This message is forwared from the Routing Research Group mailing list. A new document with int-area implications is available and titled:
"Subnetwork Encapsulation and Adaptation Layer (SEAL)" http://www.ietf.org/internet-drafts/draft-templin-seal-03.txt SEAL in part specifies an MTU determination scheme based on the Report Fragmentation (RF) concept that was first proposed by Charles Lynn on the tcp-ip mailing list in 1987 and re-introduced by Steve Deering on the Path MTU working group mailing list in 1989. Digests from those lists are found at: http://ipvlx.com/tcp_ip.txt http://gatekeeper.dec.com/pub/DEC/WRL/mogul/mtudwg-log Based on the list discussions, the Report Fragmentation (RF) method seemed to be the most attractive approach under consideration for IPv4 path MTU discovery, but was abandoned in favor of the (Don't Fragment/Fragmentation Needed) method that eventually became RFC1191. The RF method seems to have been abandoned due in part to difficulty in procuring an "RF" bit in the IPv4 header, difficulty in defining an ICMP Fragmentation Report message and the concern that not all hosts at that time correctly implemented IP reassembly. The SEAL proposal addresses all of these points (as well as others) and shows that a Report Fragmentation-based path MTU discovery capability for router-to-router tunneling in the Internet is still possible and could correct the operational difficulties associated with traditional path MTU discovery approaches. Moreover, the SEAL approach supports true diversity, especially for subnetworks that contain heterogeneous link types with diverse MTUs and/or L2 address formats. Please review and comment, Fred [EMAIL PROTECTED] _______________________________________________ Int-area mailing list [email protected] http://www.ietf.org/mailman/listinfo/int-area
