In message <290E20B455C66743BE178C5C84F1240847E63346BA@xxxxxxxxxxxxxxxxxxxxxx> l.wood@xxxxxxxxxxxx writes: > On nested checksums, the question is how they are nested; it's a matter > of scope. With a bunch of checksums checking only a payload and any > inner checksums like Russian Matryoshka dolls, the end-to-end argument > tells us that for reliable receipt of the payload, only the innermost checksum > matters. > > But here, we are not solely checking the payload, but information on how to > deliver and identify that payload - and while an outer Ethernet CRC is across > the last link, the UDP checksum, though weak, provides a check on the IP > addresses and UDP ports (via the pseudoheader check) and MPLS stack > from UDP/IP source to UDP/IP destination (and the payload, which is the bit > everyone focuses on as the performance hit as redundant and a processing > cost when the payload has its own check, and the bit that UDP-Lite can leave out). > > Nothing else checks that scope. The scope is wider, and affects the network > as a whole. Errors in these unchecked fields lead to misdirection and lead to > misdelivery. Or pollution of other ports. > > The MPLS assumption is that it's protected and checked by a strong link CRC like > Ethernet, and checked/regenerated by stack processing between hops; here, > in a path context, with zero UDP checksums MPLS has no checking at all. That UDP would be running over IP over Ethernet or POS or GFP or ... There is no layer-2 currently in use that does not have a robust FCS, generally a 32 FCS and therefore the MPLS assumption of checking at a lower layer is still valid. Curtis > "consequences for cheap hardware and for software implementations" > > I'm sorry, when was MPLS cheap? > > Lloyd Wood > http://about.me/lloydwood > ________________________________________ > From: Adrian Farrel [adrian@xxxxxxxxxxxx] > Sent: 09 January 2014 10:21 > To: Wood L Dr (Electronic Eng); randy@xxxxxxx > Cc: gorry@xxxxxxxxxxxxxx; mpls@xxxxxxxx; ietf@xxxxxxxx; david.black@xxxxxxx; tsvwg@xxxxxxxx; jnc@xxxxxxx; lisp@xxxxxxxx > Subject: RE: [mpls] draft-ietf-mpls-in-udp was RE: gre-in-udp draft (was: RE: [tsvwg] Milestones changed for tsvwg WG) > > Lloyd and Randy, > > With respect to draft-ietf-mpls-in-udp, this is why we have IETF last calls, so > thanks for the comments. > > We did take the precaution of sending this I-D for an early TSV Directorate > review because of the concern about a number of factors and the overlap with > tsvwg work, but the review came back "clean". Of course, such a review is just > one person, so this conversation is good. > > Wrt zero checksum, where do you stand on nested checksums? There is some claim > that they represent a waste of processing. I am not convinced by that when each > layer is using dedicated hardware (that can presumably process checksums at line > speed), but I am interested in the consequences for cheap hardware and for > software implementations (as have been claimed to be some of the motivations for > this work). > > Other TSV-related issues that surely pop up are: > - allocation of ports for foo-in-UDP > - congestion control > > Please note that there are a number of I-Ds that you missed in your broad sweep > of "I am opposed". You should probably look at the NVGRE and VXLAN work (which I > think is lurking around the NVO3 working group) because that is also looking at > UDP encaps of a tunnelling protocol. > > Thanks, > Adrian > > Health warnings: > I am responsible AD for draft-ietf-mpls-in-udp > I am a co-author of the gre-in-udp draft. > > > -----Original Message----- > > From: mpls [mailto:mpls-bounces@xxxxxxxx] On Behalf Of l.wood@xxxxxxxxxxxx > > Sent: 09 January 2014 08:07 > > To: randy@xxxxxxx > > Cc: gorry@xxxxxxxxxxxxxx; mpls@xxxxxxxx; ietf@xxxxxxxx; david.black@xxxxxxx; > > tsvwg@xxxxxxxx; jnc@xxxxxxx; lisp@xxxxxxxx > > Subject: Re: [mpls] draft-ietf-mpls-in-udp was RE: gre-in-udp draft (was: RE: > > [tsvwg] Milestones changed for tsvwg WG) > > > > Randy, > > > > okay, let tsvwg adopt draft-yong-tsvwg-gre-in-udp-encap, and let's get > > consensus on it. And then the authors can adopt that consensus for > mpls-in-udp, > > which overlaps in authorship... > > > > thanks, > > > > Lloyd Wood > > http://about.me/lloydwood > > ________________________________________ > > From: Randy Bush [randy@xxxxxxx] > > Sent: 09 January 2014 07:51 > > To: Wood L Dr (Electronic Eng) > > Cc: david.black@xxxxxxx; gorry@xxxxxxxxxxxxxx; ietf@xxxxxxxx; mpls@xxxxxxxx; > > jnc@xxxxxxx; lisp@xxxxxxxx; tsvwg@xxxxxxxx > > Subject: Re: draft-ietf-mpls-in-udp was RE: gre-in-udp draft (was: RE: [tsvwg] > > Milestones changed for tsvwg WG) > > > > > Because they specify zero UDP checksums, > > > I oppose publication of draft-ietf-mpls-in-udp in its current form > > > I oppose tsvwg adoption of draft-yong-tsvwg-gre-in-udp-encap in its current > > form. > > > I oppose the IETF lisp documents. > > > > lloyd, > > > > i think i understand your position. but i disagree with preventing wg > > adoption of draft-yong-tsvwg-gre-in-udp-encap, mainly because i strongly > > see wg adoption as how we get to discuss and work on a document, not as > > approval of the document. as david said, i think we need to discuss it > > so we can decide if it should be fixed. to do so, we have to adopt it. > > > > randy