InformationWeek: The Business Value of Technology

October 2, 2000

Printer ready

Editor: Karyl Scott (kscott@cmp.com)

Pushing Boundaries
IBM is pushing hard against the boundaries of data processing and data integration in a way that could ultimately affect everyone's health and well-being. Through a technology licensing agreement with Incyte Genomics Inc., IBM is applying its DiscoveryLink technology to the challenge of sifting through mountains of genetic data to better identify the causes of diseases and aid in the development of cures. The agreement will integrate DiscoveryLink's data-integration technology with Incyte's Genomic Knowledge Platform. DiscoveryLink will let researchers using Incyte's platform to access and analyze biological data to uncover complex patterns and associations.

The complexity of genomic analysis presents a staggering challenge to the data-processing tools applied to the task. Incyte's new software platform integrates data from multiple genomic-analysis tools and varied data formats to let researchers see patterns and data dependencies that are difficult or impossible to apprehend using isolated tools. Data such as DNA sequence, gene expression, and genetic variations will be able to be combined and interpreted for patterns and associations that could well lead to unexpected discoveries.

IBM's efforts are spearheaded by Caroline Kovac, VP of IBM Life Sciences. She's charged with developing, integrating, and applying IBM's IT services, software, and hardware, building partnerships and making strategic investments in companies that can have a high research and development impact in this area. IBM says it has earmarked $100 million to develop advanced research technology for the biotechnology, genomic, E-health, pharmaceutical, and agri-sciences industries.

"What's needed in this field is a knowledge-management system that lets researchers query across any and all of this data, write complex queries, and have the data appear to be coming from a common repository," says Kovac. Data warehousing doesn't do all that well with disparate data, and it also has a rigid schema, so you can't add new sources very easily. DiscoveryLink helps researchers integrate data in a flexible and scalable way.

The hardware side is also changing, Kovac says. Historically, biological computing has been done on supercomputers, but there's now strong interest in Linux clusters. Many tasks that biologists and life scientists need to do can be partitioned to run on Linux clusters. Says Kovac, "Some of the most interesting installations we're doing are hybrids with our SP2 machine and Linux clusters." --Ron Copeland

Illustration courtesy of fredrix.com

---

Online Transaction Revolution
The Universal Description, Discovery, and Integration standard being proposed by IBM, Microsoft, and others could revolutionize business-to-business E-commerce and online marketplaces by standardizing the way businesses interact over the Internet.

UDDI is a global online registry that maps businesses to the Web services they offer. A Web service in UDDI parlance could be an application that provides a real-time travel advisory over the Web or a collection of application modules that comprise a currency conversion calculator.

A global registry could be the ultimate search engine for business-to-business E-commerce, enabling companies to discover other businesses and find out what's necessary to access their online services.

The goal of UDDI is to eliminate the custom connections required for business-to-business interactions over the Internet and provide a standard way to integrate applications and business processes for online collaboration and transactions. This standardization will be especially meaningful to the many-to-many business model emerging for online marketplaces.

UDDI will let businesses dynamically publish, discover, and aggregate Web services to produce new products, business processes, and value chains, according to IBM officials. Ariba and Sun Microsystems are among the 50 other companies involved in the effort. That kind of business momentum doesn't guarantee success, but it does guarantee a serious investment at success.

UDDI is, first and foremost, a sophisticated searchable database. It contains descriptions of businesses, the Web services they offer, and interface details that describe how to connect to those services. The database, called the UDDI registry, is globally distributed and maintained in a similar way to the existing domain name system that maps domain names to IP addresses.

Any business can register its company information, the Web services they offer, and their programming interface information, with a UDDI registrar. Initially, only Ariba, IBM, and Microsoft will be registrars and will operate sites where businesses can register. Registration will be available starting Oct. 4.

By registering your business at any UDDI registry site, your service information will automatically be propagated to other UDDI sites. Initially, registration will likely be done via a Web form. The registration is stored as an Extensible Markup language data structure. Registering with UDDI and searching the registry will be free of charge at the Ariba, Microsoft, and IBM sites. Any other registrars can set their own prices and may offer other bundled services. Searching will always be free of charge.

Last week, IBM and Microsoft unveiled the Web Services Description Language for the creation and use of Web services. It builds on Microsoft's Simple Open Access Protocol contract language and IBM's Network Accessible Service Specification Language--specifications that are part of UDDI. Soap, a remote procedure call mechanism that eschews the complexities of DCOM and Corba in favor of XML, is used in UDDI by outside applications to make queries to the UDDI registry, but it can also be used by businesses to build interfaces to online services. Soap is based on XML. NASSL is IBM's Web-service specification. WSDL builds on and extends these existing specifications.

UDDI uses XML data structures to describe business services. These definitions are very general. Services don't necessarily have to be available online, nor do they have to be built using a particular computing platform. In fact, there's nothing to stop a physician or lawyer from registering their services with UDDI. Clearly, the usability of UDDI will depend on how carefully the registrations are crafted.

Back to This Week's Issue
Send Us Your Feedback
Top of the Page