Dr. Ambuj Goyal is one of IBM’s heavy hitters. A 25-year IBM veteran who joined the company as a researcher at the T. J. Watson Research Center, Goyal did early work in scalable databases that laid the technology groundwork for DB2, then led the research effort to create the Deep Blue chess computer. In 1996, Goyal was elevated to vice-president, servers and software, and director, computer sciences, where he set IBM’s long-term research direction and oversaw the work of 1,500 researchers worldwide.
Now, as general manager of information management in the IBM Software Group, Goyal ensures the success of IBM’s Information On Demand initiative and its constellation of products. But in a recent interview with InfoWorld executive editor at large Eric Knorr, Goyal still sounded like a scientist, expounding on the nature of information vs. data and the unique position IBM’s new Information Server occupies in the history of computing.
Q: Consolidating, reconciling, and delivering information strikes me as one of the most difficult problems in IT.
A: This problem has existed for 20 years. When I was doing the announcement for Information Server last October, I talked about an inflection point that has come in information management. This is similar to an inflection point that occurred in 1996 when there were many techniques to create Web sites or do Web-based business: Java Virtual Machines, HTTP servers, commerce engines. There were many different techniques. A few companies created a concept called a Web application server — IBM’s being WebSphere. So if you wanted to leverage Web technologies to improve your business or do Web programming, you just get started with a Web application server. We are at a similar inflection point in 2006. We have myriads of techniques — metadata management, ETL (extraction, transformation, and loading) tools, data creation tools, Federation tools, cleansing tools, profiling tools. People use these tools to solve the information challenge. The application server had a foundation, which was JVM.
Information Server now has a foundation, which is a metadata repository, or metadata bus. And we have restructured all the tools to come up with an information server, so companies don’t need to decide whether they need this piece of technology or that piece of technology. They have an information challenge — they just start with an information server.
Q: What about the skills to work with an information server? It seems a little difficult to find people who understand this area. Is that a problem in the market now?
A: It actually simplifies the problem because people were doing manual stuff around metadata. So what we have done is helped the business user say, “This is how I define my customer and attributes of that customer.” They can talk business language. And then an information architect, who typically used to do ETL or SQL, can take the same business language and convert it into an appropriate information infrastructure programming tool. In the past they could never do that, and governance became a huge problem. The business person really meant X, and the information architect implemented Y.
Q: But in all of these solutions, somebody is still sitting down in a meeting room, going through the incredibly tedious task of saying, “OK, we’re going to use this field and not that field, we’ll express this value this way and not that way.”
A: Right, but our view is, over decades, this is how software and middleware have evolved. There was tedious work going on connecting system A to system B. You repeated it over and over again. Then out came a product that is an outgrowth of a repeated piece of work captured into a software product.
In the late ‘90s, people used to use the WebSphere application server and would build customer portals. Many customers built and said, “Look, this is a waste of time. I would just like to have the base infrastructure to be able to do that.” So portals as a category evolved.
We believe that people are still doing tedious work when dealing with information. Information Server and its tools are converting that tedious work into an information server category. Even analysts are starting to talk about this as a category because it is solving that tedious work, and tedious work has always resulted in a new category of software.