Caught in the Middle

IT shops are caught in the middle of the transition to open and integrated – and just in case you were wondering it is tough sledding. IT shops not only have to contend with legacy internal systems many of which have poor integration capabilities; but also legacy commercial applications which have been deliberately left to become obsolete with minimal upgrades. These commercial applications often linger on because they were quite good in their time but just did not reach critical mass or have been deliberately shunted aside by their vendors newer software. Think HP Tru64 OS and OMNI database, IBMs OS2 OS and COBOL programming language, Microsofts Win/NT OS and .NET Framework 1.x, Oracles Forms and Reports plus BI-ware – ohh sorry that has been resurrected and born again as Oracles new Business Intelligence offerings.

To make matters worse the transition to Web Services, SOA-Service Oriented Architecture, EAI-Enterprise Application Integration, ESB-Enterprise Service Bus and other open and integrated system designs and architectures is complicated by 4 major shortcomings:
a)the new stuff is 2/3rd baked but key ingredients and standards are missing – look at WS-I for starters of what is missing;
b)the new stuff is barely tested and benchmarked; very few organizations know what works well;
c)the new stuff even when working leaves key security, availability, and performance gaps which will require either extensions/refinements or permanent workarounds and/or replacements;
d)open and integrated is either being fiercely resisted (Microsoft with its Web Services only ports of entry into the evolving Gates of Longhorn bastion, database vendors resisting universal standards to make heterogeneous joins and database operations a reality; etc) or benignly neglected (JavaScript vendors blithely doing their own thing almost without even considering it as one of the best candidates for a uniform, cross platform scripting language filling a huge gap in desktop, mobile, Web Services and SOA architectures) . Like Longhorn being at least 2-3 years away from emergence so are fully capable SOA and ESB architectures. Yes, parts of Avalon are being delivered or read-primarily SOA designs or steps of the message oriented architecture can be delivered today – but that leaves a lot of gaps to be filled. The whole problem is something of values is being abandoned, deserted if not deliberately obsolted without something of completely equal value being available as replacement. Sure, more than equal in the vendors eyes – a lot less so for users sticking defiantly with legacy.

Do More with Less

Besides being caught in the middle, IT shops face a relentless almost mythically absurd drive to reduce costs for ever fleeting temporary advantage in the marketplace. This is now becoming the number one moral and ethical problem in business: blind, bland “progress” is no longer our most important product. Over and over again societies are going to have to learn that the ease of use and lower cost of CFC for refrigeration is not worth the cost to society of ozone depletion , skin diseases, and cancer deaths. Or new drugs that go off like time bombs 2-5 years after introduction with pernicious (though hard to detect at the outset of introduction)side effects. Or the GUI ease of use, operation and installation with the most applications is not worth price of a continuing stream of very late to be delivered reliability, scalability, availability and now security and interoperability problems with Windows and Windows applications. Demanding systematically better is very hard to do when an immediate met need/requirement is arrayed against one, or more likely a set of cumulatively worse “future” costs.

Finally, to complete the caught in the middle theme, IT shops are being asked to do more with less in resources like dollar, manpower, and/or infrastructure just at a time when computing and IT have become the critical “command and control” system for most organizations. So catastrophe enabling solutions are being invoked: across the board cutbacks, hiring freezes, staff churn, outsourcing instead of informed redeployment of scarce resources. It is sort of like the Relief Aid for the Tsunami in Southeast Asia – the problem is not the dollar amounts or the willingness to help – but the ability to co-ordinate to deliver the right aids, at the right time, in the right places and the right quantities. Given the backlog of projects and persistently dwindling resources, IT shops are really in the relief aid business – better put the tools and practices in place for this type operation for some years to come.

(c)JBSurveyer 2005