I/O Charge Confusion

dick scherrer · Posted: Mon Jun 10, 2013 7:32 pm

Howdy,

One of my clients is on a system where they receive their Mainframe "bill" every month and they have asked that i look into the major difference in i/o rates.

Since sometime 3rd quarter '12, NO physical tapes have been mounted. During '11 - '12 nearly all physical tape was converted to virtual tape (v-tape).

The rate for "tape" i/o is about 16 times the dasd rate. Any idea why?

Also, the rate for v-tape is the same as physical tape. Does this seem proper? As far as i know, v-tape is all dasd even though there are reserved UCBs for each virtual device and these volumes can be (almost always are) managed by the Tape Management System.

Any experience / thoughts most welcome.

d

enrico-sorichetti · Posted: Mon Jun 10, 2013 9:43 pm

just guessing ...
to discourage tape usage ?
even if probably for <virtualized> tapes it does not make sense any longer.

Pete Wilson · Posted: Thu Jun 13, 2013 7:44 pm

I used to work for a place 15 years or so ago that charged heavily for Tape usage to discourage it. That was pre-virtual when all the tape handling was expensive and a real hassle. As Enrico says, with v-tape that essentially no longer applies.

I wonder if it's anything to do with the large block-size support. DFDSS uses 256k blocks on a DUMP for example, so maybe uses more virtual storage?

dick scherrer · Posted: Thu Jun 13, 2013 8:39 pm

Howdy,

Thanks for the replies - 'Preciate it!

Seems like i've stirred up a bit of a hornet's net. Somehow "nobody" realized tape was ~16 times more expensive to read/write - nor was anyone aware that v-tape was being billed at the physical tape rate.

Sometime "soon" the company that does the rate calculations is to review the current rates and come up with new rates.

This may be good or bad. . . The service has to generate enough income to offset costs and if we gain one place, we'll surely lose in another(s).

Our goal is overall cost reduction and until the "new policy" is determined, we'll probably not introduce any media changes. . .

Now, if we could just find a few Tbytes that could just go away . . .

Robert Sample · Posted: Thu Jun 13, 2013 9:11 pm

dick scherrer · Posted: Thu Jun 13, 2013 9:57 pm

Hi Robert,

What i m seeing is that there is a Lot of data that could "go away" - the Screamers are the ones who believe forever-to-date is an acceptable retention specification . . .

d

Robert Sample · Posted: Thu Jun 13, 2013 10:33 pm

And the ones who scream the loudest are the ones with the most retained data (and the ones who never go back to look at it, either) ...

Bill Woodger · Posted: Thu Jun 13, 2013 10:58 pm

Punitive charges for data-retention beyond statutory/regulatory requirements.

I had an Auditor once who insisted everything was kept as long as it was needed (obvious), and not one day longer (no so obvious). I think it was a good idea. If no-one has the "right" to require the data to be presented, then don't keep it. And certainly don't keep "part" of it.

Yes, there was data for "trends" and "historical evaluation", but that was all extracted and held as "current data", and used.

dick scherrer · Posted: Thu Jun 13, 2013 11:32 pm

One of my better senior IT managers (in a ddifferent life, it seems) refused turnover of anything that involved new data that did not have specific purge criteria. And that retention had to be justified.

d