Cloud Zone is brought to you in partnership with:

I am the API Evangelist. Not in the sense that I’m evangelizing a single API to you--In the sense that APIs are important for everyone to be aware of. I’m paying attention to not just the technical, but the business and politics of the web API movement. I share my insights by blogging on the business of APIs at apievangelist.com, politics of APIs at apivoice.com and you can find more information about me at kinlane.com. Kin is a DZone MVB and is not an employee of DZone and has posted 84 posts at DZone. You can read more from them at their website. View Full User Profile

Amazon Should Pull Plug Regularly to Show Which Companies Do Not Have It Together

04.22.2011
| 9934 views |
  • submit to reddit

I am by no means an IT architecture specialist, although I have a lot of experience in this area.

I have deployed infrastructure to support large conferences in single locations and across multiple global locations simultaneously.

I have done this for companies including Google and SAP.

I deployed this infrastructure using Amazon Web Services (EC2, S3, CloudFront, EBS, etc.)

So I have an understanding of what it takes to deploy on the Amazon Cloud, and knowledge of what it takes to scale and be redundant.

I was just listening to someone from a leading geo data provider talk about how they were unaffected by todays AWS outage in the eastern region, because they scale across multiple regions.

I heard people say they stopped using AWS because their EC2 instance failed, or EBS Volume failed….then I ask….describe your redundancy?  Nothing…

Clay Loveless of Mashery has wrote an excellent post on what it takes to operate in the clouds.   Most IT folks either don’t have the expertise to see it at this level, or won’t invest the money it takes to deliver at this level.  Says something about Mashery.

I really think that AWS should randomly pull the plug just as a fire drill, to show which technology providers built in the cloud, don’t either have the expertise, or are not willing to pay for the infrastructure it takes to stay up and running in the clouds.

References
Published at DZone with permission of Kin Lane, author and DZone MVB. (source)

(Note: Opinions expressed in this article and its replies are the opinions of their respective authors and not those of DZone, Inc.)

Comments

Michael Parmeley replied on Fri, 2011/04/22 - 9:09am

I think car companies should design their cars so the front tire has a catastrophic blowout every once in a while so we can see which people really know how to drive. Ridiculous. If you have limited resources (i.e. money) sometimes you just have to assume your infrastructure is sound and your provider is competent.

Jason Whaley replied on Fri, 2011/04/22 - 11:14am in response to: Michael Parmeley

I really wish I could upvote your comment. The real solution here is to educate users how to perform destructive testing of their systems. For AWS this really means writing some tutorials and guides on how to do this with the EC2 API. For instance... starting up a test environment for my system and see how it behaves when... I don't know, say all EBS volumes being used are detached and/or delete volumes. Or if you depend on something like SQS, see what happens to your app under reasonable load when SQS isnt available any longer (which you can simulate by changing /etc/hosts for sqs.us-east-1.amazonaws.com or whatever endpoint you use for SQS).

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.