My administrative assistant's husband is a plumber. She drives a better car than I drive. They have a summer home on Cape Cod. I have a Timberline tent. There are big bucks in plumbing.
So, too, will there be big bucks in a new kind of software "plumbing," called application integration software by my colleagues at IDC. This is software specially designed to enable disparate applications that were independently developed and deployed to interoperate. It's designed to tie back-end and front-office applications together, to link ERP and CRM systems and, someday, all the different IT systems in one supply chain.
I know IT departments routinely integrate applications as part of their deployments of new systems. And a whole branch of the IT services industry does nothing but application integration.
But this integration is different. Today's integration is generally from a single-source application to a single-target application. Or between a small handful of applications. The result doesn't scale, is hard to maintain and isn't extensible.
This software works more in a hub-spoke fashion, more or less so depending on how the integration is implemented. The most sophisticated tools are message-oriented, where special middleware sits in a hub and manages a database of business rules associated with all the different applications - the spokes - that need to interoperate.
IBM, Tibco, BEA Systems and a host of other vendors are pushing the envelope here. The simplest form of application integration works at the point of access, with screen scraping and terminal emulation rudimentary examples. This kind of software would, for example, take an input from an e-commerce portal and simultaneously send messages or inputs to multiple legacy systems (and gather responses). Attachmate, eGrabber and Hummingbird all work in this space.
A third type of integration can take place at a data integration hub that sits between the data sets intrinsic to different applications, sort of a marriage between data warehousing and data extraction software. Vendors here include Constellar, Saga, Cohera and many others. (E-mail me for a more comprehensive list of vendors.)
I'm telling you all this not because plumbing is exciting, but because plumbing breakdowns are. In the coming era of e-business and e-services, where the solutions we want to deploy have to span not just enterprises but also ad hoc consortia, such as Web marketplaces owned by competing companies, I don't see how we can go without some kind of unifying software. And the big applications vendors don't seem keen on developing tools or features that allow their software to interoperate with competitive products, so we're stuck with this growing third-party market.
You should probably be looking at this stuff now. It's going to be easy to pick which vendor or approach will be best, and there will be a learning curve associated with designing an application integration architecture. But any one of those packages will probably be better than hand-rolling integration tools or code. If you buy one of these packages, you won't be alone. A worldwide market of $800 million last year is expected to grow more than tenfold by 2004, to almost $9 billion. The bandwagon is rolling.
JOHN GANTZ is a senior vice president at International Data Corp. in Framingham, Mass. Contact him at jgantz@idc.com.