PALO ALTO, Calif. -- Can enterprise relational and object databases be empowered to do more than simply store data, then spew it out in the correct form as needed to people who access it? Some experts' consensus answer is: You bet they can. In fact, this extended work is happening with more and more frequency all the time.
What is this extra duty, you say? Gathering important business intelligence, then slicing and dicing it to order, so that business executives will have more knowledge -- and thus a better business advantage -- to make key strategic decisons.
A Software Development Forum panel Wednesday tackled the topic "What's Ahead for Databases" at the Palo Alto Research Center. Panelists included MySQL co-founder David Axmark, Jon Rubin of IBM, chief data architect Kevin Perry of Inovant (a subsidiary of Visa International), magazine editor David Stodder, and moderator Steven McHenry of SD Forum.
"How valuable would it be for a company -- let's say Visa -- to find out what its customers are thinking by using a Web crawler to troll the Internet, gather opinions about Visa cards transactions (in blogs and chat rooms, for example), and store that massive amount of data in a database for further dissemination?" asked Rubin, senior DB2 product specialist for solution development and a former Gartner Group database industry analyst. "That would be very useful information. What a job to put something like that together, for sure, but it's happening now."
As database systems become larger and add more features, the job of aggregating customer relations management information and supply-chain data will fall more and more to the job of the database administrator, Rubin said. "If the database is set up correctly, a tremendous amount of new business intelligence will become available, and afford a competitive advantage for those who know how to use it."
Are DBs still relevant as a stand-alone item?
Asked whether the database still relevant by itself as opposed to being simply a function of the larger enterprise system, Axmark said, "I'd never put my data into a form I didn't know. Right now we have Java, .NET and other languages, but there have been many languages before, and there will be many more to come after this. I'd never see the database as a commodity, it needs its own special attention -- at least in the way we use them today."
Rubin said that when business executives purchase enterprise platforms, they look at the system as a whole to do the job for the company. "But if you're the IT guy, that database in no way is a commodity, and probably not interchangeable," Rubin said, pointing out DB characteristics such as manageability, recoverability, and service-level issues related to specific databases.
Perry of Inovant said, "I consider the database a commodity, and we don't drill down much in Oracle's calls or Universal Database's calls, or anyone else's calls. We also try to avoid the features that set DBs apart ... there are exceptions for certain custom extensions, but generally we like to keep our system as simple as possible. Our value is totally in the data."
For example, drivers should be swappable, Perry said, pointing out using DB2 Universal Database in place of another database. There are times, however, when a specific feature of a particular database is desirable, he said. "For example, for one large customer, we incorporated a high-speed Informix loader for SQL, because that's what we needed at that time. But those (situations) are rare for us."
Open source vs. proprietary databases
Perry said he has trouble getting his arms around using open source databases in his ultra-high-volume business. "I need to be able to grab a neck close by that I can get my hands around, and that I can tell, 'This thing doesn't work as advertised,'" he said. "Most big companies really seek out commercial vendors, I think, for that purpose. The perception is that service will be better."
Perry got an immediate argument from Axmark. "You should already know that we at MySQL have 24/7 service and support," Axmark said. "You can get your hands around the necks of one of our guys just as well. We are proud of our service and support."
Axmark admitted that MySQL "doesn't have near the number of features that DB2 has, but we've been able to make quite a few people happy with our product," he said. "We tend to focus on the smaller, subtler features, that people use every day, rather than the bigger features you see in DB2 and Oracle."
Axmark added that although MySQL is an open-source database, "We've been a commercial company from Day One. If you pay us, we give you good service. If you don't ... well ..."
XML like teenage sex?
XML as a data format was another topic. Rubin, for one, doesn't see very much XML being used in business today.
"XML is like teenage sex," Rubin said. "Everybody talks about it, and thinks everybody else is doing it, when they're really not. XML means that the extensibility has rich new capabilities. If there's a way to get object or relational expressions in an XML format, if we can get all data on a common platform, and put it together for the large context, then there's a lot of value there. I'm not seeing it at the moment, though."
The panel agreed that DBs as we know them today aren't going away anytime soon.
"I think they will stay for a while," Axmark said. "It takes a while for people and businesses to change (to new technologies). Most of the features we have don't go into the standard, and people will add the features they want to use, when they want them. After all, people are still using COBOL."
Rubin agreed. "Things do change slowly. We can get features for developers much faster than we can change the expectations of customers," he said.
SD Forum, the panel sponsor, is a not-for-profit organization with deep ties to entrepreneurs, developers, venture capitalists, industry experts and major technology companies. It has 2,000 individual members and puts on 20 to 30 speaking engagements, workshops, and seminars each month.