Does the Nstreme protocol with Polling turned on work in a similar way as WORP?
The reason I wonder is because of our installations of WORP tend to get high delay with little traffic. This is probably because the base-station skips the polling of stations with little activity when others need the bandwith.
We’ve seen delays vary from 10ms to 200ms, which works bad with terminal solutions.
-Arve
We do delay polling for stations that don’t have any traffic, but it is dynamic and uses the max polling for three seconds after the first traffic. So it should be a problem – except for the first packet.
John
John, I suppose you meant “it should NOT be a problem”?!
arve, are you using Proxim ?
we can share experience on it. We are having problems with that…
Yes, we do use Proxim.
I’ve heard that in Proxim’s new software, 2.1, there should be a function called “WORP No Sleep Mode”. This function is meant to turn of polling on single clients, but we haven’t tried it yet.
The release notes says that you can set this with a SNMP client. In the orinoco.mib, the entry “oriWORPIfConfigTableNoSleepMode” describes the function.
You should expect less throughput with polling turned off.
-Arve
being a proxim matter, if you prefer we can chat off-site gianred123@yahoo.it