Cloud Computing Comes In For Some Nasty Weather
from the outages-weren't-in-the-forecast... dept
I've definitely been a fan of Amazon's web services offerings. By providing cheap, scalable and robust processing power and storage (and, as of recently, a simple database), Amazon has started to singlehandedly change the economics of doing certain things online. It's become quite common for startups to build their platform entirely on Amazon's systems -- and given the stability of Amazon itself, it always seemed like a good idea, especially following some high profile outages at popular data centers. However, this morning, that "robust" part got called into question as Amazon's S3 service had a major outage, taking down a variety of startups and services (and, apparently, plenty of images on various websites). While some are suggesting this shows the weakness of "cloud computing," it's not necessarily all that different than those earlier outages where you had a data center go down. Given this and the recent Blackberry outage, we're again learning that robustness isn't just nice to have -- for many services it's really a need to have, and we're not quite there yet.
Reader Comments (rss)
(Flattened / Threaded)
I am so glad... by Dennis Howlett on Feb 15th, 2008 @ 10:37am
Someone recently looked sniffy at me when I said we're using Rackspace instead of Amazon. Now who's going to look sniffy back?
(reply to this comment) (link to this comment)
Re: I am so glad... by Mike on Feb 15th, 2008 @ 10:41am
Someone recently looked sniffy at me when I said we're using Rackspace instead of Amazon. Now who's going to look sniffy back?
Heh. Well, didn't rackspace just have a major outage too?
http://valleywag.com/tech/followup/rackspace-outage-was-third-in-two-days-321909.php
Not sure you can say that one is necessarily better than the other, which I think is kind of the point. We still haven't figured out a way to really guarantee uptime.
(reply to this comment) (link to this comment)
outage by edy on Feb 15th, 2008 @ 10:53am
we've to be realistic in life. outage is part of technology.
there's no such thing in life as 100% up-time.
(reply to this comment) (link to this comment)
by Anonymous Coward on Feb 15th, 2008 @ 11:29am
Normally you can maintain a backup system, but if you system is built on Amazon what's your backup if amazon goes down?
(reply to this comment) (link to this comment)
God is 100% up time so there by Oh Yeah on Feb 15th, 2008 @ 11:30am
sorry
(reply to this comment) (link to this comment)
File under 'get a life' by Freedom on Feb 15th, 2008 @ 11:35am
Some ideas to consider....
- Not a matter of IF it will fail, but WHEN. When it does fail, chance are that it will be at the worse time.
- Even planes with two engines crash. So while redundancy sounds great, it doesn't mean that you won't have problems. In fact, sometimes the increased complexity of redundant systems will increase the chance of failure along with the time necessary to fix it (assuming specialized parts are needed).
- I'm probably being a bit naive here, but does this really matter? For instance, when we work with client that has a server, workstation or Internet outage and they can't work for 1/2 day or heaven forbid a full day, after they come to terms of their system being down it allows them to refocus and catch up on stuff that they had been putting off because of being a slave to the PC/info stream. Sometimes outages are really small opportunities to re-focus and take a step back to see the forest. (I know, I should work in the PR departments for Amazon and BlackBerry - LOL!).
Freedom
(reply to this comment) (link to this comment)
by Joel Coehoorn on Feb 15th, 2008 @ 11:52am
> "we're again learning that robustness isn't just nice to have -- for many services it's really a need to have"
I disagree. In fact, I go the complete opposite direction. This proves that robustness isn't as important as we thought after all. The businesses that went down this morning will still be here tomorrow. If any aren't, it's probably not the outage that killed them, or at most it was the straw that broke the camel's back. It turns out most businesses can survive the occasional outage. Sure, it cost them some money, but it's not the end of the world.
(reply to this comment) (link to this comment)
by Sara on Feb 15th, 2008 @ 12:23pm
I was familiar with cloud computing but never paid attention to it. Now this is the second time in two days, i've seen it come up.
the other was from a site called Metue that had some ideas on how it factors in to the microsoft yhoo deal (article is here: http://metue.com/02-14-2008/microsoft-yahoo-merger-cloud-computing-theory/)
Twice in two days , now I'm thinking I need to do some reading and pay a ton more attention to what the cloud stuff is. Time to scan the past articles here
(reply to this comment) (link to this comment)
by Anonymous Coward on Feb 15th, 2008 @ 12:47pm
A failure is inevitable because you can't predict the unknown unknown of possible incidents.
(reply to this comment) (link to this comment)
Add Your Comment