Hi Templin, > -----Original Message----- > From: Int-area [mailto:[email protected]] On Behalf Of Templin, Fred L > Sent: Thursday, April 30, 2015 3:16 AM > To: Lucy yong; Tom Herbert > Cc: [email protected] > Subject: Re: [Int-area] Why combine IP-in-UDP with GUE? > > Hi Lucy, > > > -----Original Message----- > > From: Lucy yong [mailto:[email protected]] > > Sent: Wednesday, April 29, 2015 11:48 AM > > To: Templin, Fred L; Tom Herbert > > Cc: [email protected]; [email protected] > > Subject: RE: [Int-area] Why combine IP-in-UDP with GUE? > > > > > > > > Change the GUE header to treat the first nibble as a next header > > > selector. 4 > means IPv4, 6 means IPv6 and X means GUE. > > > [Lucy] As I mentioned in several previous mails, I don't think that > > > this is a good design for GUE. Even if a compression is required, > > > the solution SHOULD use a separate UDP port to indicate IP first and > > > then check the first nibble for IPv4 or IPv6, which, in fact, is the > > > IP- > > in- UDP proposal. > > > > I will say again that the IP-in-UDP proposal as it stands does not have a > > solution > for tunnel fragmentation (made available by GUE). > > [Lucy] That is true. The IP-in-UDP proposal applies the case where no > > fragmentation at tunnel end point is needed. GUE can deal with it. > > Existing tunnel protocols (IP*-in-IP*) are deficient in not providing a tunnel > fragmentation mechanism per Section 3.1.7 of RFC2764.
You may have noticed a fact that most modern routers and switches have been capable of processing jumbo frames for a long while. In other words, fragmentation could be completed avoided in most modern networks. Even in the sparse network environments where fragmentation is still unavoidable, the default configuration which have been widely supported by most vendors (see https://tools.ietf.org/html/draft-ietf-intarea-gre-mtu-03#page-4) should be enough in most cases. That's the reason why that's implemented by many vendors as the DEFAULT configuration, IMHO. BTW, it's preferable to avoid reassembling fragments at the tunnel egress due to the negative impact on the forwarding performance, AFAIK. As such, it's not recommended to perform fragmentation on the tunnel layer and the outer IP layer. > > > However I agree with Tom that the compression concept does not align > > > with GUE principal. GUE encapsulation should not assume IP payload in > > > first > place! > > > > The proposed solution does not assume IP; it assumes a next header selector > > in > the first nibble of the data following the UDP header. > > Non IP can be conveyed when the next header selector indicates GUE. > > [Lucy] the first nibble is intended for indicating version of IP, not > > for a next header selector. Forcing other protocol to follow this > > assumption is bad. Please read RFC4928. There are many cases that GUE > > payload is IP or other and GUE header is required. The first nibble design > > splits > majority from a corner case as first logic, which is a bad design. The > "compression" version does not have the same tunnel property as GUE based > tunnel, we should not mix them at design level. > > I disagree with "bad", and I do not think RFC4928 is applicable when the first > nibble appears within a UDP encapsulation for which there is a specific user > of > that UDP port number. ECMP classifications in the network would not reach > deeply into a UDP encapsulation in order to inspect bits specific to that UDP > port. Fully agree with your points that the first nibble issue associated with the MPLS encapsulation is a totally different story. Best regards, Xiaohu > Thanks - Fred > [email protected] > > > Lucy > > > > Thanks - Fred > > [email protected] > > > > > Regards, > > > Lucy > > > > > > > > > > > > Thanks - Fred > > > [email protected] > > > > > > > Tom > > > > > > > > > Thanks - Fred > > > > > [email protected] > > > > > > > > > >> Tom > > > > >> > > > > >> > Thanks - Fred > > > > >> > [email protected] > > > > >> > > > > > >> >> Lucy > > > > >> >> > > > > >> >> Thanks - Fred > > > > >> >> [email protected] > > > > >> >> > > > > >> >> > However, if GUE payload is IP, it is OK to inspect the > > > > >> >> > first nibble of the payload to determine IPv4 or IPv6 because > > > > >> >> > this > aligns with IP protocol. > > > > >> >> > > > > > >> >> > Thanks, > > > > >> >> > Lucy > > > > >> >> > > > > > >> >> > - 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 > > > > >> > > > > > >> > _______________________________________________ > > > > >> > 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 _______________________________________________ Int-area mailing list [email protected] https://www.ietf.org/mailman/listinfo/int-area
