CNET también está disponible en español.

Ir a español

Don't show this again

Christmas Gift Guide
Tech Industry

Commentary: Getting to the bottom of grid

Firms are confused about what "grid" computing means, but that hasn't stopped them from rolling it out, Forrester says.

Commentary: Getting to the bottom of grid
By Forrester Research
Special to CNET News.com
May 24, 2004, 4:00PM PT

By Frank Gillett, principal analyst

Firms are confused about what "grid" computing means, but that hasn't stopped them from rolling it out.

Forrester recently surveyed 149 large companies in North America about their use of grid technologies: 37 percent reported that they're piloting, rolling out or have implemented some form of grid technology. Another 30 percent of firms are considering grid technology, so two-thirds of all firms are using or interested in grid technology. "Grid" has become a trendy term that no longer has one meaning.

Grid computing started some years ago as an academic effort to build supercomputing on the cheap by using many smaller processors coordinated to tackle numerically intense computer problems.


Get Up to Speed on...
Utility computing
Get the latest headlines and
company-specific news in our
expanded GUTS section.


This technical approach is only useful for computer problems that are parallelizable--problems that can be chopped into pieces for distributed processing--unlike most business problems, which require centralized processing.

But the term "grid" is now popping up all over, not because the classic grid approach has become more useful to enterprise information technology, but because "grid" is being used more broadly to apply to:

• A growing niche market. Big vendors like Hewlett-Packard and IBM offer software, solutions and services to tackle grid problems. Blade start-ups like Egenera and RLX Technologies market hardware for grids. And software start-ups like Avaki and Platform Computing have tapped into markets like engineering calculations and financial analysis.

• Data grids, which are designed to share data among systems and firms. Faced with the challenge of distributing data for processing to many computers, compute grid designers began to build technologies for sharing data widely within and among firms. Data grid technology is still early and is not as broadly used as grid computing.

• Shared use of IT resources, like clusters, organic IT, utility computing and more. Oracle included "grid" in the name of its latest database. "Grid" is often used in discussions about next-generation data center ideas like utility computing and Organic IT--and Sun Microsystems renamed its N1 initiative as "N1 Grid." Start-ups like Cassatt, DataSynapse and Paremus are extending grid concepts to new products designed for flexible support of Web services in a services-oriented architecture.

With the growing use of the term "grid" and the situations to which it applies, Forrester asked the IT and business leaders in its survey what they think grid is and whether they're using grid technologies. We offered executives a variety of choices for a definition of "grid" and for what applications of grid are of interest to them. We learned that:

• "Grid" means many things, and there is no consensus on a common meaning. Clustered computing and data grids were chosen as the most common definition, beating out parallel processing of numeric workloads by a wide margin, but no definition garnered more than 38 percent, even though we allowed multiple responses. Twenty percent reported that the term was confusing, and another 15 percent didn't know what the term meant at all.

• Data and infrastructure are top of mind for grid at more than 50 percent of firms. Executives told us that their largest interest in grid is to solve broad IT issues like distributed data and infrastructure software. Financial analysis and boosting numeric processing of conventional apps garnered major interest. Classic grid applications for scientific or engineering applications brought up the rear, with only 26 percent of firms showing interest or some implementation.

• Overall, 37 percent of firms are in some stage of piloting or implementation of grid. To understand the overall interest in grid, we took each firm's highest state of adoption for the six grid applications to get an aggregated view of grid adoption. Twenty-two percent of firms have implemented or have a rollout in progress--and 45 percent of firms are piloting or considering grid. This strong level of interest shows that firms are receptive and interested in products that offer to increase asset utilization, whether it is servers or data being shared.

• Firms that use grid technology use it for several applications. Firms that are at least piloting grid technology tend to use several applications of grid--the average number of grid applications at firms using the technology is 2.8. And almost half of firms at least piloting grid are using three or more different applications.

For vendors, the confusion over what "grid" really means complicates marketing and sales. With little prospect for clarity around these terms, vendors will get back to basic value propositions. They will focus on vertical solutions--not grid itself--to shorten sale cycles. Watch for vendors to start talking about dollars saved through deferred hardware purchases or the business value of faster results.

For customers, investigations of grid technology will lead to technology broadly applicable to the data center. They will avoid server purchases with server virtualization and slash job requisitions with management automation.

For virtualization, firms will turn to EMC's VMware on Intel and existing Unix virtualization solutions, plus the forthcoming Microsoft Virtual Server and Sun's Solaris N1 Grid Containers. For automation, chief information officers will consider products like BladeLogic, IBM Tivoli Intelligent ThinkDynamic Orchestrator and Opsware.

Ultimately, the term "grid" will fade from popular use as yet another naked technology looking for a problem to solve. But grid computing technology will survive and thrive as the enabling technology inside vertical solutions that solve real business problems.

The vision of data grids will become part of a greater vision of storage virtualization and information life cycle management. But we'll have to endure the broader use of the term grid to describe next-generation data center architectures for a while, until utility computing, on-demand or Forrester's Organic IT catches on as the name for a radical new IT architecture.

© 2004, Forrester Research, Inc. All rights reserved. Information is based on best available resources. Opinions reflect judgment at the time and are subject to change.