||Add To My Personal Library
April 11, 2008
Vol.30 Issue 15|
Page(s) 25 in print issue
Cloud vs. Utility Computing
Know The Difference To Determine What’s Right For You
A few years ago, all the big vendors began to roll out their visions of pay-as-you-go computing. This led to a potpourri of labels, such as utility computing, grid computing, and the adaptive enterprise. The basic idea was to offer computing services in the same way as utilities. You pay for the amount of processing power, disk space, etc. that you actually consume.
Customers have been able to take advantage of pay-per-use IT services for several years, whether compute, networking, or storage, says Patrick Eitenbichler, director of marketing for StorageWorks at HP (www.hp.com). At the end of the month, the service provider looks at how much was used on average and then charges by the amount consumed.
Cloud & Utility Computing
But then came along another term called cloud computing. It sounded similar and, in fact, was similar in some respects. Just about everyone decided that it was another word for utility computing. Yet these concepts have distinct differences when you get down to the details.
Users and bloggers talking about cloud computing are referring to online services that abstract the user from the details of infrastructure, including physical computing resources, location, data partitioning, scaling, security, backup, etc., notes Bert Armijo, senior vice president of product management at 3Tera (www.3tera.com). In short, cloud computing users want to avoid infrastructure. The provider is in complete control of the infrastructure.
Cloud computing comes in a couple of flavors. Amazons S3 and SimpleDB (www.amazon
.com), Nirvanix (www.nirvanix.com), Mosso (www.mosso.com), and Salesforce.coms Force.com are examples of cloud computing for developers. Carbonite (www.carbonite
.com) and box.net are examples of cloud computing services.
Utility computing users, conversely, are seeking a service that allows them to deploy, manage, and scale online services using the providers resources and pay for resources they consume. However, they want to be in control of the geographic location of the infrastructure and what runs on each server.
So which is better for what? According to Armijo, cloud computing works well as a consumer service such as SaaS, Web hosting, email, and offsite PC backup. On the developer side, it lends itself to Web design, content delivery, and building applications where scale, performance, and selection of infrastructure software arent critical.
On the other hand, Armijo considers utility computing to be better when used to build scalable Web 2.0 services and search applications. It fits well where the developer needs to control geographic distribution, scalable compute resources, high bandwidth, and performance in order to manage the user experience. SMEs looking to offload some apps are also using utility computing because it provides them equivalent control over the apps as if the applications ran in their own computer rooms.
Cloud computing is great and easy to use when performance and the selection infrastructure software isnt critical, says Armijo. Utility computing users are control freaks; they like to know what they are running on [and] how it performs; they want to tweak their system and be in full control of the resources they are using.
HP offers a different perspective. Eitenbichler says utility computing refers to a business model, while cloud computing describes the underlying IT architecture.
In terms of cloud computing, the industry thinks of a large, virtualized pool of IT resources that can be flexibly deployed to whatever application currently needs them, he says. On April 15th, for instance, the online tax service providers will require tons of IT resources; on April 16th, those same resources can be deployed to other applications. And if the cloud is big enough, youll have the flexibility to achieve service levels with significantly lower investments and higher IT utilization.
Eitenbichler is not convinced there is a whole lot of difference between the two—at least from the customer perspective. Because cloud computing sounds hip right now, there are a number of service providers who use this term for what is in reality a utility, Eitenbichler notes. But its irrelevant whether HP uses cloud computing or another IT architecture to deliver these services.
Rather than cloud vs. utility, then, he sees the argument as being utility/cloud computing vs. purchasing your own IT assets. In the end, the decision may well come down to how confident you feel about the service provider.
by Drew Robb
The View From The Clouds |
One cloud vendor is Nirvanix (www.nirvanix.com). It has built a global cluster of storage nodes referred to as SDN (Storage Delivery Network). SDN creates a global namespace transcending all the nodes. This enables Internet-enabled applications to scale instantly to meet demands for hosting and delivering large quantities of media files. Nirvanix SDN is delivered as a Web service. It operates as an alternative to services such as Amazon S3 (www.amazon.com).
A backup application called ElasticDrive (www.elasticdrive.com) is one example of a third-party product that takes advantage of this cloud-based service. It provides backup services to customers and is integrated over the cloud to SDN.
In the future, companies will forego the building, management, and complexities involved in building their own infrastructure and focus on developing their businesses, says Jonathan Buckley, chief marketing officer at Nirvanix. The box is dead will be the moniker of 2008 as more services move to the cloud.