And so the three chefs sat around this warm, steaming, freshly baked, deep-dish apple pie, enjoying its sweet aroma yet anticipating the dilemma of how to slice it fairly for sharing among their respective customers.
"I have an idea," one of the chefs said. "Whoever slices the pie gets to pick the last piece."
"But my customer is the largest consumer of pies, said another, "and should have the biggest piece After all, the pies are only as cheap as they are because my customer buys so many."
The third leader opined, "My staff has a super-duper, enhanced, upgraded microslicer with the latest BIOS release, that can help us cut exactly equal por tions. I'll get three of my people in here as soon as possible, and we'll have equal slices by the end of the week."
The other two chefs looked at each other and nodded in agreement. So the three chefs went about their business, confident that they had arrived at the most equitable solution for sharing the pie...
Information technology organizations today continually find themselves in this quandary: What is the most appropriate (i.e., accurate and equitable) cost recovery methodology to use in a corporate structure with multiple independent operating units sharing a centralized systems organization? Financial support organizations have been "caught unaware" because recovery methodologies that worked well in the glass house have proved inadequate in today's glass menagerie.
Enter the Dragon
The paradox of a decentralized, client/server architecture, which still requires, bandwidth, routers, etc., in its backbone, has left the bean counter with dollars to allocate the numerator but no basis left for further allocation. Some time ago, even data network cost recovery could reasonably piggyback on mainframe utilization. Since then, host systems have become great places to maintain humon gous amounts of data, but users like both the control of manipulating data, and the graphic user interface front ends provided by local (client or server, shrinkwrapped or home-g rown) applications.
Thus, the pendulum has begun to swing away from a model where CPU utilization was representative of network consumption toward a model where the number of entry points into the network drives cost. Unfortunately, today's networks, comprised of multicolored zebras (many different devices supplied by many dif ferent vendors and manufacturers), do not lend themselves to any reliable basis on which to employ algorithms.
Enter the Politicians
So, how to recover these informationtechnology costs that evolving technologies now leave us holding? The answer lies not only in the tools available (pur chased or home grown), but also in the corporate culture and organizational structure that exists. Consider addressing the following two models:
The IT organization has complete autonomy/authority in determining appropriate IT resource levels for the corporation, or the IT organization yields to political pressures, based on reporting structure, in designing, implementing, and supporting architectures.
In an organizational structure where it is a stand-alone entity (a "peer" of the operating units), that organization must be chartered with provisioning IT resources and services at a level appropriate for the corporation, not for the operating unit(s) with the most political clout. In order for this model to succeed, however, a basic tenet of maintaining centralized IT resources must hold true:
Pooling together the operating units' demand for T resources brings significant negotiating and purchasing power (i.e., better pricing, lower costs) to the corp ration as a whole. Making volume commitments to vendors/suppliers will benefit the corporation's ability to leverage vendors' pricing and technological direction, and will not jeopardize any operat ing unit's ability to achieve and/or maintain a competitive edge.
Many of today's charge-back models are not developed in high-level, closed-door negotiations or work sessions. …