Service level agreements (SLAs) are one of the major considerations for every buyer of cloud computing services, but often are very hard to both define and enforce. In fact, in a recent blog<http://blogs.gartner.com/lydia_leong/2012/12/05/cloud-iaas-slas-can-be-meaningless/>, Lydia Leong of Gartner has noted that most public cloud SLAs are “practically useless,” fuelling the argument that enterprise-level applications and services cannot be delivered on a public cloud infrastructure.

Tony Munro, Solutions Executive: Dimension Data Africa says that most buyers agree that a strong SLA should include provisions around uptime, response time and latency guarantees. When it comes to uptime, the network and servers should be up and running more than 99% of the time. Response rates to any emergency incident should be less than 30 minutes, and latency should be less than one millisecond (1ms) for the transfer of data packets from one cloud server to another within the same cloud network. 

The problem lies in the fact that more often than not, cloud buyers often can’t draw a significant distinction between a 99.95%, a 99.99%, and a 100% SLA as it often depends on how the public cloud provider defines their SLA measures and penalties. Despite these challenges, SLAs are critical components of any cloud offering and are necessary to help organisations gain confidence in the performance of the cloud for business-critical applications. 

The criteria below provides cloud buyers with five practical guidelines for evaluating a cloud provider’s SLAs: 

  • Uptime Levels: What level is the provider committing to and what is necessary for ensuring the performance of your applications
  • Measurement: What should you be demanding from your provider and how often should they be measuring uptime vs. downtime – monthly, quarterly, or annually?
  • Coverage: What are the five key areas that the SLA should be covering, and which are the most critical?  What should your provider be accountable for? 
  • SLA Violation: What constitutes “failure” under the SLA and how do you ensure that you are adequately covered? 
  • Penalties: How are these calculated and how can your organisation ensure that the cloud provider’s definition of an SLA violation will result in a payout that addresses the impact of the outage or downtime?

 

 

 

 

The criteria above will help cloud buyers navigate through the challenging world of Service Level Agreements so that they are able to select a managed cloud solution that aligns with their expectations and business requirements. 

Leave a Reply