Is there anyone out there who uses ODBC on a large scale (several hundred users or more) to update a Progress database? Especially a Progress database that also supports a traditional 4GL application.
In my experience ODBC is used almost exclusively to support reporting tools. I don't know of anyone who uses it more than occasionally and accidentally to update their Progress database. I'd be interested in hearing from anyone who does though.
While it isn't what you are asking for, one experience I have had related to this is the use of an OpenEdge database as a repoository for Enterprise Architecture. User counts are low, often one person, but the intensity of ODBC usage is extremely high ... and still problematic as of my last testing, although I have yet to test 10.2A with EA 7.5.
tamhas schrieb:
While it isn't what you are asking for, one experience I have had related to this is the use of an OpenEdge database as a repoository for Enterprise Architecture. User counts are low, often one person, but the intensity of ODBC usage is extremely high ... and still problematic as of my last testing, although I have yet to test 10.2A with EA 7.5.
Can you explain what you mean with problematic?
My experience with this particular combination is pretty o.k. (OE 10.1B and EA 7.1/ OE10.2A and EA 7.5), 3 users, rarely concurrent.
Performance for both reading and writing to a database on a server on the LAN are o.k., a little slower than working with a local EAP file (which is a MS Access database file). VPN access via ODBC is surprisingly fast. The delay of the ODBC VPN access compared to the delay I've previously experienced when using Progress applications in VPN client/server mode is ignorable.
Hey! Don't hijack my thread you two or I'll have to mark your responses "unhelpful"! (And you will lose 5 points each...)
I want to know about large scale ODBC use for updates. Especially in situations where the database is also being used by a 4GL based application.
I want to know about large scale ODBC use for updates. Especially in situations where the database is also being used by a 4GL based application.
One of the beauties of using an OpenEdge database as the modell repository is that you are able to read (report) and write (create elements, diagrams) to the UML modell from ABL. So Dr. Thomas and I used ODBC for read/write access to a dabase we also updated from ABL...
But I don't do anymore. I have switched to EA API access, but that's truely a topic for another thread...
Mike, over and out
Well, as noted, I cited my experience because the volume of transactions is very high, even if the user count isn't. I don't know if the performance issues I experienced relate to questionable SQL or the intensity of the traffic, so I can't guess how that might scale out to many users doing smaller transactions, but I would guess it didn't apply unless the transactions themselves were large.
Frankly, I wouldn't expect you to get much of a relevant response since I wonder why someone would have a large number of SQL clients. I can see having a lot of .NET or Java clients, but I would think most of those people were using OpenClient, not SQL.
"Why" is indeed one of my follow-up questions should it turn out that anyone is willing to confess to doing this