In an industry where backward compatibility over even four years is considered to be good, it is remarkable that IBM has achieved compatibility stretching over 10 times that period.
Today`s IBM mainframes are quite different to the original Series/360 introduced in 1964, but Douglas Neilson, IBM UK systems consultant who has 30 years` mainframe experience, says software developed to run on the original systems would still run on IBM`s latest z-series. That must be something of a record in the computer industry.
On the 40th anniversary of the introduction of the Series/360, I was curious to know how the IBM mainframe had survived so long in the face of a proliferation of competing technologies. There appears to be three main reasons.
First, when the demise of the mainframe was predicted in 1992, IBM switched to a Complementary Metal Oxide Semiconductor base to make mainframes smaller and cooler. Second, IBM recognised that changes in software demanded that the mainframe be opened up to all the new application models that had emerged. Third, IBM remained committed to the mainframe where most other players chose to leave the market.
Jan Dry, Sun regional solutions and technology manager, says IBM`s strategies drove down the cost of mainframe computing capacity to commodity level, driving their competition out of the market. He says IBM hung on despite lower margins until the other players chose to leave.
So it would seem IBM took a gamble on mainframes and that gamble paid off. But what about the future? As IBM celebrates the past, it seems to be equally concerned with the future.
IBM has adopted a policy of putting the best server technology it can develop first into the mainframes because they are where the big customers are, before cascading it down to the other families of computers. Neilson says the mainframe should be seen as the fountain of technology that permeates the whole IBM range.
Hi-tech T-Rex
As part of the strategy for the future, IBM has just released a new mainframe family, the z-890, for smaller customers. The z-890 is a derivative of the z-990, code-named T-Rex, which was released a year ago. Neilson says code-naming mainframes in development after dinosaurs is a post-modern joke in response to people referring to mainframes as dinosaurs.
The most interesting thing about the z-890 is that it is the first of IBM`s mainframes to have dedicated processors for Java. I see this as yet another strategy by IBM to keep the mainframe relevant in a changing technological environment, but will this and other similar strategies be enough to sustain the mainframe?
Neilson is confident the mainframe has a future, and Sun`s Dry agrees with him up to a point. While sharing the view that mainframes will be around for some time to come, Dry disagrees that there is potential for growing the market.
IBM recorded 200 new mainframe customers in the past two years, and would have us believe this constitutes a growing market. However, Dry says this number is not all that significant considering it is a worldwide figure and over a two-year period.
Over and above strategic planning, Neilson admits IBM has also been lucky with the coincidence of market demand with mainframe capability. He says the current e-business demands of performance, scalability, security and availability are all classic strengths of the mainframe and therefore the mainframe is assured of a future.
Competing in the high-end
Dry, however, brings an interesting outside perspective. He agrees that historically the mainframe has been extremely secure, but adds that mainframes are no longer alone in that space. He says, for example, that Sun`s Trusted Solaris range is found in many military installations where the highest level of security is required.
Dry also points out that mainframes are not alone in terms of performance, scalability, or availability when compared with mainstream high-end computers. Java optimisation and encryption capabilities are not exclusive to the mainframe.
The mainframe certainly has an interesting history, but an innovative, successful and interesting past is no guarantee of immortality.
Warwick Ashford, technology editor, ITWeb
From the Sun perspective, one of the unique characteristics of the mainframe is the lack of choice. By winning the mainframe space, IBM effectively reduced the market to a single vendor market. One supplier, one operating system, no negotiating power for the customer.
It is Dry`s belief that the mainframe market cannot and will not grow significantly. It will merely retain its user base of clearly defined customers for whom it is too costly to move to alternatives. Dry says it is unlikely that many greenfield customers will be attracted to a single supplier environment.
A persuasive argument in support of Dry`s position is that independent software vendors traditionally drive support for platforms and their investment is most likely to be in the platforms falling in the Gartner magic quadrant such as Windows, AIX, Solaris, Unix and increasingly Linux.
The mainframe certainly has an interesting history, but an innovative, successful and interesting past is no guarantee of immortality. Can these "dinosaurs" possibly still grow and live on forever, or has their time finally begun to run out? As with most imponderables in life, only time will tell.
Share