Ted MacNEIL wrote:
I think you need to expand that question to:
1. what does it cost to do a VSAM I/O to real disk (or cache in your case,
since the hit ratio is almost 100%)?
The same as the cost going to disk -- it's just faster.
2. what does it cost to do a VSAM I/O but finds a hit in a memory buffer?
The same as going to cache -- it's just faster.
This is counter to my experience. I have seen many occasions where an
increase in VSAM buffers for poorly tuned programs with over 100K EXCPs
reduced the physical I/Os, elapsed time, AND CPU time by at least an
order of magnitude. My impression is that getting a VSAM record from a
CI that is found in buffers within a z/OS address space is a minor cost
compared to the overhead of initiating a physical I/O operation to read
a missing CI. -- but if you can replace a VSAM get from buffer by
something much faster - say, an in-memory hash table lookup - you may
reduce the cost by yet another order of magnitude.
Erratic application response is much more likely to hit applications
that are I/O intensive. In such cases the cost of the I/O is not just
the CPU overhead associated with the I/O, but should also include the
cost of the response delays and productivity loss to the end user caused
by I/O contention (unfortunately hard to quantify accurately) --
perhaps also include all the cost of man-hours consumed responding to
performance complaints from the end users.
One obviously gets the best of both worlds and the greatest savings if
by some analysis and application redesign the need to read most of the
VSAM records can simply be eliminated.
3. what does it cost to do a VSAM I/O but finds a hit in memory, e.g. CMDT?
This was my recommendation!
If this region uses less than 5% of cpu, and the end-users do not complain
about R/T, then do you really have problem?
Ah, there's the key.
My original post stated that response was occasional erratic, and users were
complaining.
I am trying to fix it, but management wants a cost justification.
Hence, the question on cost for an I/O.
Since we are in an out-sourced arrangement, MIPS have a real cost.
So, does anybody have an answer?
Or, do we just continue to dance?
-
-teD
O-KAY! BLUE! JAYS!
Let's PLAY! BALL!
--
Joel C. Ewing, Fort Smith, AR [EMAIL PROTECTED]
----------------------------------------------------------------------
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
