Efficiency has a new name: grid computing

Engineers are becoming more familiar with product lifecycle management (PLM) software and its promise of efficiency. But what about grid computing? When a tier 1 automotive supplier cuts its clash analysis time from three days to just four hours, engineers take note.

Both PLM and grid computing wave the flag of efficiency with pride. But, according to Anthony Stokes of business development, grid computing at IBM, "PLM manages a product from concept to retirement; grid computing is the front process that makes it more efficient, speeding it up and improving access." In other words, the two processes complement each other. PLM is cloaked in debates over the necessity and utility of a software program designed to coordinate everything from the design and simulation of a product to the manufacturing data and customer relations associated with it-basically rolling the project manager's job into a software program. Grid computing involves a group of servers, computers, and data storage virtualized as one large computing system. By drawing on unused power from other computers within the grid, companies can reportedly benefit from more power and speed, as well as seamless collaboration.

Austria-based MAGNA STEYR-a niche vehicle supplier of concept design, production, styling, and assembly-finished implementing IBM grid computing into its system early this year. By grid-enabling its PLM application software from Dassault Systemes and using Platform Computing as the grid middleware, the company immediately began seeing results of whether several CATIA designs would fit together once manufactured. "With grid, we're seeing an increase in automating those functions that free up engineers' productivity time," Stokes says.

While MAGNA STEYR currently uses 20 grid-enabled CATIA V5 workstations for clash analysis, the company plans to integrate 200 grid-enabled workstations, which could further reduce clash detection to just minutes. They also plan to grid-enable other engineering applications and use the data grid to link all involved parties so that they may collaborate in using the latest design data.

While other grids, such as the power grid, operate as a public grid, the compute grid is privately run as intra-grids within companies. As far as security is concerned, given the seamless accessibility, the grid operates under a company's current security policies, procedures, and systems. Also, the owner of a program on the grid may limit the access rights of other grid users, essentially personalizing the access privileges. In turn, the owner of a program/system that is enabling seamless access on the grid experiences no negative effects on his own system.

Grid computing is used by a number of industries today, including the automotive and aerospace industries for collaborating on design and testing. "We've seen interest from large companies with significant infrastructures, who have great opportunities for improvement," says Stokes. "We've also had interest from niche players, where the company is focused in one area and doing that one area of design very well."

For more information on IBM grid computing, go to http://www-1.ibm.com/grid