7 Apr
2013
7 Apr
'13
9:11 p.m.
On 7-apr-2013 18:22, Guy Sotomayor wrote:
There is further irony in any resistance to the use of
mainframes,
because everyone in IT is talking about virtualisation and utility
computing (see Introducing Utility Computing, Atos Origin, October
2006). But both concepts have been around in the mainframe world for
30 years or more. By providing virtualisation, VMware is doing for
Intel platforms what VM/370 did onmainframes in the 1970?s.
Mainframe is a utility service, and has almostalways been so. Many
advocates of utility services are actually ?reinventingthe mainframe?
in many aspects of what they seek to achieve.
So, why is IBM allowing this to happen? Why aren't they there to
give VMware and others hell, perhaps even supply a bit more economical
solution for the /little people/? Why does IBM assume that people
will have the telepathic ability to learn about IBM and its products?
For a supposedly successful platform, with no problems at all (which,
by the way, that earlier Inquirer article of late 2012 doesn't seem
to support), it sure seems relegated to obscurity and extreme niches
at best.
For a supposedly successful platform, why isn't IBM interested in
making it even more successful and getting more people to use and
perhaps switch to it? Why is IBM building "x" (and "i" and
"p" for
that matter), why not /just/ "z"?
Another interesting statistic is average capacity
utilization:
Wintel-based servers: 8-15%
Unix/RISC: 28-45%
Mainframes: 65-75%
Where did you get these figures from?
It's also noted that mainframes (zSeries) can and
do run at 100%
utilization for long periods quite happily. Try that with a non-
mainframe!
100% utilization, thus with no headroom whatsoever; that's a good
idea in general...? (Regardless of whether a specific platform
would subdue or become less stable, or whatever.)
- MG