What Amazon's Outage Means For Cloud Storage

NEW YORK -- Major websites including Foursquare and Reddit crashed or suffered slowdowns Thursday after technical problems rattled Amazon.com's widely used Web servers, frustrating millions of people who couldn't access their favorite sites.

Though better known for selling books, DVDs and other consumer goods, Amazon also rents out space on huge computer servers that run many websites and other online services.

The problems began at an Amazon data center near Dulles Airport outside Washington and persisted into the afternoon. The failures were widespread, but they varied in severity.

HootSuite, which lets users monitor Twitter and other social networks more easily, was down completely, as was questions-and-answers site Quora.

The location-sharing social network Foursquare experienced glitches, while the news-sharing site Reddit was in "emergency read-only mode."

Many other companies that use Amazon Web Services, like Netflix Inc. and Zynga Inc., which runs Facebook games, appeared to be unscathed. Amazon has at least one other major data center that stayed up, in California.

It's not uncommon for Internet services to become inaccessible due to technical problems, sometimes for hours or even days. But Thursday's outages were notable because Amazon's servers are so commonly used, meaning many sites went down at once.

Amazon did not respond to requests for comment. It has not revealed how many companies use its Web services or how many were affected by the outage.

No one knew for sure how many people were inconvenienced, but the services affected are used by millions.

Amazon Web Services provide "cloud" or utility-style computing in which customers pay only for the computing power and storage they need, on remote computers.

Seattle-based Amazon has big plans for AWS. Although it now makes up just a few percent of the company's revenue, CEO Jeff Bezos said last year that it could eventually be as large as Amazon's retail business. Competitors include Rackspace Hosting Inc. and Microsoft Corp.'s Azure platform.

Some people consider cloud computing more reliable than conventional hosting services in which a small company might rent a handful of computers in a data center.

If one of them malfunctions, the failure can take down a website. But "clouds" like AWS use vast banks of computers. If one fails, the tasks that it performs, such as running a website or a game, can immediately be taken over by others.

When a company needs more capacity, maybe because of a surge in visitors to its website, it only takes minutes to rent more computers from Amazon.

But cloud computing isn't immune to failure, either.

Lydia Leong, an analyst for the tech research firm Gartner, said that judging by details posted on Amazon's AWS status page, a network connection failed Thursday morning, triggering an automatic recovery mechanism that then also failed.

Amazon's computers are divided into groups that are supposed to be independent of each other. If one group fails, others should stay up. And customers are encouraged to spread the computers they rent over several groups to ensure reliable service. But Thursday's problem took out many groups simultaneously.

Outages with Amazon's services are rare but not unprecedented. In 2008, several companies lost access to their own files for about two hours when one of Amazon's data centers failed. The companies included DigitalChalk Inc., which delivers multimedia training over the Web.

In general, Amazon Web Services have been more reliable and, above all, cheaper than many other hosting systems, said Josh Cochrane, vice president of product development at Palo Alto Software in Eugene, Ore.

But the firm's websites and Web-based applications that create business plans were all brought down by Thursday's crash.

"It's a pretty vulnerable feeling," he said. "This is a really big message to us that we need to revisit our strategy."

That might include spreading the applications more widely over Amazon's network, so that problems at one data center won't bring down everything, he said.

Amazon engineers struggled throughout the day to rectify the problem. Leong said the problems are of a type that's not covered by Amazon's money-back guarantees.

Comments

[bkerensa]

02:52 AM on 04/25/2011

Linode for the win

[flamenco]

10:52 AM on 04/23/2011

I feel sorry for those who had the outage.

This may be partially positive in that it may force some improvements in cloud hosting, which is so over-hyped that I think it needed to come down a few pegs.

[HK79]

03:43 PM on 04/22/2011

Whats wrong with a dedicated server nowdays i still use one.. works fine for me.

[poster1122]

02:53 AM on 04/23/2011

The big reason is cost of ownership. Most web businesses aren't IT related. A small site that sells flowers, shoes, nicknacks, etc doesn't have (or want) IT as a core competency. As such, they want to focus their energy on their business and not the implications of using this database or that, what class o servers that will support their growth for the next 2 years, etc. They want to be focusing on getting suppliers for the flowers, next season's fashion trends, that sort of thing. They want to treat IT infrastructure like a utility like power or water; they want it, but they don't want/can't afford to absorb the cost of building that infrastructure.

Also, having your own dedicated server doesn't provide you a guarantee of better uptime. Drives crash, storms knock out power, T1 lines go down, etc. The difference is that in a private server scenario, you need to provide the staff to handle all these things. That can be a pretty big barrier to doing business, especially when you're first starting out.

[flamenco]

10:57 AM on 04/23/2011

Very valid points. Yet, cloud hosting can be just as expensive as dedicated - it varies a lot. Either way, you're usually on the hook for needing some serious server admin skills, unless you pay more for a user-friendly control panel and more support. You can shop around and find that either way could work.

For sites that don't get a ton of traffic, shared hosting on a decent host is very reasonably priced, and may work fine. The key phrase is "decent host".

Poster is right about this, though - no hosting is 100% crashproof.

[BlueDog1]

02:43 PM on 04/22/2011

Well you can always call the Amazon 800 number in India and they will provide vender management for you....................................I think in the old days we called this a Lemon suck.

[Keith Alban]

01:55 PM on 04/22/2011

April 21, 2011: the day the machines take over. (http://terminator.wikia.com/wiki/2011/04/21) Playstation Network down for days to come, Amazon, 4SQ, Reddit.... my laptop died... Hmnnnn.

[LightShadow62]

01:06 PM on 04/22/2011

Cloud storage; just another line item on your monthly bill.

[Stephen-in-Seattle]

11:13 AM on 04/22/2011

"revisit out strategy".... Don't put all your eggs in one basket. Don't put your web sites on the same data center provider.

[jt1980]

03:20 PM on 04/23/2011

The unfortunate thing is if you start having redundant cloud providers, then your administrative cost is just as high as co-locating and managing the servers yourself.

[macrocosm]

10:56 AM on 04/22/2011

Well none of my sites went down. In fact in over two years of use ive not had a single minute of down time, not that a little would be all that unimaginable. Amazon is the best hosting service I've ever had the pleasure of using, I'm sure they will have no trouble bouncing back from these small issues.

[BlueDog1]

02:42 PM on 04/22/2011

I hope Amazon gives a left turn signal before turning if not you could have your head torn off.

[oneyippie]

10:47 AM on 04/22/2011

You can expect more incidents like this, affecting even more websites and company data if you use the "cloud". Cloud computing takes your data away from your control. You have to be nuts to let that to happen. Plus once in the cloud you lose privacy as gov't agencies are allowed access to your data thru the cloud. Think not? Think again... It's why it was invented, like Facebook and Google.

[Stephen-in-Seattle]

11:15 AM on 04/22/2011

Cloud computing also creates a monthly revenue flow, which maturing companies like Microsoft need.

[Angelus2007]

12:48 PM on 04/22/2011

I generally agree with you, but MS doesn't need the montly revenue. Windows will still be the standard for a business OS for years (if not decades) to come. Due to security issues, businesses will still use Windows on PCs even if people at stop. Which will still create huge revenue as there are upgrades and new licenses bought.

[hypnotoad72]

04:37 PM on 04/22/2011

100% correct.

And just wait for cloud providers to shut down in "the new normal".

Leasing space and hoping the cloud provider fully fulfills their promise of uptime (good luck, most profit-driven entities do as little work as possible for the most amount of profit) is a pipe dream.

Hoping they keep things secure is another issue.

As is intellectual property ownership; some cloud entities, in their terms of service, will give themselves a helping hand to anything a user uploads.  But "the service is free"... (not really, marketing, ads, and using what people upload to promote themselves or sell to others is the form of payment.  Like bartering.)

And, on top of everything you and I have said, people still think it's of big benefit to move the government into the cloud:

http://www.zdnet.com/blog/service-oriented/clouds-consolidation-culture-making-federal-it-less-horrible/6921?tag=content;search-results-rivers

:facepalm:

Government might be in a mess, but relying on the cloud, ran by a for-profit enterprise, is just :facepalm: x100.

[Yorksgal]

12:08 AM on 04/23/2011

Of topic, but whenever I see your posts with Picard - I can't help reading your posts in English, English. So it always gives me a smile. Thank you.

[lensman3]

10:28 AM on 04/22/2011

Wikileaks was originally stored on Amazon's cloud. Pressure from the U.S. government cause Amazon to kick them off. Pressure from the U.S. Government also caused the US credit card companies to not transfer monies to Wikileaks and the DNS company in California kicked them off. Wikileaks had to make other arrangements.

Because of being kicked off of "cloud" servers, I would suggest to businesses and those individuals wanting to store info in the "cloud" hold off and let the "cloud" companies prove themselves as reliable and not fold to pressure from DOJ.

[Comeplayinmyreality]

10:08 AM on 04/22/2011

LOL, i guess no one else got that this was according to the SCC, that Skynet was going online on the April 19th at 8:11 pm and became self awared on the April 21st. just a little bit of movie nerd trivia.

[baronmerlot]

10:40 AM on 04/22/2011

Lol...thanks, I had forgotten that one.