The Web is atwitter with discussion of the technical hobgoblins that are bedeviling Amazon’s S3 Web storage platform today. S3, which a lot of significant Web-based services rely on to provide the disk space that they need to store stuff, has been glitchy or altogether inoperative for at least the last six hours. There’s some good information and perspective over at SmugMug’s official blog–SmugMug being one of the services that uses S3 and has therefore been wonky today.
I happen to be working on a review of Web-based image editors for PC World. and found that a couple of them weren’t working today. One, the excellent Picnik, says that S3 is to blame. (I have a hunch that the other one also leverages S3, but I’m not positive.) The outages are a hassle–hey, I have a deadline to meet–but they’re also a sobering and useful reminder that Web-based services are extremely dependent on a lot of complicated technology and infrastructure working properly. Picnik may be problematic today, but my copy of Adobe Photoshop Elements, an old-fashioned piece of desktop software, is working just fine.
Not that desktop software is invulnerable: If my motherboard or hard drive croaked right now (and they could!) I’d be denied access to Elements until I solved things myself. All things being equal, it should make sense to throw technical challenges like keeping software working at a big company like Amazon or Google or Yahoo or Microsoft.
But it’s fascinating and sobering to see that Amazon–a company with oodles of resources, armies of techies, and, one hopes, a sophisticated game plan for keeping its services chugging along even when things go wrong–can fall victim to technical gremlins like this. The Amazon Web Services site refers to the Amazon platform as “robust,” and it is, mostly…but “robust” is not a synonym for “failsafe.”) With S3 as popular as it is, an awful lot of customers of an awful lot of services were inconvenienced today, even if they didn’t know that S3 was to blame. (In many cases, they didn’t: The error messages I got at Picnik didn’t mention S3, and I only learned they related to it when I checked out the Picnik blog,)
S3 suffered another major outage back in February, so this current one isn’t unprecendented in the least. I wonder if any of the companies that use S3 will reconsider the proposition. And I hope that Amazon explains exactly what happened and what steps it’s taking to prevent it from happening again.
(Wacky postscript: I uploaded an image of the Amazon Web Services logo to use with this post, and couldn’t figure out why it wasn’t displaying. Yup–WordPress.com uses S3, too.If you see an empty rectangle at the top of this post, S3 is still sickly; if you see the Amazon Web Services logo, it’s a good sign that it’s feeling healthier. I’ll say it again: When Amazon S3 goes boom, so does the Web…)
Back in March, Google 
In my last