Infrastructure Pattern Matching

By Nick Gall  In a column last May, I wrote about infrastructure development (ID) as a new discipline supporting--but distinct from--application development (see "Help Wanted: Mi ddleware Mechanics," www. NetworkComputing.com809/809colgall.html). ID focuses on the infrastructure upon which applications are deployed: the networks, servers, storage and middleware. Since then, I've spoken with many IT organizations that have ID groups (in function, if not in name) in various degrees of maturity. These groups range from embryonic ad hoc teams focused on an immediate infrastructure issue, such as the deployment of enterprise resource management suites from vendors like SAP, Baan and JD Edwards, to formally chartered departments with continuing infrastructure oversight. Marketing Infrastructure to Business Regardless of their size or maturity, all of these ID groups are seeking an answer to the same question: How should IT market infrastructure to business? Traditionally, IT has employed two techniques in carrying this off--the pork-barrel approach and the Chicken Little approach. The pork-barrel technique seek s to bury infrastructure investments in the massive budgets for major application initiatives. For example, many organizations' TCP/IP infrastructures were initially paid for as minor line items in SAP R/3 deployment budgets. The Chicken Little approach, on the other hand, seeks to strike fear into the hearts of business users by suggesting that unless an infrastructure upgrade/overhaul is done immediately, some IT disaster, like outages of mission-critical systems, security violations or bottlenecks slowing key order-processing systems, will surely befall the organization. These two techniques can eat away at an IT organization's credibility and undermine its standing with business decision-makers. The pork-barrel approach forces IT into a reactive role in which infrastructure decisions are made in a piecemeal application-by-application manner. The resulting infrastructure will be fragmented, overlapping and difficult to manage. The Chicken Little approach can spontaneously combust if business users d ecide not to fund the supposedly critical overhaul/upgrade and the predicted disasters never occur. Articulating Useful Business Drivers Leading IT organizations are adopting IPM (Infrastructure Pattern Matching), the emerging ID-perspective strategy, for justifying infrastructure investment in business terms. The key assumption underlying this technique is that business process patterns can be mapped to IT infrastructure patterns at levels high enough to be useful in communicating infrastructure. And these patterns also are detailed enough to provide meaningful guidance to information technology decision-makers. In other words, IPM gives IT organizations a better way of communicating with business users, doing away with the pork barrelers and the Chicken Littles. Although it follows common wisdom in IT organizations today that decisions must be aligned with business drivers, no one seems to know how to articulate functional business drivers. In most of the IT plans I've read, business drivers a re actually platitudes used to gain competitive advantage, provide adequate response time and scale to meet business growth. Such bromides provide no insight into how to achieve such goals within IT.

Other Columnists this issue Corporate View By Brian Walsh On The Edge By Art Wittmann Other Articles by Nick Gall Infrastructure Pattern Matching Business Is Getting The Middleware Message Next-Generation TP Monitors: Are You Ready? Three's A Crowd With Object Lessons Help Wanted: Middleware Mechanics