isham research
A personal note
And with a heavy heart. As those bored enough to have read my biography will know, I owe a great deal to IBM's mainframe architecture. Working with IBM, and against it, and with it again - without it ever paying me a penny - I sustained a career and fed a family for close to four decades.
This note was immediately triggered by a story from South Korea, but there have been others.
A few years ago, one of the good IBMers dropped something very heavy on my famously sandalled foot and - while I was hopping around in agony - asked me what I thought critical mass was for the mainframe. In other words, what was the point at which the number of customers/installations became so small that investment by IBM in future developments would unsustainable. The thought had never occured to me and through the waves of pain I was unable to speak - but after a few beers (and one or two Southern Comforts) that evening I came up with a number for him the next day.
Sadly, that number is history. Around four-fifths of IBM's mainframe revenue now comes from 200-300 customers.
So the point comes at which further investment in the platform should be questioned. zSeries remains the most robust and secure corporate computing platform available. Those who run major corporations on it should not worry - adding a few lines or a couple of processors to an existing massive CICS/DB2 environment will not make it any harder to migrate when the time comes. No criticism of IBM is implied, either - for many the alternative platform will have an IBM badge; pSeries, iSeries, xSeries enjoy the same support that zSeries did.
But the crunch comes with wholly new applications - those that have no immediate synergy with existing ones.
In such cases, a mainframe decision must carry with it the implied cost of migration within the medium term.
Phil Payne
isham research
Committed mainframer
7 September 2006