It happens
It is not fair to pick on Amazon who runs, by all account, a first rate operation for their cloud computing service. But, being the best also means any glitch is highly visible as was the case in Ireland and Virginia. I am reminded of the incident, about 10 years ago, where one of our site in the east coast was down because some gas work around the datacenter severed the connection physically. (And, what is it with East Coast locations that would go down?)
Thinking about SLA
In light the Cloud outage, it is worth reminding oneself the idea behind SLA, service level agreement. It is NOT mainly a technical document that specifies how the hardware are used. It is NOT mainly a legal document that aims to put fear of god into the operators. It is FIRST and Foremost, a document that articulates your business requirement and how this (cloud computing) service can support your objects. What this also means is that you have to articulate a plan B that would support the business goals.
Then, you are ready to talk SLA and make intelligent trade-off's.
* Lightning strike zaps EC2 Ireland: http://www.zdnet.com/blog/saas/lightning-strike-zaps-ec2-ireland/1382
* Short-Lived Amazon Cloud Outage Takes Down Several Sites: http://allthingsd.com/20110808/short-lived-amazon-cloud-outage-takes-down-several-sites/?mod=mailchimp
* Were Amazon's Outages Inevitable? http://www.technologyreview.com/blog/editors/27072/?nlid=nldly&nld=2011-08-11
No comments:
Post a Comment