If the world of computing were actually materially advancing, we would fully expect the number of machines to collapse to equal the number of data centers (perhaps plus one). So that number would typically be two (or perhaps three): one (or two) for the primary data center, and one for the alternate data center. Any more than that means there are capacity constraints and/or inefficiencies yet to be conquered.
That's *exactly* what's happening/has happened as mainframe computing continues to push the envelope furthest and farthest. The System z machines are getting so powerful, so efficient, so thoroughly virtualized, that the vast majority of customers only need "a couple" to run their entire businesses, and with the highest possible qualities of service. And IBM is quite happy with that outcome, because it represents the highest achievement in real-world computing, the maximum bang-for-the-buck. It's exactly what customers want (maximum efficiency), and it's exactly what IBM is building. Citigroup said just that in the zEnterprise announcement: they can do much more work on many fewer machines, and that's such an obvious and clear measure of business computing efficiency ("do more with less").
Yet that's exactly what's NOT happening (or at least not happening fast enough) in the rest of computing -- and that's a big problem for individual businesses and for the planet. Remember those old movies which talked about how "the computer was so big, it took up a whole room"? Well, unfortunately we've gone exactly backwards: the rooms have gotten bigger, not smaller, and now they're stuffed to the gills with lots of machines, each satisfying (if we're lucky) a tiny piece of the overall business's needs.
This is progress? Heck, no!
This article originally appeared in the IBM-MAIN forum. The views expressed are solely those of the author and do not necessarily reflect those of anyone else or of any particular company