Timeout stumper?

Guys,

I have an app that goes to a fast cluster and then runs a query on a DB behind that cluster. I am having a problem with it timing out at exactly 1 minute. Does this ring a bell with anyone? I see that UDP times out in one minute but it is TCP so don't think that would be a problem. I have sniffed the traffic on both sides and don't see anything crazy. Setup my log to show resets and don't see anything there? The developer tells me that the query does return to the cluster machine but it times out before it returns to the user. The query takes like 90 seconds. They have other queries that take less than a minute and they return fine so it sounds like a timeout SOMEWHERE. But where???

Thanks

Joe