RE: Fastest access: user space, data area, data queue

[qsrvbas@xxxxxxxxxxxx]

Albert:

One thing to keep in mind about objects such as user spaces, user indexes and 
data queues, is that some of the normal checks aren't done. Locks, for example, 
aren't verified.

One job can ALCOBJ *EXCL, but another job can still access the object and 
update it without problem. Only when the second job actually executes ALCOBJ 
(or otherwise explicitly requests to establish or test for a lock) does that 
part of object access run for those objects.

That kind of processing is overhead when you consider an object such as a data 
area. Lock testing isn't so important when you're dealing with your own objects 
in QTEMP for example. IBM basically lets you run a little faster with a user 
space, but the risk goes up a little. It's up to YOU to make sure your access 
is appropriate.

In short, it means that accessing a user space can be faster than accessing a 
data area. And accessing a data queue for a message can be faster than 
accessing a normal message queue. Whether it's a meaningful difference is up to 
you.

Tom Liotta

midrange-l-request@xxxxxxxxxxxx wrote:

>   7. Fastest access: user space, data area, data queue (Albert York)
>
> I was wondering is there is any kind of significant difference in speed when 
> accessing a user space, data area, or data queue entry. 
>
> Thanks,
>
> Albert