Sunday, May 11, 2014

Enterprise Architecture: Musing on Buring platforms and Cycle of vendor upgrades

The recent announcement by Microsoft on discontinuing support (even paid extended, extended support) for Windows XP and the debate in the media has broader implications on burning platforms at hand. To those of us engaged in vocation of Enterprise Architecture, the topic of burning platforms (a.k.a IT Debt @Gartner) continues to manifest itself in different forms and at different times. It is an event that seems to surface with regular enough frequency.  And the issue is not just restricted to an IT vendor. On the contrary, the Microsoft XP debate just surfaces a larger issue at hand: the lack of disciplined Application Portfolio review and Management (APM).

The need to upgrade a collection of application platforms (open group definition may have a cascading effect across the ecosystem. E.g for a big bank to plan for upgrade of ATM (link) systems from XP to Windows 7 (or 8) it is not just a shift in versions of software running on the ATMs. Implications to hardware (obvious need to suddenly also upgrade hardware), networks (potential increase in bandwith), integration…. the whole nine yards. A version upgrade of OS requiring an enterprise - or at least division - wide APM exercise. Such an APM exercise focused on addressing a burning platform can generally be a knee jerk reaction that addresses “A” problem than the symptom. We don’t hear much of this topic, especially the strategic implications of burning platforms for several reasons:
One reason is because it is just not sexy to talk about issues surrounding “keeping the lights on” Many Architects would rather spend their time “architecting” and designing new systems and engage business, than in planning the upkeep of ecosystem on hand. Such upgrades are generally relegated to IT/IS service teams and generally written off as operational “cost of doing business”.
Ø  In a resource constrained environment where most organizations are looking to stretch every dollar; not taking a strategic view of vendor roadmaps and upgrade plans and its implications on one’s existing architecture is not only counterintuitive and surprisingly wasteful. 
Another reason this topic doesn’t get a lot of airtime is because many corporate Enterprise-IS/Business-Architects and IS leaders are busy reacting to individual occurrences of such burning platforms; and are busy negotiating workarounds with vendors and technology partners.
Ø  No point in rocking the boat by airing dirty laundry?
Perhaps the real reason is a clear lack of communication. There is a lack of systematic and periodic effort in trying to understand and account for the total cost of such changes. 
Ø  Without being forewarned, Business counterparts react to the challenge like a deer caught in a headlight.
An idea worth exploring: A simple dashboard that can easily show the complexity and highlight the TCO implications to a CIO’s business counterparts would be a cool tool to have