Hi Stewart, > -----Original Message----- > From: Int-area [mailto:[email protected]] On Behalf Of Stewart Bryant > Sent: Wednesday, April 29, 2015 3:32 PM > To: [email protected] > Subject: Re: [Int-area] Why combine IP-in-UDP with GUE? > > I confess that I have only skimmed this thread, but as far as I can see no > one has > mentioned OAM. If we are designing a general purpose encapsulation there > really needs to be an OAM indicator so that OAM can fate share with the data > that it is monitoring.
IP-in-UDP is just intended to be a poor-man's UDP encapsulation for IP and act as a potential replacement of IP-in-IP and IP-in-GRE in Softwire networks. Here the UDP itself is deemed as a tunneling protocol (see Section 3.1 of RFC6936, https://tools.ietf.org/html/rfc6936#page-10). As for a general purpose UDP-based encapsulation which have taken fragmentation, security, OAM, checksum, congestion control and even metadata etc into account, it could be GUE, GENEVE, VXLAN-GPE... > I don't think it is a factor in this discussion unless people think that we > will be > doing super DPI, but > RFC4928 notes an issue when using the first nibble of a packet as a type > identifier in an IP/MPLS environment. In a nutshell this says avoid 0 and 1 > as well In the IP-in-UDP case, a to-be-allocated destination port value is used to indicate that the UDP tunnel payload is an IP packet. As for whether the encapsulated IP packet is IPv4 or IPv6, it would be determined according to the Version field in the IP header of the encapsulated IP packet. Therefore, it seems a bit different from the first nibble issue associated with the MPLS encapsulation. Best regards, Xiaohu > as 4 and 6. > > - Stewart > > > > _______________________________________________ > Int-area mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/int-area _______________________________________________ Int-area mailing list [email protected] https://www.ietf.org/mailman/listinfo/int-area
