Alberto Medina, Mark Allman, and I have a draft paper on "Measuring the Evolution of Transport Protocols in the Internet" that has a section (Section V.B.) on Path MTU Discovery. >From the paper: "Table X shows that PMTUD is used and succeeded for slightly less than half of the [web] servers on our list. For 31\% of the servers on our list, the server did not attempt Path MTU Discovery. For 18\% of the servers on our list, Path MTU Discovery failed, presumably because of middleboxes that block ICMP packets on the path to the web server." ... "Since a non-trivial number of network elements discard well-known ICMP packets the results of our tests do not offer hope for protocol designers proposing to use new ICMP messages to signal various network path properties to end systems (e.g., for explicit corruption notification, handoff or outage notification, etc.)." We would appreciate any feedback, either on the paper, or on other measurement studies that would be useful on sometimes-unanticipated interactions between transport protocols and routers or middleboxes. The paper and test code are both available from "http://www.icir.org/tbit/";. - Sally http://www.icir.org/floyd/ _______________________________________________ Ietf@xxxxxxxx https://www1.ietf.org/mailman/listinfo/ietf