RE: This is a software design question - ILE related

Well thousands of people having the same open data path that would be added
to each program that is bound to the module is a waste of resources.  But
then if you do not have enough servers running to process all the data queue
request, the system will fall behind.  We use the data queue approach for
similar type of request and have a quick time stamp check, if on the queue
for more than xx seconds, start another server after completing the current
request.   If a server times out, we use 60 seconds, more that xx number of
times, it ends.  We use a data area to track how many are currently running,
min number to be running and max number allowed.

I prefer the data queue request idea.  Never thought about it for zip code
lookups.  We also have other files that are widely used for lookup of
values.  Since I have to re-do our current zip code stuff,  I am going to
try the data queue server approach.

Good luck and let me know what direction you go and how it works.


Christopher K. Bipes    mailto:ChrisB@Cross-Check.com
Sr. Programmer/Analyst  mailto:Chris_Bipes@Yahoo.com
CrossCheck, Inc.        http://www.cross-check.com
6119 State Farm Drive   Phone: 707 586-0551 x 1102
Rohnert Park CA  94928  Fax: 707 586-1884

If consistency is the hobgoblin of little minds, only geniuses work here.
Karen Herbelin - Readers Digest 3/2000

-----Original Message-----
From: Nathan M. Andelin [mailto:nathanma@haaga.com]
Sent: Monday, July 31, 2000 10:46 AM
To: MIDRANGE-L@midrange.com
Subject: This is a software design question - ILE related


The example I'll use is hypothetical.  But, the question applies to a real
application I'm working on.

With ILE you're encouraged to encapsulate code according to subject into
reusable components.

With that preface, here's the hypothetical:

An ERP application has separate "service programs" for encapsulating code
related to "employees", "vendors", "customers", "sales leads", etc.  One
thing common to each module is "person" related data (i.e.. name and
address).  Each separate module needs to use a "zip code" file to look-up
the "city" associated with a zip code.

For modularity, one idea is to create a single service program that opens
the zip code file, and exports a function that returns the city for any
given zip.  All other modules in the application can simply "bind" to, and
use that function.  It saves little code by not including the record
retrieval code in each module.

Another idea is to turn the "CITY LOOKUP" module into a server rather than a
service program/sub-procedure.  A set of data queues could be used to send a
zip code - receive a city (in the context of a single request).  The
application might have the following structure:

The ERP application may have thousands of concurrent users.  Which design
approach would you take, and why?  Or, would you do something else?


Thanks.


Nathan M. Andelin
+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---