Leap forward hits snags
THE on-going migration to client/server platforms is the most significant architectural shift the industry has seen since the adoption of IBM's 360 mainframe systems starting in the early 1960s, according to a report by management consultants McKinsey and Company.
An analysis of the financial performance of computer vendors confirms the trend but, the report says, technological complexities, higher than expected costs and shortages of key expertise are slowing the process.
McKinsey's 1993 Report on the Computer Industry says most companies are delaying major implementation decisions on client/ server architecture, while still testing it in non-mission critical applications.
Few transaction processing or other mission-critical applications are taken beyond the pilot stage, it says.
McKinsey cites four principle obstacles that have delayed companies' migration to the long-touted client/server architectures.
First, the cost of switching from existing 'legacy' systems has proved to be more expensive than had been anticipated. The short-term cost is extremely expensive, because existing software and hardware need to be scrapped.
The report says while a legacy system still performs adequately, the overall expense of converting the system negates the primary motivation for implementing client/server architecture - the need to cut costs.
Furthermore, even client/server pioneers have been surprised at hidden costs involved in getting clients and server to work together.
Secondly, the industry still lacks key applications on client/server platforms, and effective support tools for the new architectures. Customers running mission-critical applications remain doubtful of the overall system reliability of client-server, andsuspicious of security features.
Thirdly, skills necessary to implement client/server systems are in short supply, on both the vendors' support side as well as the customer side.
Finally, the existing distribution channels are inappropriate for open systems, and are largely an extension of channels used either for legacy systems or for standalone PCs.
For example, a vendor's direct sales force is unlikely to recommend the kind of ''best of breed'' solutions that open architectures allow. Systems integrators, McKinsey says, are well suited to large-scale application but tend to be overkill for incremental projects, a common characteristic of client/server projects. Traditional PC sales channels simply lack expertise.
Despite these obstacles, McKinsey emphasises that it sees no ''retreat'' from the client/server direction.
Four out of five companies McKinsey spoke to are headed in the client/server strategic direction, with most of them saying their efforts should lead to a fundamental change over the next five to 10 years.
The McKinsey report, released last week, is global in perspective and its research found that the Asia-Pacific region is the only computer market in the world that continues to grow at double digit rates.
The report, compiled from data on 152 companies, says demand for computer products and services in the Asia-Pacific has grown at an average of 14 per cent compounded annually from 1987 to 1992.
