Data, Data EverywhereBy Michael Vizard | Posted 2006-08-07 Email Print
A tech management crisis is upon usand there's no bulletproof solution in sight.
For the past several years, the information-technology industry has been diligently working toward making it easier to access data locked in various applications as part of an effort to help I.T. organizations derive more value from the data they collect. While this remains a noble goal, an ugly reality is starting to rear its head: the pending data management crisis.
Back in the bad old days of client/server systems, we could reasonably expect that all the data we needed to keep track of would be in a few disparate databases. But with the rise of XML and Web services, this data is not only in databases, but in file systems, applications and a wide range of personal computing devices. This is making it more difficult to establish a data warehouse where the "one truth" concerning the state of a business process can be found, because data that has a material impact on the results of that business process is basically everywhere.
The end result is that as I.T. systems increasingly collect more data, it's becoming tougher to reflect the true state of the business. In fact, recent surveys on information management practices by CIO Insight, a sister publication of Baseline, find that 53% of the CIOs surveyed said the majority of the data they collect goes unused; 52% said that users frequently complain about not receiving the data they need.
And to make matters worse, as the number of interdependent business processes continues to multiply across any number of composite applications linked to multiple data sources, it will become increasingly difficult to create relevant reports about the true state of the business, no matter which government mandate is in place.
To try to deal with this problem, some application vendors such as Cognos have developed their own meta-data structures that keep track of related data sets residing in diverse applications. And middleware vendors such as IBM and BEA have created next-generation data management tools that sit on top of their respective application server products.
But even with these tools in place, the industry will still need to create a set of standards for sharing meta-data so that the meta-data system used by, say, Cognos can share information about how and where data is stored with any number of other data models. Hopefully, some standards body will eventually address this issue, but probably not before most I.T. organizations experience some pain because of it.
In the absence of any real standards, there is hope in the form of pure middleware companies such as webMethods. At one time, companies like webMethods were the darlings of enterprise application integration. But as that function became a commodity feature of application servers, these vendors needed to reinvent themselves. Given that necessity is the mother of invention, companies like webMethods are discovering that the next great opportunity in the enterprise lies in business process integration across multiple applications and data sources.
For example, Ingersoll-Rand relies on webMethods to create a platform that not only insulates the business from specific application interfaces, but also gives the business a starting place to determine its one true state because changes to applications and data are tracked in real time.
As Ingersoll-Rand CIO Barry Libenson puts it, for the first time in history you don't have to have a Ph.D. to create a distributed computing environment. But the downside is that you can't stop the proliferation of applications, so what's required is a higher level of management abstraction that integrates business processes rather than applications.
Instead of having a series of discrete supply chain and CRM applications, what we have emerging today is a series of interconnected components from those applications that constitute a composite application reflecting a specific business process. For example, the front end of the supply chain application is actually a module in the CRM application. And when a process occurs in the supply chain application, it in turn kicks off an update to the general ledger application, which then sends an alert to a business intelligence tool that is monitoring the overall process.
While all this may give the average I.T. person pause, the good news is that anybody with expertise in data management pretty much has guaranteed lifetime employment. The bad news is that he or she might be long retired before these issues actually get resolved.