IBM prepares for 40th anniversary of 360 mainframe with good news

Rethink IT, April, 2004 by Barry Graham

MARKET OVERVIEW

As discussed in recent editions, IBM is expected to make a major series of mainframe announcements this month on the 40th anniversary of the original IBM 360 mainframe launch.

The new developments will include a z800 replacement, a Java processor for the z990 and z800 replacement system and major software enhancements (particularly in the software cost area). The latter software cost improvements and the z800 replacement will be aimed very much at the alternative platforms and attempt once more to emphasise the cost advantage of the mainframe.

Such a marketing message will be very believable as, in the fourth quarter of 2003, worldwide server revenues for systems costing over $250,000 (according to IDC) were as follows:

                                                          Revenues ($m)

IBM mainframes                                                    1,800
IBM other                                                         1,030
HP                                                                  906
Sun                                                                 420
Other (Unisys, Siemens, Fujitsu, Bull etc)                          744
Total                                                             4,900

In terms of market share this made IBM the clear leader with a market share of 58% and their mainframe business alone equating to almost 40% of the total server market for systems costing over $250,000. This compares to as little as 10% in the worst days of the mainframes history during the late 1990s.

                                                           Market share

IBM mainframes                                                      37%
IBM other                                                           21%
IBM total                                                           58%
HP                                                                  18%
Sun                                                                  9%
Other (Unisys, Siemens, Fujitsu, Bull etc)                          15%

IBM is also growing market share at the expense of its competition as can be clearly seen by the year to year growth figures for the fourth quarter of 2003.

                                               Annual percentage change

IBM total                                                           30%
HP                                                                   9%
Sun                                                                -16%

Consequently the IBM theme for the April announcements can confidently be predicted as one of 'the mainframe is continuing to evolve and its role is increasing for larger users worldwide'.

This will force users once more to compare the relative benefits and costs of the alternative platforms. To help in this process this and the following edition of the monthly bulletin will review the relative costs of the different platforms, discuss the benefits of the mainframe and review some of the contradictory information being provided by IBM's competition.

THE DINOSAUR MYTH

With all of the expected mainframe announcements, and a major push again from IBM to extol the virtues of the mainframe, we can expect many of our subscribers to once more revisit the platform choice issue. To help them in this task we have below updated our 2002 Dinosaur Myth paper which concentrates on the TCOs of the different platforms.

We have always maintained that in comparing the TCO of the platforms what really matters is the number of users, performing whatever functions are necessary to the organization, that a computer can support, with a reasonable level of service. Therefore, the key yardstick of a computer's cost effectiveness is the TCU (total cost per user) rather than the TCO, measured over a reasonable time-span to eliminate any high up front rather than long term costs--say five years.

This is significant as a mainframe can support many concurrent users performing different tasks and changing the tasks being performed on a real time basis. To help them achieve this multiple application capability, mainframes have evolved hardware mechanisms for the efficient sharing of resources among large numbers of concurrent users. They also have very sophisticated software resource management capabilities, which allow the users to have their work completed on a priority basis such that even when fully loaded the key applications get the capacity needed to perform the task in hand. These resources can be reallocated literally second by second to achieve this goal. Unix systems and PCs do not have such sophisticated mechanisms, although front some vendors claims users could be forgiven for believing differently.

This ability for all work to be completed on one system is crucial to minimizing the TCU. For example, let's say 10 applications are each used by all of the staff in an organization. In this case, if each application required up to 10Mips of capacity, but in total no more than, say, 20Mips were needed at peak load, then on a mainframe 20Mips would suffice, whereas in the Unix or PC case a minimum of 10 systems of 10Mips each would be needed, assuming they could run effectively at 100% utilization--five times more capacity in total than on the mainframe.