IPX vs IP performance

Can anybody tell me if there would be a network performance benefit moving from ipx/spx to tcp/ip protocol under Netware 5? I'm sure I recall that someone said that less network traffic is created when using IP. The Netware server runs multiple database intensive applications that can struggle when the network becomes busy so we're looking at all ways to increase performance.

Thanks in advance.
Who is Participating?
I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

I wouldn't do it on NetWare 5, myself, since that's an obsolete version, and some services still relied on SPX calls IIRC,  but it is possible to do all IP on 5.0.  

What SP is on the NW5 server(s)?  What database application is running, and is it SPX/IPX-dependent?  You may not be able to move off IPX for your database traffic.

IPX is a "chatty" protocol - it broadcasts service advertisements and keepalive packets all the time, on top of the regular traffic.  IPX also seems to be a better routable transport for NetBIOS than TCP/IP is.

To properly configure a "pure IP" network with NetWare, you must configure SLP (Service Location Protocol) to enable services to be found, since there is no "sap broadcast."  SLP is client-driven - a client service agent will multicast (not broadcast) looking for service agents that would then reply with locations of services.

That's another thing with NW5 - it uses SLP V1.  Starting with NW5.1, IIRC, SLPV2 was implemented to improve the efficiency of SLP, by providing the Directory Agent, a Service Agent type that other Service Agents report to, that registers and maintains a list of all the services on the network, and their locations.  If a client knows the location of the DA, it can unicast to the DA and get access to the entire list of service locations.

Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial

5.1 shipped using SLPV1.You could move it to SLPV2, but it wasn't officially supported.
IPX on a LAN is superior in terms of speed and efficency, but not so on a WAN.
There is as much as a 20% difference in terms of throughput when measuring IPX vs. IP on a LAN.
IP has superior ECC, but it adds overhead to the total system throughput(great for a WAN ,but not as critical on a LAN.

What are you running in terms of infrastructure (switches, hubs, 10 or 100mb)?

There are a number of reasons a Novell network will slow down.

Collisions (if you are running hubs instead if switches)
Not enough RAM in the servers. (Cache buffers which should be at > 60%)
Slow disks (RAID not tweaked for DB performance).

limsmanAuthor Commented:
Thanks for your responses. I'll get back when I've checked some things out.
Upgrade your Question Security!

Your question, your audience. Choose who sees your identity—and your question—with question security.

limsmanAuthor Commented:
ShineOn: Truth is I don't know what SP we're running on NW5 as I'm not the network administrator, but as you've said, it is unwise on NW5 due to some services using IPX. We are running several databases on the server. They are running on Extended Systems Advantage Database Server (ADS) and also Paradox. ADS, our largest DB is client/server based, but Paradox is not of course. I'm sure both databases run under IP. Your comments have been very helpful and I've passed them on to the network administrator who will better understand the details you have gone into.

pgm554: Thanks for your valued comments which have again been passed onto the network administrator. In answer to your questions, we are running a mix, mainly 100BASE-TX through switches over a fibre backbone, but one segment running 10BASE2 possibly over 10BASE5 backbone. Not sure if they are still hubs or switches in that area since upgrading is in progress on that segment. Just last week we upgraded the RAM in the server in an attempt to improve things. Disks are RAID but not sure how they are configured (i.e. towards DB performance).
Anything having to do with 10BASEx is a bottleneck.
And depending upon how your switches and hubs are deployed can cause issues.

There are tweaks that can be applied to the server for DB efficiency


Here is an example of tweaking the NW set parameters in the autoexec.ncf:

SET Maximum Concurrent Disk Cache Writes = 500
SET Dirty Disk Cache Delay Time = 0.5
SET Maximum Concurrent Directory Cache Writes = 100

SET maximum directory cache buffers = 4000
SET minimum directory cache buffers = 2000
limsmanAuthor Commented:
Once again, thanks for your comments pgm554. My manager has already agreed that the 10mb segment could cause a bottleneck which is the reason for the upgrade. The tweaking parameters are certainly worth checking out.

I'll close this question off now and split the points between the two experts that have helped, with the accepted answer going to ShineOn as the first to answer the original asked question.

Thanks again.
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today
Novell Netware

From novice to tech pro — start learning today.

Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.