We have hundreds of BMS map load modules that were last assembled and linked years ago when we still used RMODE(24)...
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
and AMODE(24) for everything. We still use these maps every day, but we shifted all of our COBOL code above the line.
Is there any performance benefit to re-linking these maps and moving them above the line, too? For example, might it decrease below the line memory contention?
Like everything else, I'm afraid that the answer is that it depends; which is not terribly helpful.
In your case, you seem to be running quite happily and not too bothered about the below-the-line storage consumed by these maps. There is a fairly strong argument along the lines that if you are not constrained by 24bit storage, why worry about changing these maps: after all, there are lots of them!
The other side of the argument is that these maps, although small, will each occupy about 1 page of below-the-line storage and 100s of 4k pages soon adds up. This eats up the available storage, and could cause problems in the future. In general, below-the-line is bad, so get rid of them.
What do your change control chaps/chapesses think about redoing the maps? Is this a trivial change or is it complex? Would everything have to be re-tested and planned, or would things be more relaxed?
I think I would be inclined to convert high use and big maps provided that change control was not onerous in your existing CICS regions/releases. If maintenance is changing the applications which use the maps, I'd probably insist that the relevant maps were moved above the line as part of this maintenance.
Nevertheless, if you are going to generate the recompiles automatically (I visualize a little Rexx Exec retrieving the names of the mapsets and then generating jobs to do the recompilations/relinks), it is probably easier to redo the whole lot.
Personally, and this is more of a Sysproggy approach than my Strategy persona should permit, I'd be very inclined to put off this conversion until you install a new release of CICS (plug for that nice CICS TS 2.2) and recompile all your maps to include the ADS extensions as well as moving above the line. These extensions will permit you to use the CICS Bridge facilities which are a way of emulating applications within CICS (you run a transaction at a 'Bridge Terminal' and manipulate things at a BMS field-level). You might never want to use these functions, but if you are redoing all the Maps it is sensible to include this info whilst you are at it - merely to prevent a future mass-recompile. This approach has the benefit of avoiding the change-control procedures as , no-doubt, all validation of these recompiled maps would be subsumed as part of your validation of the new release of CICS.
All of which is not terribly satisfactory: all you can do is talk to yourself in a Socratic dialog. I've given you the Thesis and Anti-thesis, the Synthesis is up to you!
Dig Deeper on IBM system z and mainframe systems
Related Q&A from Robert Crawford
For better mainframe capacity planning, how do I convert CPU hours to MIPS? And is there a way to calculate the relationship between MIPS and MSUs?continue reading
I have two years of experience in mainframe technology, currently working as a mainframe developer. I want to change to Java technology.continue reading
I want to replicate DB2 from the mainframe to an AIX box since it's cheaper and the copy can be used for testing. Is this possible?continue reading
Have a question for an expert?
Please add a title for your question
Get answers from a TechTarget expert on whatever's puzzling you.