IBM ironically appears to be following in the footsteps of arch-rival HP by consolidating onto a single line of microprocessors for all of their proprietary servers. This substantially reduces the absolute amount of R&D necessary to support IBM’s three proprietary server lines, as well as increasing the number of systems and MPUs that will amortize the R&D costs. iSeries customers have already reaped the rather substantial benefits of sharing hardware with the performance oriented pSeries systems. Now, IBM is taking the next step and extending the benefits to zSeries users.
While much about the POWER6 is unknown, it is possible to make intelligent performance estimates based on a combination of the known micro-architectural features and performance of existing MPUs for relatively straight forward benchmarks, such as SPECcpu2000. If IBM can execute as well as they have with the POWER4/5, the POWER6 should certainly be competitive when it arrives.
The migration of IBM mainframes to PPC MPUs is a historical landmark for the industry. For a long time, pundits and IBM employees have maintained that the zSeries was the one low-volume architecture capable of being self-sustaining, by virtue of the incredibly high margins, the large associated consulting and software business and the unwavering IBM commitment to binary compatibility. However, while the architecture may have been self-sustaining, this most recent move will certainly enhance the profitability of the mainframe by substantially reducing hardware costs. Strategically such a move will also strengthen the PPC architecture as the profits from mainframes could be used to subsidize more R&D. It is also interesting to ponder the fate of Linux on the mainframe, as Linux is fully supported on PowerPC systems. Is the Integrated Facility for Linux truly needed any more, what will the advantages over a pSeries system be? That is just one of many questions which IBM will no doubt answer in the coming year.
I would like to anonymously thank everyone who supported me on this article. Without your help, I could not have done it.
Copyright 2005 David Kanter. All rights reserved. No portion of this article, in part or whole, may be reproduced, copied, transmitted, stored, downloaded, in any manner in anyway for any purpose whatsoever without the express written consent from the author.