Adviser: Know your provider's infrastructure

Become An Insider

Sign up now and get FREE access to hundreds of Insider articles, guides, reviews, interviews, blogs, and other premium content. Learn more.

Strangely enough, last month's column about the importance of service-level agreements (SLA) was posted on April 21, the day of the big Amazon EC2 outage. While that outage certainly helped to illustrate the importance of having good SLAs, I swear that the timing was merely a coincidence.

Amazon's outage serves as a good reminder that what you really want is uninterrupted service, and SLAs alone aren't enough to get you there. Operational outages like Amazon's can definitely reduce service availability, so it's important to do what you can to understand and mitigate the likelihood of unplanned downtime in advance.

Amazon's own summary of how the outage happened makes it clear that it was not the result of external forces, which is something that folks often worry about with the cloud. Instead, it was a routine upgrade that was "executed incorrectly." The resulting problems were made worse by a ripple effect of interrelated automated systems and processes that weren't configured to anticipate or handle the initial error.

This type of outage highlights the question of whether or not the application of technology in the cloud has outpaced the ability to effectively manage it. So let's take a closer look at infrastructure operations management issues and some ways to contractually address them when adopting a cloud service.

The virtual nature of cloud computing makes it easy to forget that the service you get is dependent upon a physical . And the infrastructure behind the scenes of a public cloud computing service is a lot more complicated than a traditional data center.

In addition to general computing components such as virtual machine monitors, data storage and associated middleware, a public cloud infrastructure has to deal with things like workload management, data replication and recovery, and resource metering. And to make matters worse, all of these have to interact effectively, while they change over time as feature improvements and bug fixes are continuously rolled out.

To continue reading this article register now

5 power user tips for Microsoft OneNote
  
Shop Tech Products at Amazon