On Thursday 09 April 2009 15:31:10 Daniel Senie wrote:

On Apr 9, 2009, at 7:15 AM, Robert E. Seastrom wrote:

...

Interesting. When I got my Sprint EVDO card (u727) a year and a half ago, they were pretty nasty about gunning down (bidirectional spoofed RST coming out of the middle of the network somewhere) any TCP sessions that were idle for ten minutes or more.

We observe this same kind of behavior with firewalls in the path watching for dead sessions they can clean up. Appears they send RSTs to both end points when they decide a session has gone away, as that'll let end hosts figure it out sooner. Same workaround of turning on keep=alives once a minute solves this too. The behavior in the case of firewalls makes sense, as state tables have to be cleaned up eventually.

The UMTS world has a lower-layer protocol called HARQ in the radio air interface which functions a little like TCP; the idea is to detect dropped packets on the radio link and retransmit them before the TCP interval times out, thus providing faster recovery. I wouldn't be surprised if there is a similar mechanism to police the use of spectrum; and a lot of mobile operators see "Internet" as an application. Somewhere around I have the incredibly long referral string Vodafone sent my blog server not long after they started real Internet service; a Squid, a Novarra, a 724 Solutions machine of some sort, and I think something else too.