What’s In A Name? Utility vs. Cloud vs Grid

In recent weeks there has been a rolling conversation on tech blogs about the definitions of utility computing and cloud computing, with contributions from Geva Perry at GigaOm, Simon Wardley, James Urquhart of Cassatt, James Governor at Redmonk and IBM’s Gerrit Huizenga. John Willis has sought to classify cloud computing providers.

And what about grids? The IEEE has published a classification scheme for grid systems (link via 3Tera’s Bert Armijo). Are all these definitions and classifications helping or just adding to the confusion about these terms? This week Derrick Harris of GRIDToday looks at the parsing of technology terms and how “grid computing” and “high performance computing” scored poorly in a recent Forrester report gauging enterprise interest in various technologies.

Computing At Scale: Parallel Programming

Bill McColl spent 20 years as head of research in parallel computing and scalable systems at Oxford University Computing Laboratory before leaving in 2006 to found Parallel Machines, a Silicon Valley startup developing scalable parallel systems. In recent weeks Bill has had a series of interesting posts on his blog, Computing At Scale:

• In a post on Parallel Programming Models, McColl looks at models for architecture-independent parallel programming that were developed in the 1990s, which are alive and well in 2008.
• In The End Of The Relational Era, McColl looks at recent developments in the database world. “Recent evidence suggests that we may indeed be reaching the end of the relational era, an era that has lasted for more than 30 years,” he writes.

BIll’s blog is new to me, and has gone through some quiet periods, but looks like a good resource if you’re involved in web scalability.

HP Launches Cloud Platform, Data Center Tools

HP today announced Data Center Transformation portfolio, a suite of services to help companies modernize and consolidate their existing data centers, or shift their operations to a cloud-based service hosted by HP.

The most intriguing new piece of HP’s strategy is the company’s entrance into cloud computing with HP Adaptive Infrastructure as a Service (AIaaS), which lets customers host applications in HP data centers optimized for Microsoft Exchange, SAP applications and other critical business apps. The HP platform joins a growing number of software as a service (SaaS) offerings, most notably Amazon’s utility hosting services and Salesforce.com.

“Within a matter of hours, customers can rapidly access additional computing power to meet their fluctuating needs,” HP said in its announcement. “With HP AIaaS, customers can realize improved service levels and convert traditional capital investment into an ongoing operating expense because all assets are owned and managed by HP.”

Big Spike for The Big Switch

Tim O’Reilly has noted a sudden surge in sales of The Big Switch, Nick Carr’s book about the emergence of utility computing and its impact on Internet life and enterprise IT. Tim writes:

We were curious about this huge jump. We see some recent activity in the blogosphere, but nothing jumps out at us as driving this spike.

Chris Anderson, the author of The Long Tail, wonders in the comments if the jump might be due to a single big buy for a conference. Interest in Nick’s writing have been helped by two recent posts about Microsoft’s web apps and data center strategies. But those didn’t appear until March 1-2. What explains the pop in sales starting the week of Feb. 24? In following the blog links, I came across a piece of interesting timing.

EMC’s Cloud Offering Taking Shape

As the Internet fills with computing clouds, which will be the largest? Much of the speculation has focused on the usual suspects - Google, Microsoft, Amazon - along with newer contenders like Facebook and Salesforce.com. With online storage likely to account for a significant chunk of the cloud, storage giant EMC has emerged as a wildcard. It’s been clear for months that EMC is assembling a cloud offering, with its acquisition of Mozy representing an initial beachhead.

Yesterday’s acquisition of a small startup called Pi has raised the visibility of EMC’s infrastructure ambitions. Pi is short for “personal information,” and is developing an online service to “enable individuals to control how they find, access, share and protect their increasing volumes of digital information.” Pi founder and CEO Paul Maritz, a Microsoft veteran, will become President and GM of EMC’s new Cloud Infrastructure and Services Division.

Chuck Hollis, Vice President of Technology Alliances at EMC, said that although EMC has bought many companies, the Pi deal “isn’t your everyday acquisition.”

Sun Preps Cloud Platform to Vie With Amazon

Sun Microsystems (JAVA) is preparing a new utility computing platform designed to compete directly with Amazon’s S3 and EC2 services. The new service is code-named Project Caroline, and will be formally unveiled at the JavaOne conference in May, according to The Register. Sun appears to be positioning the service to be more accessible to startups and small businesses than its current cloud offering at Network.com. Here’s Sun’s description of the project:

Project Caroline is a hosting platform for development and delivery of dynamically scalable Internet-based services. It is designed to serve an emerging market of small and medium sized software-as-a-service (SaaS) providers. Anticipating needs driven by new SaaS business models and processes, Project Caroline helps SaaS providers develop services rapidly using high-level programming languages like the Java(tm) programming language, Ruby, Python, and Perl, to update in-production services frequently, and to automatically flex their use of platform resources to match changing runtime demands.

Sun Distinguished Engineer Bob Schiefler is scheduled to make a presentation about Project Caroline at the JavaOne event, which is available online and provides developers with some details on how the service will work.

Joyent Backup Services Down for Three Days

Two online storage services operated by utility hosting provider Joyent have been offline for the last three days, apparently due to corruption problems with the ZFS file system. Strongspace and BingoDisk have been offline since Saturday night (Jan. 12). On Tuesday Joyent CEO David Young said the extended downtime was caused by complex corruption issues with ZFS, a new file system for pooled storage originally developed by Sun Microsystems for its Solaris 10 Operating System.

“We got bit by a massive ZFS bug,” Young wrote in an advisory to customers. “That’s the long and the short of it. The ZFS corruption got onto/into the backups. The good news is we can unravel the corruption. The bad news, given the fact that Strongspace and BingoDisk ran on a Thumper (aka SunFire X4500), was that we have to use other Thumpers to stage the uncoding of the ZFS mess. Moving so much data around to decode the ZFS corruption has taken time.” UPDATE: See our follow-up story for more. Joyent was using an older version of ZFS, and the bug in question was fixed nearly a year ago.

It will likely take more time to sort out the issues and recover user data, Young said.

Reliability in the Cloud: SLAs Will Matter

PRINCETON, N.J. - As Web-based “cloud computing” services become essential to Internet users and businesses, the operators of these services will need to improve their reliability and offer service-level agreements (SLAs) to define user expectations. That’s the prediction of experts at “Computing in the Cloud,” a workshop organized by the Center for Information Technology Policy at Princeton University.

Some cloud-based services could become so vital that they become candidates for government regulation, according to panelists at the event, which was organized by Princeton professor Ed Felten, a popular blogger with expertise in Internet security and intellectual property.

While we are in the early stages of cloud computing - services that run in a Web browser and store information in a provider’s data center - some of these services are already essential to their users. “People depend on these new technologies we’re creating,” said Jesse Robbins, a technologist and blogger at O’Reilly Radar. “When they go away, there’s the potential for people to be hurt.”

Jonathan Rochelle of Google (GOOG), the project manager for Google Docs and Spreadsheets, said the reliability of cloud-based services is improving and will improve more. “Capitalism is going to drive improvement,” said Rochelle. “These (improvements) will come because consumers demand it.”

The Role of SLAs

SLAs, which define a service provider’s responsibilities for performance and uptime, are a big part of the solution. “Right now there’s a lot of considered avoidance of SLAs,” by some cloud service operators, said Rochelle, who predicted this would change.

Amazon (AMZN), which is seen as an innovator in cloud computing with its Amazon Web Services offerings, only recently adopted an SLA for its S3 storage service promising 99.9 percent uptime. Amazon implemented the agreement after an API outage on its EC2 utility computing service wiped out customer application data.

In some cases, SLAs must evolve to address unanticipated scenarios in a Web 2.0 world in which collaboration is a central theme. “There is no obvious ownership model on some of this stuff,” said Rochelle, who cited an instance in which a document on Google Docs was jointly developed by 10 collaborators, only to have the account owner remove the document and close the account.