I know there are a lot of potential variables, which is why I was hoping someone had already made something. :)

Box 1 (PII-300, 256mb RAM, single IDE drive)
1. Win2K-AS, SP2, Post-SP2-Security Patch 1
2. 6 IP numbers on a single NIC, an IIS site on each IP number.
3. Each IIS site running High/Isolated
4. Single DBC on the local filesystem with a single 3.3-million-row table, indexed on some integer and some char columns

Box 2 (PII-300, 256mb RAM, single IDE drive)
1. Win2K-AS, SP2, Post-SP2-Security Patch 1
2. 6 IP numbers on a single NIC, an IIS site on each IP number.
3. Each IIS site running High/Isolated
4. Single MSDE DB on the local filesystem with a single 3.3-million-row table, indexed on some integer and some char columns (or would varchar be better?). Windows Authentication.

Test 1 - Indexed Retrieval
Using the respective MSSQL or VFP7 OLEDB driver, use an ASP page to perform some simple SQL-Selects on some indexed columns in the big table. Then do the same but add a Group By clause.

Test 2 - Table Scan
Using the respective MSSQL or VFP7 OLEDB driver, use an ASP page to perform some simple SQL-Selects on some non-indexed columns in the big table. Then do the same but add a Group By clause.

Test 3 - Self Join
Using the respective MSSQL or VFP7 OLEDB driver, use an ASP page to perform a SQL-Select on the table, joining on column(s) in the same table.

Test 4 - Insert
Use WCAT to post something random to an ASP page, and using the respective OLEDB driver, do n inserts.

Test 5 - Update
? Not sure how to test that one.

After that I would put the DBC and MSDE on a seperate system and do the same tests over the network.



>There's a *large* range of variables that affect this speed. What is your environment?