Report: Are AWS and HP cloud SLAs useless?

On December 10, a Network World article reported on a Gartner analyst Lydia Leong's blog post which suggested the public cloud services of Amazon Web Services and HP have the "worst SLA (service level agreement) of any major cloud provider." 

The Network World article, which was republished onAsia Cloud Forum's sister publication Computerworld Hong Kong, criticized AWS and HP's public cloud services on two fronts:

1) That AWS and HP impose strict guidelines which require users to architect their cloud systems to run on multiple availability zones (at least two AZ required for AWS, and at least three AZs required for HP). Not only would this incur additional costs for users, Leong suggested that HP's requirement of a third AZ will "further depart from people's comfort zones";

2) That the SLAs of AWS and HP were "unnecessarily complex" like "word salads," and were "limited in scope." For example, Network World reported that AWS's most recent outage impacted its Elastic Block Storage service, but it was not covered by its SLA.

A week later, in a separate blog post, titled "Some clarifications on HP's SLA", Leong clarified that, to be considered downtime, "HP's SLA is intended to cover a single-instance failure," while "AWS requires that all of your instances in at least two AZs be unavailable."

As for the wording of cloud SLAs, Leong emphasized that the "best SLAs are plain-language comprehensible," without requiring one to go through "a tangle of verbiage to figure out what they intend."
 

In response to Leong's criticisms, an AWS spokesperson said:

 

As for HP, though it did not provide any specific response to Asia Cloud Forum in relation to Leong's comments, it pointed us to a blog post by an HP senior product manager Blake Yeager dated December 11, which essentially clarified HP's cloud services SLAs.