I have just been reading an IDC White Paper that takes a ten year look ahead at IT technology. First and foremost IDC sees an era of self managed systems. Systems will be so sophisticated and complex “they will have to manage themselves”. AI and expert systems will be applied to managing systems. The “operator” may be a automated control system doing many of the chores of load balancing and system recovery in a self diagnosing and self recovering fashion.
Next, IDC foresees end users exercising more power over data processing funding demanding measurable performance standards in service level agreements. This will in turn result in “tightly defined levels of reliability, availability and serviceability … organizations will no longer be able to afford problem solvers. A new job description will arise , that of problem preventer “.
IDC also expects hardware vendors to agressively pursue software flexibility. “In order to maintain market presence and share, most vendors will aggressively integrate hardware and software”. Also software will see a new application development paradigm “where developers utilize new language tools to dramatically cut the time to develop new applications …[its called outsourcing]”.
Fiber optics, according to IDC , will be the at the heart of storage farms and remote access capabilities to “storage plantations”. In general, IDC expects Moores law for computing power will continue to apply to high speed storage devices of ever increasing capacity.
Finally, IDC sees mainframe computers as not going away – “mainframe computers will continue to provide corporate-wide access, connectivity, control and management of invaluable corporate data assets. In the increasingly competitive business climate of the 90s, mainframes will do this by answering the call for distributed processing power over worldwide networks”
The Cats out of the Bag
With the last quote with emphasis I have let the cat out of the bag. This is an IDC White Paper from 1990. It was preoccupied with the hardware world and the hegemony that IBM had over IT development and processing with mainframe processing. It was a sponsored article (remarkably similar to the current Microsoft “Get the Facts on Linux” campaign), more akin to an Armonkish aching to preserve that ascendancy. Within two years, Microsoft would turn on their IBM benefactors, sabotage OS/2 development by withdrawing from it and produce Windows 3.x to defeat OS/2 with key technology and lessons learned on the OS/2 effort. By 1995, the IBM hardware monopoly had been replaced by the Microsoft software monopoly. There is no hint of these events in the IDC white paper.
Next, IDC has a very benign look at application development – not at all catching the great war between higher priced, “-ties” delivering mainframe and Unix systems (IDC missed this too – and of course is excused from missing Linux which was yet to be created – but also missed the developing currents of Open Source with shareware, freeware, and university sponsored “come and get it” ware). The “-ties” delivered very high availability, reliability, scalability, security but at a price. Arrayed against them were the “Just Good Enoughs” with no show stoppers, GUI usability, easy installation and lowest cost from Microsoft,but also Apple and legions of 3rd party PC developers. “Just Good Enough” apps, then File Servers, then Client Servers systems along with a succession of desktop OS which were one year declared the best thing since sliced bread and then when the next version of the OS came along were dissed by the very same folks who had once praised profusely the earlier OS or other software tool/app. Indeed a little deliberate , discrete cut-off of backward compatibility was used to hurry the demise of earlier software.
IDC also missed the great CHAOS in software development as project failure rates (not delivered or delivered and not used)throughout the 90s persisted at 50% for most of the decade. Despite CASE-Computer Aided Software Engineering, ERD-Entity Relationship Diagramming tools for database planning, OO Design tools excellent early OO Analysis and Design theories finally coalescing in the late 1990s in the UML-Unified Modeling language, or the people insights of Brooks, Yourdon and others ==> application development literally foundered. Certainly missing was identifying the critical importance of managing of the rate of change, packaging development stages into manageable steps and controlling the effects of IT system change on the total organization to minimize negative feedback.
But give IDC credit, they certainly got the importance of and growth of huge storage systems and the associated data warehousing systems. They also got right the move towards self diagnostic and increasingly automated remote system control software. In fact IBM and others are just beginning to deliver self-healing, self-balancing and self-recovering systems of good promise. So IDC was ahead of the curve in its vision here.
But IDC prediction of governance by end users with bechmarked measures and service level agreements got started but never quite took hold. Perhaps this was the tacit agreement – for accepting mission impossible assignments participants werent held to performance requirements/standards . Perhaps this is implicit in the White Paper itself – for the mission impossible of defining mainframes as the dominating influence of 1990s IT, perhaps a few major trends will just get sidelined.
In general, forecasting 5-10 years out is a risky business. One needs disruptive, iconoclastic thinkers who are irreverent towards the status quo – and more responsive to some of the under currents and critical trends and their consequences. Someone who is dedicated to sleuthing, to seeing where the evidence leads unpleasant as it might be. Its easy to see how a character like Cassandra might popup in a Greek play – they were obsessed by having oracles peer into the future.