Yahoo! SVP Shelton Shugar, speaking at the 4th International Cloud Computing Expo, outlined some of the detail involved in serving 600 million users per month while handling hundreds of petabytes of data.

"We are building a private cloud to be used within Yahoo!," he said. Hundreds of applications have already been migrated to the cloud, he said, noting that "we are (now) focusing on the data processing environment and serving environment."

"This is a long-term activity. started a few years ago," Shugar said. "And prior to cloud we might have had to literally forklift machines and move them when we want to move capacity. With cloud serving we are building a flat network and can basically point and click and move capacity from one place to another."

Shugar outlined five areas of emphasis:

Edge services - where the company deploys around the edges of the netwrok to speed content delivery.

Cloud serving - involves hosting applications. automated provisioning, and "we allow through a declarative language to define an app through tiers of an architecture," Shugar said. "Tiers can be as big or small as you need and can adjust over time. We can use the language to define connectivity, and can clone a collection of tiers."

Storage - All of the online servers have RESTful APIs, "so are easy to use," Shugar said. "We can support a variety of consistency models, and there's a lot of similarity between and among our services." He also mentioned the company's "MOBstore" (mass object store) capability, which takes objects ranging in size from a megabyte to a couple of gigabytes, which are assumed to be immutable. He also noted the use of the company's "Sherpa" funcitonality," so I dont' have legions of developers trying their own schemas, etc. They write a few API calls and it's done," he said. "It's a great way to drive innovation."

Batch processing - This is the data warehouse, and where Hadoop comes into play. "We started developing this a few years ago, and ran on about 20 machines," Shugar said. "Today we run on tens of thousands of machines, with 80 petabytes of storage. We have upwards of 1000 people trained on it, and hundreds of applications on it. It let's us pull together thousands of computers and say hey 'you're one file.'"

Shugar also described "a new language called PIG (not an acryonm) that's a higher-level procedural language. The amount of code you write is very small, while adding metadata, workflow, data management, etc."

Data collection - "This is a set of services that allow us to collect content from our web servers, platforms, and search," Shugar said.. "It cleans it up all the data, pulls out out the bot traffic, etc."

Shugar went on to provide some use cases. One example was the home page. As Shugar noted, "when you type in yahoo.com you are going through our cloud. We look at user behavior through Hadoop and determine which are the most interesting articles for folks, a feedback loop to improve the consumer experience by selecting content the consumer is interested in." Through this method, "clickthroughs are up," Shugar said. "The consumers win because they get what they want, we win because we get more advertising."

Cloud Computing Expo 2010 East to Attract More Than 5,000 Delegates from 48 Countries
SYS-CON Events announced today that the 5th International Cloud Computing Conference & Expo will take place April 19-20-21, 2010, at the Jacob Javits Convention Center in New York City.

The International Cloud Computing Conference & Expo series is the world's leading Cloud-focused event and is held three times a year, in New York, Silicon Valley and in Europe. Over 200 corporate sponsors and 10,000 industry professionals have participated in Cloud Computing Expo since its inception, more than all other Cloud-related events put together.

The three-day event will offer a rich array of sessions led by exceptional speakers about the business and technical value of cloud computing with more than 80 sponsors and exhibitors on the 70,000 sq. ft. show floor and over 5,000 estimated delegates from well over 48 different countries.

4th International Cloud Computing Conference & Expo
At the 4th International Cloud Computing Conference & Expo, November 2-4, 2009, being held in the Santa Clara Convention Center, Santa Clara, CA, more than 2,000 delegates will find out how cloud computing is transforming the way that enterprises everywhere build and deploy applications.

Now held three times a year -- in New York, Prague, and Santa Clara -- the Cloud Computing Conference & Expo series is the fastest-growing Enterprise IT event in the past 10 years, devoted to every aspect of delivering massively scalable enterprise IT as a service. The event is co-located with our 7th International Virtualization Conference & Expo.

4th International Cloud Computing Conference & Expo Keynotes
Cloud Computing Expo 2009 West keynotes, November 2-3, 2009, in the Santa Clara Convention Center, Santa Clara, CA, will be delivered by Richard Marcello, Unisys President of Systems & Technology; Shelton Shugar, Yahoo! SVP Cloud Computing; and Richard Sarwal, Oracle VP of Development for Enterprise Manager.

Delegate Registration Options
Cloud Computing Expo (www.CloudComputingExpo.com) delegates can pre-register for the conference online with $800 savings here: https://www3.sys-con.com/cloud1109/registernew.cfm

Sponsorship Opportunities
Sponsorship and Exhibit opportunities can be obtained from Carmen Gonzalez by email at events (at) sys-con.com, by phone 201 802 3021.

SYS-CON's International Cloud Computing Conference & Expo, held each year in California, New York and Prague is the leading event covering the fast-emerging Cloud Computing market for Enterprise IT professionals. Co-located with the International Virtualization Conference & Expo, the combined event delivers the #1 i-technology educational and networking opportunity of the year for those seeking to establish a market lead anywhere in the multiple layers of the Cloud Computing ecosystem.

About SYS-CON Media & Events
SYS-CON Media (www.sys-con.com) has since 1994 been connecting technology companies and customers 24x7x365 through a comprehensive content stream at www.SYS-CON.com -- focused on 35 main subject areas, from Cloud Computing to Web Security -- interwoven with market leading full-scale conferences produced by SYS-CON Events. The company's internationally recognized brands include among others Cloud Computing Expo (www.CloudComputingExpo.com), Virtualization Conference & Expo (www.VirtualizationConference.com), Government IT Conference & Expo (www.GovITExpo.com), Cloud Computing Bootcamp (www.CloudComputingBootcamp.com), Cloud Expo Europe (www.CloudExpo-Europe.com) and now UlitzerLive! New Media Conference & Expo (www.UlitzerLive.com).

SOURCE: Ulitzer, Inc.

Contacts:

Delegate & Attendee Inquiries
Corinna Melcon
Corinna (at) sys-con.com
http://corinnamelcon.ulitzer.com
201 802 3019

Conference Faculty Inquiries
Jeremy Geelan
Jeremy (at) sys-con.com
http://jeremygeelan.ulitzer.com
201 802 3026

Sponsorship and Exhibit Opportunities
Carmen Gonzalez
Carmen (at) sys-con.com
http://carmengonzalez.ulitzer.com
201 802 2021

Expo Floor Sales
Megan Mussa
Megan (at) sys-con.com
http://meganmussa.ulitzer.com
201 802 3020

Website and General Inquiries
Krisandra Russo
Krisandra (at) sys-con.com
http://krisandrarusso.ulitzer.com
201 802 3020