chriswebb on TwitterChris Webb on FacebookSubscribe to the RSS FeedLinkedIn

Partly Cloudy: Amazon’s S3 Service Goes Down

King CloudWe’ve been hearing for some time that the future of our data storage is “in the cloud.” Many of us already use cloud storage for email (GMail), files (XDrive) and backup (Mozy) among many others. But what happens when you can no longer reach your data?

Several companies found out today when Amazon.com’s S3 service experienced an outage and left applications without their data tier. I first experienced it this morning about 8:30 EST when those of us on Twitter noticed that avatars were not loading. Later reports of missing images on Facebook, and other applications started rolling in as startups and application developers realized their apps were broken due to the Amazon S3 outage.

So, what to do when you rely on the cloud? Simple - don’t rely on the cloud completely.

SmugMug’s Don MacAskil has it right, and discussed his approach on TechCrunch’s report of the outage:

We do rely on S3 for our primary storage, but we do maintain our own “hot cache” of data in our datacenters, too, which is less than 10% of our total storage. Our customers weren’t affected by this morning’s outage.

PBwiki’s Nathan Schmidt agrees:

Never build your architecture to require low-latency, high-availability access to S3 or its competitors, because you won’t get those - that’s not what it’s for, that’s not what it’s optimized for, and you’re never going to be able to peel back those layers of abstraction and long-haul network.

(Photo credit akakumo)

  • Andrew
    Nice picture)
  • Ade,

    Thanks for commenting.

    I would not want to speak for Don, but you raise some good questions. Since I don't know the ins and outs of his infrastructure, I cant answer for him.

    Don's an avid blogger over at SmugMug, so you might want to ask him over there, or perhaps he will comment here. Don, you want to chime in here?
  • Ade
    I'm not sure I completely understand SmugMug's response from a technical perspective. You have to make a decision about which 10% you're going to store locally, and I'm going to assume that it's at least files that were used more recently. But what happens when someone pulls up a photo that was uploaded years ago and it's not part of that 10%?

    Or even if they store high-res photos on S3 and store thumbnails and previews locally, if someone requests a print of that photo, they'll need to wait until S3 is back up to fetch the high-res files.

    Logic just seems to dictate that if you're storing content there that isn't being stored anywhere else, then either you don't need that content in the first place, or you're going to be affected in some way if it goes down.
blog comments powered by Disqus