IBM gets down to business with Grids

Grid Iron

ComputerWire: IT Industry Intelligence

IBM Corp yesterday announced ten new grid computing initiatives that are the culmination of six months of marketing and partnering work after the establishment of a cross-divisional organization last June that handles all development, alliances, marketing, strategy, and sales for grid technologies within IBM,

writes Timothy Prickett Morgan


The initiatives target potential customers in precise niche markets where grid technology is being embraced by early adopters. For each niche, IBM has put together a complete offering, including open source software technologies, its hardware and middleware, and products from dominant third party grid computing experts who have been in this market longer than Big Blue.

IBM, ever keen on making money by fostering new markets that will generate a big services business, has spent the last six months trying to get grid computing from its academic and research organization nursery out into the larger commercial data processing and collaboration market because it is here that grid computing can help IBM push its hardware, software, and services, and if it is lucky, make a few billion dollars in sales in the next couple of years.

As part of the initiatives, IBM has signed master partner agreements with Platform Computing, the Toronto-based vendor of platform-agnostic grid middleware that is one of the early pioneers of the idea, and DataSynapse, a New York-based grid specialist in the financial services market.

IBM has also announced alliances with Avaki, Entropia, and United Devices, who provide middleware in the grid market. IBM's grid offerings also lean heavily on the open source Globus toolkit for building grids and the Linux environment. IBM is pushing grids based on its xSeries servers running Linux and its pSeries servers running AIX.

White Space

Grid computing seeks to lash machines together -whether they are within a single facility or scattered around the world - in such a way that massive sharing of data and processing capacity is possible. Grid computing also seeks to make use of unused processing cycles on desktops and workstations, which Dan Powers, vice president of grid strategy at IBM, calls "white space." This is a good term for this excess capacity, and it is the first time we've heard it referred to in this manner.

IBM is keen on stimulating this new grid market, even though the next-generation of the Globus toolkit, version 3.0, is now only in the alpha phase. Globus Toolkit 3.0 will be the first platform to implement the so-called Open Services Grid Architecture, which marries the grid architecture to the idea of Web services.

Tom Hawk, general manager of the IBM grid organization, was adamant that grid computing is not a future technology, that it is here today, and in the market niches that IBM has chosen to target, this seems to be the case. But for general businesses, these initiatives, as well as those put together by IBM's competitors to chase the grid market, are just the test beds for what grid computing might be.

Discount brokerage house Charles Schwab, which is a big IBM mainframe and Unix server shop, was rolled out as an example customer for the new initiatives.

Schwab has set up a grid that lashes together xSeries servers running Linux and a version of the AIX LoadLeveler grid scheduling program that IBM Research has ported to Linux (but which is not yet commercially available) and a task scheduling program that IBM Research calls Tags that rides on top of Globus.

This grid has allowed a financial services application that runs on this cluster of machines in the company's San Francisco data center to do its work more efficiently and therefore cut down on the processing time for this application from four minutes to 15 seconds. While shaving transaction times down by a few minutes may not be a big deal in Peoria, in the financial services world, this can be the difference between making some money and making a lot of money.

Grid Unlock

For 2003, IBM is focusing on five different markets with its grid initiatives.

  • IBM is offering an enterprise optimization bundle and a business analytics bundle and for the financial services market. The enterprise optimization bundle is akin to what Schwab has implemented, which seeks to improve the flexibility and efficiency of the existing IT infrastructure by connecting machines and applications through grid technology. The analytics grid aims to pump up the accuracy and speed of the statistics that drive financial transactions so companies can make decisions faster; it also aims to make IT systems more resilient, which is problematic since financial systems have long since been distributed across many distant and incompatible machines. This is a perfect environment for grid computing.
  • IBM is offering a similar analytics grid offering for the life sciences, and an information accessibility grid that also seeks to make the exchange and querying of data in non-standard formats more efficient.
  • A similar information access grid offering is being packaged up for government customers.
  • IBM has targeted two different grid bundles at the automotive and aerospace industries, one each for engineering and design collaboration.

In addition, IBM has also established a grid implementation workshop for companies who want to check out this new technology, and has set up grid implementation centers in Montpelier, France, Nakahari, Japan, Poughkeepsie, New York, and Santa Teresa, California.

IBM is bringing all of these different technologies together, says Powers, because "customers want commercial products with support since they cannot got to the open source community to get support." IBM wants to be the throat to choke in grid deals, which is how IBM Global Services makes its money a lot of the time.

© ComputerWire

Biting the hand that feeds IT © 1998–2021