I hate it when I do dumb things and this one was a doozy.
I've been fighting an ongoing case for several weeks now. Things have been extremely busy at work and we've all been going at it like gangbusters to get our work done. Things are so hectic we have to throw some ideas out, read logs or network trace and then scramble to the next case.
It's a young man's game. I'm not young anymore.
Regardless, the case I have been fighting with involved application sharing and slooooooooooow processing. I tried network traces, perfmon traces, client traces... Everything I could think of.
Finally I had the customer break the clients down to a simple switch with only those two machines on it, with them in the same subnet and one connection back to the server. It still was slooooooooooow.
What. The. Fishsticks.
So, I decided to beat my head against the wall this evening and try their antivirus program as invasive as I could make it on my lab to try to make performance suck. IT STILL WORKED!!!!
What the crap? How can their experience be so bad? What could possibly be making that thing suck so hard? What could be limiting the performance? It was like the bandwidth was throttled down to nothing...
You may have guessed it by now. In Lync you can limit the bandwidth you allow between clients in certain scenarios. This is so you don't have Fred the Janitor hogging all the bandwidth with Sue in Accounting doing video chats when Mr. Big is needing to do his conference call with the shareholders.
My customer had limited the bandwidth for his clients through a policy. A quick look back through the logs and I confirmed it.
"What took you so long to find something so obvious?" you man ask me. It is all in presentation. The description of the failure was two VERY specific scenarios. I got so wrapped up in those, I didn't see the obvious point staring me in the face.
Never, ever, take anything for granted. *sigh*