View previous topic :: View next topic
|
Author |
Message |
Amit Manas Dubey Currently Banned New User
Joined: 15 Dec 2006 Posts: 22 Location: Mumbai
|
|
|
|
Hi All,
I have a VSAM file LRECL = 6500, Key length = 22, No of records in the file = 1064949.
I need to know what impact would it have on I/O (Random read) for 1 single record, if we increase the Record length by 600 bytes ?
Will it have any difference at all, and if yes, approx how much ?
Best regards,
Amit. |
|
Back to top |
|
|
stodolas
Active Member
Joined: 13 Jun 2007 Posts: 632 Location: Wisconsin
|
|
|
|
Of course there is some impact. However, it is probably somewhere between unmeasurable and minimal. |
|
Back to top |
|
|
dick scherrer
Moderator Emeritus
Joined: 23 Nov 2006 Posts: 19244 Location: Inside the Matrix
|
|
|
|
Hello,
Quote: |
it is probably somewhere between unmeasurable and minimal |
Yup, i'd go with this. . .
Given that the index structure has not changed and that the number of index entries has not changed, the time to locate the specific index entry should stay the same as it is now. Reading one 6.5k record versus one 7k record should be "between unmeasurable and minimal". |
|
Back to top |
|
|
Bill O'Boyle
CICS Moderator
Joined: 14 Jan 2008 Posts: 2501 Location: Atlanta, Georgia, USA
|
|
|
|
An optimal CISIZE, for both Data and Index will also have a significant affect on performance as well as the AMORG parameters utilized and BLSR.
If you're truly issuing RANDOM Reads, then ensure that your program specifies ACCESS IS RANDOM as opposed to ACCESS IS DYNAMIC.
HTH....
Regards,
Bill |
|
Back to top |
|
|
|