>So again, my app functions great (Get Records = 15 seconds), as long as there is only 1 user. Surely this is NOT normal VFP behavior or is it? I have been coding FoxPro for 25 years and have never known of this. It sounds like some of you may be trying to tell me, Yes, if 2 or more users have your database open and even if only 1 of those users is querying the database, it could be 4 times slower for the user that does query the database. Is this possible?
>
>One more thing. My app IS slow. I have been working toward a move to TSQL for at least 2 years (and a 100K dollars), but am still months away from converting the 1st installation. The system has 100 tables and almost 1000 views. I do have one more test to conduct. I am now afraid this may be effecting my app at all my customer sites and that I am just unaware of it and because my users know the app is slow they are just putting up with it. I will conduct that test hopefully by tomorrow and will post results.

In my experience there is a significant performance hit going from EXCLUSIVE use to any sort of multi-user SHARED use. Going from 1 user SHARED to 2 users SHARED should not show that great a difference (which if your numbers are as you say are more like 10x rather than 4x slower).

You're clearly willing to, and have, put a lot of effort into testing around this issue. The next step, which I think you alluded to earlier is to set up your own test environment. Ideally you'll want to do 2 tests:

1. With your own (probably small) dev/test data set

2. With a copy of the customer's data

That will tell you:

- if a 4x or 10x slowdown exists in another environment
- whether there's any difference in results between dev/test and some actual customer data

In turn that'll let you know whether to chase gremlins at that particular customer and also give you some idea what other customers should be seeing.

***

I was going to suggest testing with the Server 2008 box actually serving up a copy of the data. This would work around some foibles of using XP as a server, such as the 10 incoming connections limit: http://support.microsoft.com/kb/314882 . Of particular interest: "Per development: The connection limit refers to the number of redirector-based connections and is enforced for any file, print, named pipe, or mail slot session. The TCP connection limit is not enforced, but it may be bound by legal agreement to not permit more than 10 clients." So you can potentially get situations where, say, 10 other machines establish a single mapped drive to the XP "server". If the XP machine is also acting as a print server, and one of the 10 wants to print (or needs to in order to complete some function), it can't get a connection.

However, I think you'll get better, cleaner data faster by setting up your own test environment.