Denis:

I set up a batch with two 700 line invoices. It took less than one minute to post them. Granted no one else is on the network but 30 minutes is way to long no matter what else may be happening.

I suspect that Andrew's response this morning is likley the right answer.

I'm running Dell optiplex workstations and Dell PowerEdger server with a Gigibit switch. The server is running Windows 2003, Raid5 configuration and 4GB of RAM. Workstations are XP Pro with 2GB RAM. Network cards are all Gigibit as well.

Hi Guys...
Thanks for the suggestions...When I tested yesterday, I put the workstation,and both servers onto thier own gigabit switch, by themselves, no other equipment or traffic. I have since found the servers and the workstations have 100mb NICs, so we are maxing out at 100mb. I disabled the antivirus on both servers, with very little noticeable difference. I guess the 100mb NICs are the first suspect now... unless you can propose something else. My plan will be to replace them Tuesday and re-test. Thanks again!

Good day... I wanted to pursue this further, so I took a brand new workstation (XP Pro with gigabit NIC and 4 gb ram) and installed Adagio on it. I connected the workstation through a gigabit switch to my Windows 2003 server (8gb ram) with a gigabit NIC and copied the customers data to it. (No other devices on the network) Both the server and the workstation were running Mcafee total Protection - Small Business. I then posted the transactions and timed how long it took to post the batch. The batch had two invoices, one with 371 line items, and the other with 444. It posted in 6 min 45 sec. I then removed the Mcafee from both the server and the workstation and posted again. The result - posted in 6 min. 37 seconds. I then connected through a 100 mb switch and posted. 57 min 10 sec. Finally, I copied the data to the local workstation C drive and posted again. 10 seconds.
From all of this, I have learned:
1. A gigabit switch makes a major difference in operational speed of the network.
2. Anti virus software ( at least Mcafee ) makes very little difference.
3. The fastest performance hands down is with the data stored locally.

What I still don't know is: Would running Adagio in a virtual desktop client on a terminal server and storing the data locally on that terminal server still be a bad idea performance wise, considering the vast difference in posting speeds for large invoices? Other than installing a new gigabit switch for my customer, I can see no way to improve performance, other than moving Adagio and data to a terminal server and running the software as remote desktops.
I would greatly appreciate your thoughts on this... Thanks!

Did you actually remove McAfee, or just disable it? We have found that some anti-virus products don't actually stop interfering if they are just disabled.

The fastest perfrmance is always with the data local, since the OS can manage the file bufferes for multiple workstation access. When the communication is "over the wire", the data must be delivered and refreshed continuously to make sure that chnges are actually reflected.

I'm still surprised at the discrepancy in posting times between local and the dedicated network connection. Is the server doing anything else?

The terminal server approach will be faster because of the fact that all the copies of the data are being managed locally.

Denis:

What type of hard drives are you using in your server and workstations?

For 5 users or less you should be able to get away with using a terminal server for both programs and data. I've never done one for such a small number of users and I have not actually tested this but in theory this should work.

I actually removed the Mcafee from both the server and the workstation and rebooted both devices. The server is an Intel S3210SHLX server board, Xeon processor, 8 GB Ram, Raid 5 Seagate 500mb SATA drives. The workstations are Intel DG43GT motherboard with 3 gb Core 2 Duo, 4gb ram (Win 7 ready)and a single 500 gigabyte Seagate drive. Both devices have onboard Intel Gigabit ethernet ports. The server was doing nothing else... just sitting there idle.
I am inclined to give the terminal server route a whirl to see how it performs... unless anyone thinks it's a bad idea..

Denis Matte
DMS Services Inc.

Denis:

SCSI vs SATA makes a big difference. SCSI would give a significant performance enhancement.

When I did the testing I had Symantic Endpoint (aka Norton Corporate) AntiVirus fully turned on. The trick with any AntiVirus software is to make sure it is not checking the following folders;

Data
QRASP
Softrak

I've also found the retail and/or OEM version of most antivirus programs to be a problem.

Hi Denis,

Chatting over lunch, we're really surprised at the length of time it took to post those two invoices "over the wire". We'd be expecting no worse than a minute or less, give the 10 second performance on a local copy.

Something else is going on (but I can't be much more help than that, sorry).