Intelligent Enterprise

I take the topic of this blog literally: Will the enterprise enter the "mashup" maze? IBM thinks so. IBM would not call it a maze. They call it the IBM Mashup Starter Kit and recently put it up for (free) download. It's part of their angle on Web 2.0 applications, in fact one of the leading parts (along with webtwoifying Lotus Notes). What IBM is doing with mashups is interesting for a number of reasons, but perhaps the most significant is that IBM is pushing mashups at all.

It's not that mashups (however defined) aren't being taken seriously in the enterprise – for the most part they're not. It's that IBM, the primus inter pares of enterprise vendors, has chosen mashups as a key product for development.

The term mashup is very trendy, which as usual translates into many different ways of describing it. Most simply: A mashup combines program components and data from different sources into a single integrated Web application. The best-known application is Google Maps, which is used to compose maps with specialized data, for example a real-estate map. By now there are thousands of other mashup applications and dozens of companies rushing to produce or perfect their mashup makers. IBM is one of them.

I won't attempt a history creating composite applications… dream or reality; they've been a gleam in many a developer's eyes for decades. Whatever the inspiration — Object Oriented Programming, CORBA, etc. — for a variety of reasons eventually the gleam guttered. The current hope/belief/intuition is that mashups may succeed in becoming a significant way to build applications because: Now there are relevant and widely accepted standards (Web services standards in particular), there is enough cumulative know-how about applications built from components, and there is the vast pool of interconnected data and program sources provided by the Internet.

Predictably, mashups are already being sliced and diced into new categories, for example: Consumer, Data and Enterprise. There will be more categories, accompanied by various methodologies and products. This is where the maze comes in; there will be many choices, perhaps too many. Complexity of this sort was certainly one of the killers of previous composite application development. Then there is the matter of inherent complexity in managing components and data from multiple sources.

Two potential killers here are security and control, or lack thereof. It takes very little imagination to see the possibilities of mixing mission-critical corporate data with ad-hoc internal data or even with public domain data. How do you spell IT nightmare? This is where IBM comes in. If any company understands the sensitivities of enterprise IT, it is IBM. So when IBM begins evangelizing Web 2.0 and mashups, it is almost an article of faith that security, control and management are part of the mix.

IBM has developed a mashup server (IBM Mashup Hub), which like an application server will have the capability of doing mashup management. However, the capability is embryonic. In un-IBM fashion, IBM is offering its mashup software in rolling updates. With these updates, IBM has promised to make mashups secure for enterprise IT. It will be interesting to watch how it gets there, because quite likely that will be a way through the mashup maze.