Cloud Hosting

It's no secret that Amazon's Elastic Cloud Compute (EC2) cluster is big. It powers massive services such as Netflix, Instagram, and Reddit. But just how big is anyone's guess. The retail giant won't say. So, dissatisfied with the company's silence, one man decided to find out.

Huan Liu, a researcher with Accenture Technology Labs, crunched the numbers, and according to his estimates, Amazon Web Services (AWS) boasts just under half a million servers445,000 to be exact.

That may still be a guess, but it's certainly an educated one. Mr. Liu used a process called DNS translation to map out the internal IP addresses of each active AWS instancemade possible, in short, because EC2 publishes all public IP addresses used.

As of March 12, 2012, Mr. Liu estimates the US East data center, based in Virginia, is the largest in size, with 5,030 server racks. Assuming that each rack boasts 64 blade serversan educated guess, as there's no way to know for surethe total would amount to 321,920.

Mr. Liu was also interested in mapping the growth of Amazon's cluster over the past six months, as an indication of the cloud computing industry's overall health. From August 23, 2011 to February 23, 2012, Mr. Liu observed that the US East data centerwhich suffered an outage last Apriladded roughly 110 server racks each month.

For comparisons sake, thats more than Amazons entire So Paulo centre in Brazilthe company's smallestwhich Mr. Liu estimates to have just 25 racks. However, while the growth rate looks roughly linear, Mr. Liu wrote on his blog, recently it is showing signs of slowing down.

Its important to realize that Mr. Lius results, while precise, are still based on assumptions; he admits it is possible that other AWS services, such as S3, SQS and SimpleDB, could run on dedicated racks, which would exempt them from his calculations. Also, if there was no active instance on a rack, it would not be counted in Mr. Lius results.

Read the original post:

New estimate pegs Amazon's cloud at nearly half a million servers

Everything works well in the cloud, until it doesn't.

Consider the Microsoft so-called "Leap year" bug that crippled that company's Azure cloud services last month. Bill Laing, vice president for Microsoft's server and cloud division, described the system failure in a blog post and said that Microsoft will overhaul its disaster recovery efforts, as well as other aspects of the business.

Fortunately, in this case, services were restored and the outage was (relatively) short-lived.

Nonetheless, it's a certain reminder of how things can -- and will -- go wrong in cloud services and that each organization is responsible for their own business continuity.

Also see "The cloud security survival guide"

Until recently, that was a concern of David Wellington, IT specialist at TAPS.org. Since 1994, TAPS (Tragedy Assistance Program for Survivors) has offered support to more than 35,000 surviving family members of fallen U.S. servicemen and women with casualty assistance officers and military chaplains. As TAPS moved increasingly to cloud-based services, it wanted to make sure it had reliable access to its data -- even if the cloud services provider went down.

About a year ago, TAPS began moving away from its on-premise productivity and office software to Google's Gmail and Google Apps. "We were relying on our own services and virtual private networks, but it was clunky and sometimes the connection simply didn't work. Switching to cloud services made sense and is easier for our people to use," says Wellington.

"We want all of our employees, wherever they are, whenever they need it, to be able to access the information they need. We also wanted to know that it's backed-up, so that we're always ready to help family members and serve those survivors," he says.

Having access to their data "no matter what" meant having backups that were not reliant on the cloud services provider itself being available. While it may seem strange, at first, backing up applications and data that are in the cloud -- where service providers are widely expected to take care of backups and security for their customers -- Wellington and TAPs aren't taking any chances.

To back up their cloud-based data, TAPS turned to startup Backupify, a provider of cloud-based data archiving, search and restore services for online services such as Google Apps, and social networks such as Facebook and Twitter. Late last month, Backupify also released its Snapshot for Salesforce, which saves a copy of one's Salesforce backup on Backupify servers and also provides the ability to download the Salesforce.com data for onsite backup.

More here:

Be Prepared For When the Cloud Really Fails

Amazon's beautiful Sterling, Virginia, data center. Photo: Eric Hunsaker/Flickr

Its one of Amazons best-kept secrets. How many computers does it take to keep its Elastic Compute Cloud platform afloat?

And now, a researcher with Accenture thinks he has the answer: 445,000. Thats the number that Huan Liu came up with when he did a bit of internet sleuthing. Its a fairly big site; its pretty impressive, he says of the entire EC2 operation.

EC2 is Amazons pay-as-you-go computing service. Its become a popular way to spin up computing power for a corporate skunkworks project or a startup, but its also the back-end for serious online sites, including Netflix and Dropbox.

Lius analysis found that Amazons main cluster of data centers, located in northern Virginia, is truly massive: he guesses that Virginia is home to about 322,000 servers. But he also found that Amazon has a relatively small footprint in other parts of the world. For example, he guesses that there are only 1,600 EC2 servers in Sao Paulo, Brazil. Its hard to compete with Amazon on scale in the US, but in other regions, the entry barrier is lower. For example, Sao Paulo has only 25 racks of servers, Liu wrote in a blog post discussing his findings.

Liu, a research manager with Accenture Technology Labs, took advantage of the way that Amazon organizes its EC2 domains to come up with his estimate, which strikes us here at Wired as a bit of a lowball guess.

Because Amazon relies heavily on virtual computing that is, it can host several software-based virtual servers on a each computer figuring out the number of machines in Amazons data center is a very tough task.

But Liu used a few tricks to link all of Amazons Domain Name System and IP addresses to actual server racks used by the Internet giant. Then, by guessing that each server rack has 64 machines in it, he came up with his total numbers.

He tells Wired that hes pretty confident about the number of racks that Amazon uses. As to whether the company crams 64 or 128 servers in each rack? Well that, nobody knows for sure. Its an educated guess, he admits.

The estimate also leaves out the servers that are powering Amazons Virtual Private Cloud, a hosting service for servers that are kept off the Internet, and which couldnt be measured using Lius techniques.

Go here to read the rest:

Amazon Cloud Powered by 'Almost 500,000 Servers'