We are coming up for another big mainframe anniversary (the IBM System/360 was announced on April 7, 1964; although, as usual, the next release, the 370, actually got it right). But I do wish the mainframe would finally go away - the Death of the Mainframe has been announced regularly and prematurely since the 1990s. And it is still here and running many of the world's most important IT systems. [Shome mishtake, shurely? Ed]
Well, the mistake is probably being made by people who think that just because distributed server industry is trying to catch up with what the 370 mainframe could provide in the way of virtualisation, multi-processor and secure computing 30 or 40 years ago, a modern mainframe has as clunky a user interface (and as exploitative software licensing) as the last-century 370 was famous for. Not true, but readers should research that for themselves, don't just believe what I say. Things have only got better since the 20 year anniversary of Stewart Alsop's announcement of the death of the mainframe.
I was at the CA Expo customer event last week and attended the excellent mainframe track. The enthusiasm of the audience was palpable. The enthusiasm of the CA Technologies people describing tools such as CA Chorus and CA Applogic for System z and how they bought the mainframe into general enterprise IT automation as a first-class player was palpable. In fact, the keynote by Michael Madden (General Manager Mainframe, CA Technologies) on the evolving well-governed business-led automation that's a feature of mainframe environment would have been almost as useful for people working entirely in the distributed space. The modern System z is "just" an enterprise server with the sort of rock-solid virtualisation, security, multiprogramming, performance and utilisation levels other kinds of server aspire to - and excellent support for disruptive innovations such as Big Data. The cloud integration that CA Technologies is delivering is helping to make System z simply another part of the emerging next-generation enterprise IT infrastructure.
Nevertheless, CA Technologies still has some room for improvement, in two places at least. First, it was a pity that the mainframe track was off in a silo by itself. I am sure that CA Technologies does believe in the future of the mainframe, so why not mark it an integrated part of the full event?. And, second, customers from the floor told stories of putting forward a conservative business case for, say, Linux on System z and having it rejected by the CIO in favour of a more expensive, less effective distributed solution. Why? Well, perhaps because CIO's are scared of lock-in to a solution with just one hardware supplier; but perhaps also because they associate the word "mainframe" with the mainframe of 30+ years ago, because they are frightened of a technology they aren't familiar with; and. most importantly, because no-one has explained, effectively enough, the relevancy of System z for their applications (and perhaps IBM could be more help here too).
In my governance role I hate waste and, frankly, not exploiting something like the modern System z looks like waste to me. Perhaps the vendors of System z technology and software could help by marketing more directly at CIOs who aren't already System z customers - and by killing off that pesky "mainframe" title.