-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On
Behalf Of wtrovijo
Sent: Tuesday, July 03, 2007 6:32 AM
To: IBM-MAIN@BAMA.UA.EDU
Subject: Re: Another one bites the dust

> > How many "Administrators" are they hiring to replace you?
> 
> I'm still here, just doing different stuff (SAN management and 
> high-performance computing at the moment).  The new administrative 
> system is such a huge boondoggle that it's hard to tell how many 
> people are involved in providing the kind of support formerly provided

> by a single sysprog.
> --
> Matt Simpson --  z/OS Support
> 

Based on Matt's numbers I was trying to figure out how big - and "cheap"
- will be ours. Our current 2tb, 5k tablespaces , 45k getpage/second DB2
zOS application will be replaced by sap in the next 2 years...
<SNIP>

Based on my experience of converting from MCP[1] to See-Bull (or Sieble
as you prefer), a VSAM master file having records of 4K (with areas
multiply defined) turned into a 1MB record. Now, the WHOLE MCP install
(all programs, data, page, spool, mirror-ed DASD, etc.) all fit neatly
into 3 TB. Moving ONE, count-em, ONE application required a 5TB RAID
unit, various and sundry servers, PLUS 4 full blown, "got it all"
REGATTAs (each was physically larger than the z/800 1b0). Each REGATTA
produced MORE BTUs than the z/800 (to the point that if you were cold in
the computer room you could walk behind one of them to actually get
warm!).

YMMV. 

But I was absolutely dumbfounded at the amount of money thrown at moving
off a mainframe based system. And the end result was, the CICS based
application that was being replaced had SUBSECOND response while the
See-Bull system had, well, 5-15 second response. So let me put this in
perspective: a customer calls to order an item (whatever it might be).
The call handler will need to go through 7-11 screens (CICS) for a total
time of 5-6 minutes to handle the call. The See-Bull system required 5
screens (browser based) and took 11 minutes.

Again YMMV.

And this was sent to Carrico for posting to ReBoot Hill (why? Because it
was, in my estimation, a US$100M fiasco that resulted in the See-Bull
system being abandoned when they realized that the computer room would
have to be expanded for the third time to handle all the hardware...)

[1] Prior post -- MCP is my name for z/OS. It is The Master Control
Program vis-a-vis IBM revenue for SCPs on z/Architecture, right?

Regards,
Steve Thompson

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Reply via email to