The information is piling up. While pundits may argue justifiably that the swell of cyberstuff has netted us little more actual truth, none can deny that this sea of easily-retrievable data is flooding upon us in an irreversible tide. Slotting information into database pigeonholes seems a nice quick-manage solution. But how do you make your databases function as a working library rather than a labyrinthian archive?
Guidance is available from the New Jersey Chapter of the ACM/IEEE, which is sponsoring “Adventures in Databases” on Thursday, February 16, at 7:30 p.m. at Sarnoff. The speaker is Tom Laszewski, an executive with Oracle. Call 908-582-7086 for more information on the free event. There will be a pre-meeting dinner at Ruby Tuesday’s restaurant at 6 p.m.
For the past two decades, Laszewski has been a firsthand player in developing the systems that not only generated the data explosion, but that have made it more manageable. He grew up in Stevens Point, Wisconsin, where his hard-working family saw no need for college. But he prevailed on them, and went off to Wisconsin University and earned a bachelor’s degree in management information there in 1986.
“At that time,” Laszewski recalls, “all information management involved the old VCM and main frame databases.” Then, as he was obtaining his master’s degree in computer information systems, Novell, and soon after, Microsoft, created systems for database storage on PCs. “If only I had bought stock in either of those firms back then,” laments Laszewski.
After graduation, Laszewski joined Sybase and Electronic Data Systems as a developer. For the past 10 years he has worked as technical director at Oracle’s Partner Technology Group, creating many of the database advances of which he will speak.
“We stand on the cusp of a whole new body of technology that will make not just databases easier to manage, but make our information more accessible, digestible, and secure,” says Laszewski.
AI + DB = >$. The mundane platforms of database management may seem scarcely the place for the exotic applications of artificial intelligence, but it is from this marriage that some hefty cost savings are born. Most companies employ a lot of staffers, whether in-house or outsourced, to constantly update and tweak their computer systems. These workers devote valuable time to reorganizing and speeding up the info-flow process.
Increasingly such fixes and improvements can be handled by a new wealth of self-managing databases (or as IBM with its euphemistic flare labels the phenomenon — “atomic computing”). Databases can now be programed to re-index and reorganize automatically. They can even be trained to seek out and proactively enact shortcuts, turning a two hour process into two minutes.
Laszewski admits that it takes a while to train both human and machine to their new relationship. While the docile computer can be trained fairly easily, people are often shocked by the latest patches as they are installed. “It takes a bit of adjusting for both it and us,” says Laszewski.
Fickle dedications. The old idea of clustering has now taken another step forward. Laszewski jokes that the new age of clustering should not be confused with the original schemes of linking all PCs to search space for signs of extraterrestrials. Rather, in this more down-to-earth application, a collection of dedicated computers can now shift dedications as seasonal needs require.
The business that has 20 machines in its plant may already have a clustering style of software that shares all basic information throughout all divisions of the company. If so, accounting can find out what shipping and receiving are doing. But now, by establishing what Oracle calls an enterprise grid, dedicated machines in any department can shift and become dedicated to another department’s tasks.
This means that during the pre-Christmas rush, when the shipping department needs extra capability and machines, it can commandeer them. Then as the year-end crunch hits accounting, that department can get the help it needs. Technically, such sharing is already possible on old connected computers, but with this new application, the network traffic is vastly diminished.
Security race. Not too many years back, database access involved the client interacting directly with a monolithic system, for example, a Unix box. But now in the era of middleware, we talk to databases via websites and even E-mail. Through this prism of middleware, the old user name connector is fragmented into a complexity of web names, database names, and user IDs. This complexity also allows various names to more easily be picked up and back traced.
The current solution seeks security through a return to simplicity. Some database and systems handlers now offer a single “meta-user name,” which doesn’t get disclosed on each transaction even as it provides access to E-mail, websites, and databases — all in one. Such meta-IDs are encrypted all along the line, end to end, thus foiling most all hacking attempts.
Shattering structure. Traditional documents are structured — that is, the data is fixed in rows and columns, all neatly indexed. Then, thanks to E-mail transmissions, unstructured documents began entering and leaving many departments, and creating a muddle. But now information may be fed into databases shaped on how the client wants it cross-referenced, with a mix of traditionally structured and unstructured languages.
The data flood presents the same problems as any bulging library. We need more entryways — many of which have already been built through the Internet and E-mail. But once inside, we need wider hallways to accommodate the increased traffic and smoother ways past the stacks and stacks of stuff we don’t need, to the nuggets of information we really want — right now. According to Laszewski, technology is one step ahead of the info-overload problem. Now, if we poor humans can only keep abreast.