IBM compares new z10 mainframe with 3-4 yr old Sun servers and makes their own old mainframes look slower. Now I worked with the mainframe for over 10 years, COBOL and CICS and all thoses thing were great. However, these type of comparisons devalue the mainframe. We need to compare new products from one supplier (IBM) with the equivalent new and current products from competitive suppliers, Sun etc. Quote from IT Jungle article "Last week, with the launch of the System z10 Enterprise Class mainframe, the mainframe is as close to parity with big Unix and proprietary RISC machines as it has been in a long, long time." But can IBM keep this up, the mainframe lost competitiveness about 10yrs ago and the UNIX mainframes will not stand still.

Not sure about the proprietaty claim either, SPARC CPU designs are open source and have been for over 10 yrs, see OpenSparc. Also Fujitsu use the same CPU's. So Sun has other companies using their Chips, I do not know of another company using the IBM chips or IBM CPU designs being open source. I am afraid that Sun is not proprietary but many other companies are. This is an old pre-conceived view.

Also, I thought that the industry has started to understand that CPU speed i.e. GHz is not a good indicator of performance, with all the claims that IBM is making I would like to say that the equivalent amount of T5220 (8 core, 64 threads with 1.4GHz) servers that take up the same floor space, power and cooling as a IBM z10 will be able to can consolidate more smaller systems while maintaining performance and improving availability and management than a z10. Solaris has in-built virtualisation with LDOMS and containers that is free and open source. Now what would the IBM z/OS costs for this be, people really need to look at this area?

Strangely I just read in this article that IBM has caught up with large UNIX mainframes with their z/OS mainframe. However, this article got some very simple technical details wrong. Large Sun Enterprise Servers (which have all the features of mainframes and more) e.g. the M8000 and M9000 series use up 64 x 2.1GHz and/or 2.4GHz dual core CPU's. You can have up to 64 of these dual core CPU's in one system. Now what we can do in these Sun UNIX mainframes is mix CPU speeds, so if by any chance we have a new faster CPU becoming available in the next 6 months, you can add that to the system while it is running. Now that is innovation and investment protection.

So this claim from the article; "That extra clock speed is important since the 4.4 GHz clocks IBM can now deliver in the z10 give it an edge over the 3 GHz or so clock speeds of X64 alternatives and the 1.6 GHz to 1.8 GHz clock speeds of the Itanium and UltraSparc-IV+ processors used by Hewlett-Packard and Sun Microsystems in their big iron Unix boxes." Is just becoming less relevant.

So IBM is comparing a new z10 system with some old Sun System. These M-series large UNIX mainframes have been available for 10 months now.

As a colleague Jeff Savit wrote in a very good article, IBM is comparing their new products with competitors old products a very sneaky trick and not very honest. Thus this shows that we have to be very cautious about any of the claims in this announcement. This article is a must read for anyone who wants to understand the IBM z10 systems claims in comparison to Unix mainframes.

There is another very strange performance comparison going on here, so that the new z10 can look better, This article implies that the old systems are now specified as being a lot slower than they really were. From the same article here is the quote. "IBM's documents for the z10 say that the fastest z6 core is expected to deliver 62 percent more performance than the fastest z9 core, which had a much lower clock speed. (That would seem to suggest that the 1.7 GHz z9 processor core running full out should have been rated at 568 MIPS, which is lower than the numbers a lot of people have been throwing around for three years.)" So not only does IBM not compare latest z10 mainframe server with old Sun servers, but they compare and reduce the performance numbers of the older mainframes. If I was a IBM z/OS customer I would ask for a refund on the old system software licences as the MIPS value has just changed, since the old z-series was not as fast according to the MIPS indicator as IBM said it was.

Summary, the truth is coming out out slowly, so buyer beware, look at these mainframe claims very, very closely and be suspicious.

I have the greatest respect for IBM, but there are just too many strange claims being made here. After all IBM together with Toshiba (I think) made the cell processor that we have on the Playstation 3. Now that IBM cell chip is really interesting. I cannot say the same about the z10 mainframe with all thes strange comparisons going on.