We are in the process of changing over our Novell Server to NT server with the idea of eventually eliminating Netware and IPX.

Presently, the workstations are all NT, on an Ethernet LAN.

The NT server, with the Revelation NT product is up and running as a test. The Novell server is running the VLM product.

On those workstations we have added IP to, we are experiencing a a general performance slowdown, both in terms of loading the product @ the workstation and in running large data extracts for reports. For example, running a verlifylh on our data volume usually takes about 25 minutes and has grown to over 1.5 hours now, with no significant or suddent increases in the data population. (Approx 368 MB in 24 tables.)

Is there a significant slowdown under IP for the AREV product in general? Or because of the mixed protocols, is there degregation? Or should I be looking more @ Peter Walters April & May 1998 related topic?

Any ideas would be helpful.

Thanks

Dana Corson

Dana..

I think you're experiencing "Protocol-Stacking" issues on your NT Desktops. What seems to happen is this:

NT/TCP-IP gets about 4 times as much process time as anything else.   In general, IPX-SPX uses a default 1K packet size.  TCP-IP can use   up to 10K packet sizes on NT - maybe even larger.  Therefore IPX-SPX   gets runt's udder on the desktop.  I've seen significant I/O   degradation when multiple protocols are active.

I don't know how to change the priority of service protocols. Maybe a good Guru can help…

Don Miller

C3 Inc.

We are hoping that's the case. When we finally make the break from IPX to IP we will see what happens. (Next week sometime.)

Thanks.

Dana Corson

View this thread on the forum...