Mountain View, CA
|reply to MxxCon |
Re: Another example why "The Cloud" is a bad idea.
MxxCon , I am in full agreement that the likelihood of a highly diversified cloud service (like AWS) going completely offline -- that is to say, EVERY geographic region going down -- is pretty unlikely.
However, there have been a few documented cases of entire AWS geographic regions going down:
* 2012/03/15 -- EC2 east region -- reference
* 2012/03/26 -- Amazon EC2 -- reference
I haven't seen anything on the outages mailing list (I'm subscribed) about AWS issues since then. I'd have to check NANOG as well to see if there were reports there too.
Making life hard for others since 1977.
I speak for myself and not my employer/affiliates of my employer.
Yes, a region can go down, but again, AWS offers people all the tools necessary to create an infra that can survive a region failure. If they did not implement and such an outage was unacceptable to them, its their own fault.
And I doubt that "a local accounting firm" would be using a bare-bones EC2 anyway. It would be either S3 or some seller.
If it's a seller that had 3 day long outage, then again, don't blame Amazon or "cloud" for actions of one incompetent company!
[Sig removed by Administrator: signature can not exceed 20GB]
|reply to koitsu |
Any engineer worth his salt who is hosting with AWS would set up redundancy in two, very geographically redundant, regions.
The fact several high profile sites didn't and went down along with a single AWS region just proves they should invest more money into infrastructure talent.
There is nothing wrong with "the cloud" as long as your provider is half-way competent.
Disclaimer: I am a SaaS cloud engineer. We don't host with AWS and we have geographic redundancy via BGP.