Digitale Medien und Netze 1 rohc Robust Header Compression 51th IETF August 2001 London Chairs: Carsten Bormann Mikael Degermark Mailing List:

Digitale Medien und Netze 2 51 st IETF: Agenda (from feet) u1. WG chair admonishments u2. Real agenda Blue sheets Scribe

Digitale Medien und Netze 3 Hello! This is an IETF Working Group uWe are here to make the Internet work (Fred Baker) s Together! (Harald Alvestrand) uRough Consensus and Running Code (Dave Clark) uWorking Group is controlled by s IETF Process (RFC2026, RFC2418) – read it! s Area Directors (ADs): Alison Mankin, Scott Bradner s Charter ( -- read it! s Working Group Chairs: Mikael Degermark, Carsten Bormann s Technical Advisor: Erik Nordmark uWork is done on list s And on IETF meetings, interim meetings, informal meetings, … s Mailing list is official channel, though

Digitale Medien und Netze 4 RFC 2026: Internet Standards Process uStandards track RFCs: s WG consensus (as judged by WG chairs) s WG last call s IESG approval (based on AD recommendation) s Quality control! s IETF last call uInformational RFCs uBCP (best current practice) RFCs

Digitale Medien und Netze 5 RFC 2026: IPR issues (1) u(10.2) No contribution that is subject to any requirement of confidentiality or any restriction on its dissemination may be considered […] uWhere the IESG knows of rights or claimed rights […] the IETF Executive Director shall attempt to obtain from the claimant […] a written assurance that upon approval by the IESG of the relevant Internet standards track specification(s), any party will be able to obtain the right to implement, use and distribute the technology […] based upon the specific specification(s) under openly specified, reasonable, non-discriminatory terms.

Digitale Medien und Netze 6 RFC 2026: IPR issues (2) uContributions (10.3.1(6)): “The contributor represents that he has disclosed the existence of any proprietary or intellectual property rights in the contribution that are reasonably and personally known to the contributor.” uI.e., if you know of a patent application for a technology you are contributing, you have to tell. Or just shut up entirely!

Digitale Medien und Netze 7 ROHC: Charter (4) Goals and Milestones uMar: I-D on Requirements for IP/UDP/RTP HC. uMay: I-D of layer-2 design guidelines. uMay: I-D(s) proposing IP/UDP/RTP HC schemes. uMay: I-D of Requirements for IP/TCP HC. uJun: Requirements for IP/UDP/RTP HC submitted to IESG (Inf.) uJul: Requirements for IP/TCP HC submitted to IESG (Inf.) uJul: Resolve possibly multiple IP/UDP/RTP HC schemes into a single scheme. uAug: I-D on IP/TCP header compression scheme. uSep: Layer-2 design guidelines submitted to IESG (Inf.)  TCP g/l uSep: IP/UDP/RTP HC scheme submitted to IESG (PS) uDec: IP/TCP HC scheme submitted to IESG (PS) uJan: Possible recharter of WG to develop additional HC schemes. Done in last-call Working To do

Digitale Medien und Netze 8 51 st IETF: Agenda (Thu) u1530 Chair admonishments and agenda (10) u1540 WG document status(15) u1555 Input from ROHC Bay-Cough s 1555 Results from Roke ManorBormann (10) s 1605 Discussion/Implications(10) u byte s 1615 Document statusJonsson (5) s 1620 CDMA2000 and GERAN mappingsJonsson (5) s 1625 Discussion(10) u1635 TCP s 1635 requirements issues & discussionChairs (30) s 1705 Draft updates; discussion; merging?Authors (30)

Digitale Medien und Netze 9 51 st IETF: Agenda (Fri) u0900 Signalling Compression s 0900 Overview/Taxonomy _____Chairs (10)_____ s 0910 Proposal Shootout Authors (35) s 0945 Discussion(30) s 1015 Requirements(20) s 1035 Way forward(10) u1045 Way to Draft Standard s 1045 Carving up the standard(10) s 1055 ROHC MIB discussionQuittek (20) u1115 Rechartering(15)

Digitale Medien und Netze 10 WG documents in publication: RTP ROHC uPublished: s RFC3095: Framework and four profiles (was: draft-ietf-rohc-rtp-09.txt) s RFC3096: RTP requirements (was: draft-ietf-rohc-rtp-requirements-05.txt) uAlready part of 3GPP Release 4 s Alongside with R99’s inclusion of RFC2507 (not RFC2508!) uAdopted by 3GPP2 s Release C end 2001

Digitale Medien und Netze 11 Lower layer guidelines udraft-ietf-rohc-rtp-lower-layer-guidelines-01.txt uCompleted WG last-call in December uStill stalled s AD input: Prescriptive text not in style for Informational s Did not attempt IETF last-call (to avoid stall for RTP ROHC!) uHow to proceed? s 1) submit as BCP – well, not really based on wide-spread experience s 2) edit and submit as informational -- who does this? s 3) punt and reuse material in implementers guide

Digitale Medien und Netze 12 ROHC over PPP udraft-ietf-rohc-over-ppp-02.txt uSon-of-2509 (PPP negotiation in IPCP) s Makes ROHC immediately useful in PPP world s Also: Example for negotiation needed by other types of links uChanges from –01 s Explicitly allow multiple instances of IP-Compression-Protocol uFinal touches: s Define merging rules for IPCP and IPV6CP option results uWG last-call –03 next week

Digitale Medien und Netze 13 ROHC Implementers’ Guide? uCould take up lessons from Interop Events u(Could take up L2 guidelines. No.) uAt Draft Standard level, could move some of the material into one of the Standards Track documents uWho? s Mikael Degermark, Péter Krémer

Digitale Medien und Netze 14 LLAROHC – way forward? uRelatively simple extension to ROHC uMapping for CDMA2000 defined, GERAN forthcoming uWould be most useful in September 3GPP2 release uLast-call LLAROHC now s draft-ietf-rohc-rtp-0-byte-requirements-00.txt  Informational  draft-jonsson-rohc-lla-rtp-01.txt++  PS uAnother revision on CDMA2000 mapping s Add GERAN s Try September 2001 for draft-ietf-rohc-lla-mapping-00.txt   Informational

Digitale Medien und Netze 15 TCP ROHC requirements uAlready on ROHC charter! s But we didn’t get around to it, yet uThere is no TCP HC out there that does SACK, ECN, … s The world is looking at ROHC to fix this s Attempt to be future-proof! èROHC TCP must be applicable in the wide Internet s Encumbered solutions won’t cut it!

Digitale Medien und Netze 16 ROHC TCP Requirements uLink properties != RTP s No residual errors, but may have packet loss uShould work well for short-lived TCP transfers s In particular, for sequences of connections between same pair of hosts uMust deal with current and future TCPs s SACK, timestamp, ECN, Diffserv, SYN/FIN/Initial TCP negotiation, etc uRobustness: s Should not disable [might even help] TCP mechanisms s fast retransmit, fast repair, New Reno, etc s MUST NOT generate damaged headers (that can pass TCP chksum!) s TCP sequence numbers and IP ID less predictable s Solve known problems with TCP Checksum s Window scale option – satellite links (loss of 64K undetectable) s window field decrement + seq no increment (rfc1144)

Digitale Medien und Netze 17 ROHC TCP Requirements Issues uWhat level of link-layer retransmission?  A: Some  scheme should work well with persistent link-layer retransmission, but also with higher link-layer loss rates uShould we expect residual bit errors? s A: No, but there is a non-zero probability, and we should make sure we don’t increase it uReordering between compressor and decompressor? s A: Yes. How much? s Need to fix framework in the process (reordering between CIDs is fine already, but need to work on reordering within) s Q: TCP ACK acceleration? Probably not.

Digitale Medien und Netze 18 TCP – way forward? uFreeze –02 requirements document uAdd TCP Model document s What kinds of TCP exchanges do we expect; point to traces s Not a “benchmark”, but a checklist s Present at Salt Lake City TSVWG uNeed lower-layer guidelines document? s How much L2 reliability is good for you? s See PILC ARQ document! uStart work on merging TCP schemes s State management s Assume EPIC for encoding?

Digitale Medien und Netze 19 Signaling compression – make it a WG item? uPro: s It’s needed! (Call setup time will be bad without it) s Fits in ROHC framework *if* done hop-by-hop s No changes to end systems, more redundancy to look at s Hop-by-hop makes it easier to compress between calls uHmm: s Might be better done end-to-end (or in SIP proxy) s What about IPCOMP, TCPFILTER and friends? s Not really header compression (do we care?) s Is hop-by-hop still useful once SIP gets secure? uCon: s IPRs?

Digitale Medien und Netze 20 Signaling compression – hbh vs. e2e uE2e means UE to P-CSCF, *not* UE to UE/phone uHbh means UE to first IP node in RAN uHbh (link-layer) issues: s Security – but almost the same problem applies with “e2e” s Handover requires work s Need to require certain SIP port usage s ROHC contexts are unidirectional so far – need mechanisms for bidirectionality uE2e (transport-/app-layer) issues: s How to negotiate compression algorithm? SRV??? s Link layer has negotiation scheme already s Need ROHC for IP/UDP compression, anyway

Digitale Medien und Netze 21 Signaling compression – way forward? uAgree e2e vs. hbh s Do e2e now uMerge requirements documents s Issue: memory size? uGenerate signaling Model document s What kinds of SIP exchanges do we expect s Not a “benchmark”, but a checklist uStart work on merging and simplifying schemes s Start with an apparently unencumbered scheme

Digitale Medien und Netze 22 Rechartering: Goals and Milestones (1) uROHC-over-PPP s WG last-call –03 August 2001, submit September 2001 uLower-layer Guidelines: submit for Informational RFC s WG last-call (again) –02 August 2001 uImplementers’ Guide: –00 in September u0-byte IP/UDP/RTP s Try for 3GPP2 deadline (September 2001) s Requirements and Assumptions: I-D May 2001 s Solution ID (LLA) –01: July 2001 s WG last-call LLAROHC –00 August 2001, submit August 2001 s Informational ID (CDMA2000/GERAN): September 2001 s WG last-call October 2001

Digitale Medien und Netze 23 Rechartering: Goals and Milestones (2) uSignaling compression s Shoot for 3GPP R’5 deadline (RFC in December 2001) s Focus on call-setup time and in-call stealing issue (SIP/SDP) s Requirements and assumptions draft: merge August 2001 s Include draft-hannu-rohc-signaling-cellular-02.txt as problem stmt s Start merging solutions documents now! s WG last-call October 2001 s Test interoperability November 2001

Digitale Medien und Netze 24 Rechartering: Goals and Milestones (3) uTCP: new dates s Requirements and assumptions frozen: August 2001 (London) s Call-for-freeze to ROHC, PILC, TSVWG s TCP model document: –00 Sep, –01 for SLC (November 2001) s draft-ietf-rohc-tcp-00.txt: September 2001 s WG last-call March 2002, submit April 2002 uEPIC s Need to be done before TCP if we want to use it for that s Separate notation document draft-ietf-rohc-epic-00: August 2001 s Decide: Interoperable implementations by Dec 2001? Slip some more?

Digitale Medien und Netze 25 Rechartering: Goals and Milestones (4) uROHC over reordering channels? s Do some of the work in TCP uMIB s draft-ietf-rohc-mib-00: October 2001 s WG last-call Jan 2002, submit Feb 2002 uDraft standard by 2Q2002 s Separate documents (Framework, 4 profiles): October 2001 s Merge implementers’ guide: Feb 2002 s WG last-call Apr 2002, submit May 2002 uSCTP s Start collecting requirements now, revisit in SLC

Digitale Medien und Netze 26 Rechartering (5) uRemember: This all has to go through the ADs… Back-burner: uUDP-lite profile? s Try for 3GPP deadline (December 2001)? No. s Requirements, Specification: I-Ds April 2001 s WG last-call August 2001, submit September 2001

Digitale Medien und Netze 27 AOB Two other drafts re ROHC: udraft-minaburo-parameters-00.txt s Negotiates use of a feedback option for link quality feedback udraft-kenward-seamoby-ct-rohc-reqs-00.txt s Context Transfer Considerations for ROHC