This story's headline has been updated to more accurately reflect IBM's business model.
NEW YORK, April 3 – Calling its more well-known work in the area of supercomputing “old news,” Caroline Kovac, vice president of IBM’s life sciences unit, said Tuesday that Big Blue is aggressively targeting the data management challenges of life science research.
While IBM most recently made news in the life science computing sector by installing a 7.5 teraflop computer at NuTec Sciences in December and a 700 gigaflop system for MDS Proteomics in January, Kovac said, “Data management is every bit as important, and maybe more important [to the life sciences] than supercomputing.”
IBM said the push into data management would not impact their hardware business.
IBM intends to challenge Oracle’s domination of the industry by promoting its DiscoveryLink data management and integration system to pharmaceutical and biotech companies as well as academic research groups.
DiscoveryLink combines IBM’s DB2 database platform with a relational connect technology and wrappers that permit the system to integrate a number of public and proprietary tools and data sources. The core of the system is an optimizer in DB2 that rewrites the queries of the native systems’ search tools. The result, said Kovac, is a “federated database” that permits users to access distributed and heterogeneous data sources without porting them into a central repository.
Current DiscoveryLink customers include Merck, Incyte Genomics, and MDS Proteomics.
Janet Perna, IBM’s general manager of data management, said that a pilot demonstration of DiscoveryLink at an undisclosed pharmaceutical company was able to run Oracle eight times quicker than the native Oracle system could on its own, due to the DB2 optimizer.
However, IBM faces competition in this space from providers of data warehousing solutions, who claim that a centralized repository of curated data serves the needs of large, distributed research organizations far better than the federated approach provided by DiscoveryLink. Viaken Systems, for example, recently inked a deal with EMC and Exodus Communications to meet its customers’ growing need for increased data warehousing capacity.
Proponents of data warehousing claim that because the data in a federated approach is not cleaned, sorted, and organized as it is in a data warehouse, access to the data is much slower.
Kovac noted, however, that the best solution may be a combination of data warehousing and a federated system such as DiscoveryLink, which she said is better for data sources that are constantly changing.
Rather than competing with information management companies in the life sciences arena, such as Lion Biosciences or InforMax, Kovac said IBM would seek partnerships with these companies. She said that IBM would be able to offer its infrastructure experience to these smaller companies, that typically don’t have the capability for robust deployment of their tools in the distributed environment that pharmaceutical companies require.
Instead, Kovac said that IBM’s primary competitor in this segment is Oracle. Perna noted, however, that growth of DB2 has been over 70 percent annually over the last several years and that over 1000 Oracle customers have turned to DB2 in the past year.
IBM’s traditional competitors in the high-performance supercomputing market, Compaq and Sun, have no data management solution at this time, Kovac noted.
Since launching its life sciences initiative in August, Kovac said, IBM has announced a new partnership nearly every month. She forecasts that this rate will continue into the future as IBM strives to capture its share of the IT market for the life sciences, forecast to reach $40 billion by 2004.