I love Amazon but AWS is clearly having growing pains
Posted by Margaret Dawson on August 25, 2011
I have been a fan of Amazon.com from its launch in the mid 90’s, when I was living in Asia and loved using Amazon to send gifts to friends and family back home or to treat myself to a book or CD I couldn’t find in Taipei. I also worked at Amazon and saw firsthand the incredible commitment to the customer and the innovative platform that has taken the company from a bookstore to the leading global eCommerce platform for the planet.
So, it is with this foundation that I am struggling with my love-hate relationship for Amazon Web Services. Why is it that we cannot even imagine Amazon.com the eCommerce site “going down” even in the height of the holiday crush, and yet many of us have experienced firsthand the pain of AWS not keeping its promise? And I say this having experienced my own Website being hit by Amazon outages and being surprised, frustrated and wondering where was the failover?
Gregg Lamm, with Puget Sound Business Journal, wrote a great overview recently, on the issues Amazon has been having with AWS, particularly with the recent crash of its data center running EC2 (Elastic Compute Cloud). This one fortunately only lasted a few hours, unlike the company’s disastrous cloud crash in April, when multiple eCommerce sites were down for days, losing potentially millions of dollars in sales. At that time, not only did the main cloud service go down, Amazon’s failover apparently didn’t work either. So, if companies had a separate redundancy plan, they fared better than those relying solely on Amazon.
As Charles Babcock wrote in Information Week: “this part of the incident illustrates that AWS has not so far safeguarded the services in one availability zone from mishaps in another. The details in the postmortem make it clear that what started out as a problem of electrical supply to one availability zone spread different problems to other zones, even though a customer’s use of a second availability zone to run a backup system is supposed to ensure high availability.”
What is going on here?
Many media articles were quick to scream “the sky is falling” on cloud computing and claim this is why companies should not put all their eggs in one basket (or all computing resources in the cloud). I agree that the cloud’s key value proposition is augmenting and extending existing infrastructure not replacing on-premise with the cloud lock, stock and barrel - but this is hardly a reason to give up on the cloud or start sounding the alarm bells.
And to be fair, there is some onus on the consumer of any cloud service to make sure you have architected your solution for success. Platform’s as a Service, like Azure, don’t magically enable you to create great applications – you need to have a strong design, architecture and code base. Same is true for IaaS. Are you taking shortcuts when moving infrastructure to the cloud rather than designing and implementing a fully redundant, defense-in-depth network architecture?
These events are also good reminders of the need for strong due diligence when selecting a cloud vendor. The biggest names in the industry are not necessarily the best suited to meet your needs. At the end of the day, any cloud solution is still running on somebody’s data center, so you need to understand that environment. Also, we are starting to see many cloud service layers, where a SaaS application or eCommerce app is running on a different company’s Infrastructure as a Service (IaaS). Make sure you understand the layers, SLAs and capabilities across a cloud service chain.
Outages by Amazon and others are not going to disappear, but you can help minimize the impact on your company. And I am sure Amazon is doing everything it can to avoid such mishaps in the future. The company has proven time and again its ability to both overcome obstacles and show great agility when needed.
Tags: Amazon Web Services, AWS, Cloud services, EC2, IaaS, Microsoft Azure
No Comments »
No comments yet.