|
View previous topic :: View next topic
|
| Author |
Message |
awyn
New User
Joined: 11 Feb 2008
Posts: 49
Location: chennai
|
|
|
|
hi ,
can u tell me how to do mainframe performance testing and what are the different aspects should be covered to do mainframe performnace testing.. |
|
| Back to top |
|
 |
expat
Global Moderator
Joined: 14 Mar 2007
Posts: 8797
Location: Welsh Wales
|
|
|
|
| What are you looking to find out ? |
|
| Back to top |
|
|
awyn
New User
Joined: 11 Feb 2008
Posts: 49
Location: chennai
|
|
|
|
| i want to know in what aspects the performance of mainframe is considered.and how the different different teting like volume, stress all should be considred for mainframe jobs |
|
| Back to top |
|
|
vasanthkumarhb
Active User
Joined: 06 Sep 2007
Posts: 275
Location: Bang,iflex
|
|
|
|
Hi,
Mainframe includes technologies such as COBOL, VSAM, DB2, JCL etc
What exactly u need to know in which area???? |
|
| Back to top |
|
|
dick scherrer
Moderator Emeritus
Joined: 23 Nov 2006
Posts: 19244
Location: Inside the Matrix
|
|
|
|
Hello,
Is this a school assignmrnt or is this for some part of your job with some company?
Your answer to that will generate different kiinds of answers. |
|
| Back to top |
|
|
Pedro
Global Moderator
Joined: 01 Sep 2006
Posts: 2547
Location: Silicon Valley
|
|
|
|
Sorry, I am not a performance management professional *...
But I recall that there are products that simulate a workload over a long period of time. The workload is a variety of read / update requests. The objective is to figure out how much of something can be done and to get the average.
And then to be able to have the same workload to test various configurations or various hardware.
* I think performance testing is a career unto itself in larger software development companies. It can be somewhat complex to setup and to measure. |
|
| Back to top |
|
|
dick scherrer
Moderator Emeritus
Joined: 23 Nov 2006
Posts: 19244
Location: Inside the Matrix
|
|
|
|
Hello,
| Quote: |
| performance testing is a career unto itself in larger software development companies |
It is also in some organizations with very large environments. Between the cost of bigger faster hardware and the licensing practice known as "tiered pricing" (the license costs more when run on a more powerful system) very large $ savings may be realized by having good "local" performance measurement (and tuning).
The current tendancy to just throw slop together is making this more important. I've been asked to help with several "long-running" batch processes to see if they could be improved. Looking at the jobstreams, i found multiple executions of the sort product and other utilities that passed 100s of millions of records multiple times just to prevent writing one application program.
The thought process went something like "if i had a file with... i could. . .'. This was repeated a few times slicing the requirement into many unneeded passes of the data. The "reason"s to bypass coding a solution were many - 2 of which were:
. that "jcl" could be promoted very quickly while "programs" could not.
. the developer was completely unqualified as a programmer and was unable to write the code. |
|
| Back to top |
|
|
|
|
