rohc Robust Header Compression 49 IETF December 2000
rohc Robust Header Compression 49. IETF December 2000 San Diego Chairs: Carsten Bormann <cabo@tzi. Org> Mikael Degermark <micke@cs. Arizona. edu> Mailing List: rohc@cdt. luth. se Digitale Medien und Netze 1
49 th IETF: Agenda (from 30000 feet) u 1. AD and WG chair admonishments u 2. Real agenda üBlue sheets üScribe Digitale Medien und Netze 2
Hello! This is an IETF Working Group u We are here to make the Internet work (Fred Baker) u Rough Consensus and Running Code (Dave Clark) u Working Group is controlled by s s s IETF Process (RFC 2026, RFC 2418) – read it! Area Directors (ADs): Alison Mankin, Scott Bradner Charter (http: //www. ietf. org/html. charters/rohc-charter. html) -- read it! Working Group Chairs: Mikael Degermark, Carsten Bormann Technical Advisor: Erik Nordmark u Work is done on email list rohc@cdt. luth. se And on IETF meetings, interim meetings, informal meetings, … s Mailing list is official channel, though s Digitale Medien und Netze 3
RFC 2026: Internet Standards Process u Standards track RFCs: WG consensus (as judged by WG chairs) s WG last call s IESG approval (based on AD recommendation) s s s Quality control! IETF last call u Informational RFCs u BCP (best current practice) RFCs Digitale Medien und Netze 4
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 […] u Where 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 specification(s) under openly specified, reasonable, non-discriminatory terms. Digitale Medien und Netze 5
RFC 2026: IPR issues (2) u Contributions (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. ” u. I. 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 6
IPR issues: ROHC WG policy u IETF IPR policy defined in RFC 2026 u For expedience: Include IPR statements in the contributions (I-Ds, slides) s Upon advancement to RFC, these IPR statements will be replaced by a pointer to http: //www. ietf. org/ipr u Unencumbered technologies facilitate interoperability and are therefore generally preferable Of two roughly equal proposals, select the unencumbered one s Desirable: Default configuration is unencumbered s Digitale Medien und Netze 7
ROHC: Charter (1) u Cellular links: expensive, limited bandwidth u IP/UDP/RTP and IP/TCP packets benefit considerably from header compression u Existing schemes (RFC 1144, RFC 2508) do not perform well over cellular: high error rates and long link roundtrip times s do not compress TCP options such as SACK or Timestamps s u Goal of ROHC: develop header compression schemes that perform well over links with high error rates and long roundtrip times. s must perform well for cellular links built using technologies such as WCDMA, EDGE, and CDMA-2000. s should also be applicable to other future link technologies with high loss and long roundtrip times s Ideally, it should be possible to compress over unidirectional links. s Digitale Medien und Netze 8
ROHC: Charter (2) u Good performance: minimal loss propagation s minimal added delay s And, of course, the size… u Target: generic TCP and UDP/RTP compression s applications of particular interest: voice and low-bandwidth video s u ROHC may develop multiple compression schemes e. g. , for specific link layer technologies s additional schemes may be added in consultation with the ADs. s u Must: assure that when a header is compressed and then decompressed, the result is semantically identical to the original; s perform well when end-to-end path involves more than one cellular link; s support IPv 4 and IPv 6. s Digitale Medien und Netze 9
ROHC: Charter (3) u First task: Create more thorough requirements documents u Maintain connections with other standardization organizations developing cellular technology for IP, such as 3 GPP and 3 GPP-2 s ensure that output fulfills their requirements and will be put to good use u Develop a solid understanding of the impact that specific error patterns have on HC schemes, and document guidelines to L 2 designers regarding what L 2 features work best to assist L 3/L 4 HC u Address interactions with IPSEC and other security implications. u. Remember: Only IESG can change the charter! Digitale Medien und Netze 10
ROHC: Charter (4) Goals and Milestones u u u Done Mar: I-D on Requirements for IP/UDP/RTP HC. in last-call May: I-D of layer-2 design guidelines. Start now May: I-D(s) proposing IP/UDP/RTP HC schemes. To do May: I-D of Requirements for IP/TCP HC. Jun: Requirements for IP/UDP/RTP HC submitted to IESG (Inf. ) Jul: Requirements for IP/TCP HC submitted to IESG (Inf. ) Jul: Resolve possibly multiple IP/UDP/RTP HC schemes into a single scheme. Aug: I-D on IP/TCP header compression scheme. Sep: Layer-2 design guidelines submitted to IESG (Inf. ) TCP g/l Sep: IP/UDP/RTP HC scheme submitted to IESG (PS) Dec: IP/TCP HC scheme submitted to IESG (PS) Jan: Possible recharter of WG to develop additional HC schemes. Digitale Medien und Netze 11
49 th IETF: Agenda (Tuesday) u 1545 Chair admonishments and agenda u 1555 WG document status s In last-call s s Bormann (5) draft-ietf-rohc-rtp-lower-layer-guidelines-00. txt Svanbro (1) draft-ietf-rohc-rtp-requirements-03. txt Degermark (1) draft-ietf-rohc-rtp-06. txt Bormann (10) New doc needed on ROHC performance? u 1615 News from ROHC testing s (10) (20) Japan Telecom testing update (3) (15) Sato (15) u 1630 ROHC over PPP: Carsten Bormann (15) draft-ietf-rohc-over-ppp-00. txt s discussion s Bormann (5) (10) Digitale Medien und Netze 13
49 th IETF: Agenda (Wednesday) u 0900 ROHC future I: packet encoding draft-price-rohc-epic-00. txt s Discussion s Price (20) (10) u 0930 TCP s s s 0930 TCP vs. RTP requirements 0935 draft-ietf-rohc-taroc-00. txt 1005 TCP via EPIC (on mailing list: “draft-price-rohc-epic-tcp-00. txt”) 1025 TCP way forward 1030 Signaling compression Degermark (5) Zhang (20+10) Price (10+10) Chairs (5) Hannu (20) u 1050 ROHC future II: 0 -byte solutions 1050 draft-mccann-rohc-gehcoarch-00. txt s 1105 discussion s u 1119 Bakeoff? u 1120 Need for Rechartering? P. Mc. Cann (15) (14) (10) Digitale Medien und Netze 14
WG document status: In last-call u draft-ietf-rohc-rtp-lower-layer-guidelines-00. txt (Oct 12) s No last-call comments yet u draft-ietf-rohc-rtp-requirements-03. txt (Nov 20) s Few last-call comments (see next slides) u draft-ietf-rohc-rtp-06. txt (Nov 29): RTP ROHC Main deliverable s 156 pages (should be 100) s Bulk of last-call comments (see next slides) s Editor left out performance assessment material s s Separate document needed? Digitale Medien und Netze 15
WG Last-call: Short term time schedule u End-date given in last-call: 2000 -12 -14 about 1400 Z u But the year has only 49 weeks! u But then, this is a 14 -day WG last-call : -) u Editor probably needs a few more days before submitting to the IESG IETF last-call could start before Christmas s Next IESG meeting probably in 3 rd week of January s u If all runs really well, in the RFC-ed queue end of Jan Digitale Medien und Netze 16
ROHC-RTP-requirements-03 last-call issues u Editorial s Make definition of loss/damage propagation consistent with rest u Issue: Handover requirements 3 a: Handover loss events change to “Loss events of length 150 ms should be handled efficiently and without additional packet loss or erroneous headers being introduced by ROHC”. s 3 b: Handover context recreation add “ROHC should not introduce packet loss during such events”. s Digitale Medien und Netze 17
ROHC-RTP-06 last-call issues (1) u Editorial: s s s s Make definition of loss/damage propagation consistent with rest 5. 2: can’t have Add-CID on feedback Two sections numbered 5. 2. 6 5. 6. 4: replace by just pointing to 5. 6. 3 5. 7: Fix GRE section references (5. 8. 4. 5, not 5. 8. 4. 2) 5. 7. 5. 2: Tsc is always 1, set to 0 in ext 3 5. 7. 6. 11 feedback example: code: s/2/1/ Issue: Location of “Impairment Considerations” section Digitale Medien und Netze 18
ROHC-RTP-06 last-call issues (2) u Not-quite-so-editorial: s s s 5. 3. 2. 2. 3: No CRC-based repair of SN in context in R-mode 5. 3. 2. 2. 4: b for Full->Static, c for Static->no context 5. 4. 2. 2: handle IR-DYN packets like UOR-2 packets 5. 6: Add reference to mode bits in IR/IR-DYN 5. 7. 5. 1. : RND flag (no ACK should be required in O-mode) Appendix A: v 4 vs. v 6 (fixed text supplied by Lars-Erik) Digitale Medien und Netze 19
ROHC-RTP-06 last-call issues (3) u Document dependencies: s Remove MIPv 6 destination option support (!) s s s HA works well enough without special support BU etc. occur infrequently GRE: replace RFC 1701 reference by RFC 2784 (and attendant changes) u Change requests: s Issue: IR-DYN/UOR-2 feedback option (withdrawn) Digitale Medien und Netze 20
Performance document u Extensive thread in mailing list about descriptive text u Removed most of this from – 06 u Preserve text in informational document? 1. Do it now to help initial implementors make decisions 2. Do it later when people have implemented it 3. Refer people to INFOCOMM et al. Digitale Medien und Netze 21
ROHC over PPP u Son-of-2509 (PPP negotiation in IPCP) Example for negotiation needed by other types of links s Progress this independently of main document s s Need PPP protocol identifier! (or two? ) u Channel setup s PPP negotiation sets up channel s s MAX_CID, MRRU, MAX_HEADER LARGE_CIDS flag (issue: IPv 4 vs. IPv 6!) Set of profiles in use PPP: No need to identify special uncompressed profile-0 context s Always can use PPP protocol identifier 0 x 21 instead of profile 0 Digitale Medien und Netze 22
EPIC – how to use? u Do we want to take this up for further ROHC work? u Need a way to use this in standards Could standardize the output of the EPIC processor (duuh) s Define EPIC processor input language? s u Hard to do the all-layers trick here… Will have to cooperate with other bodies s Are we the right body to “package” EPIC for them? s Digitale Medien und Netze 23
ROHC TCP – why develop separately? u The requirements for robustness may be less stringent s Can do retransmission at link layer (see PILC) u Less stringent time constraints on development u Different protocol than RTP (obviously) u New problems: Options like SACK, timestamps u Solicit wider input wrt next generation TCP compression s But is this maybe still a researchy topic? Digitale Medien und Netze 24
ROHC TCP Requirements u Different link properties s No residual errors, but may have packet loss u Robustness: s Should not disable [might even help] TCP mechanisms s fast retransmit, fast repair, etc MUST NOT generate damaged headers (that can pass TCP chksum!) s Must deal with current and future TCPs s SACK, timestamp, ECN, Diffserv, Initial TCP negotiation, etc TCP sequence numbers and IP ID less predictable u Might want it to work well for short-lived TCP transfers? u Solve known problems with TCP Checksum Window scale option – satellite links (loss of 64 K undetectable) s window field decrement + seq no increment (rfc 1144) s Digitale Medien und Netze 25
TCP – way forward? u Need requirements document s How much can you guess about TCP implementations u Need lower-layer guidelines document s How much L 2 reliability is good for you? u Start work on TCP scheme State management s Assume EPIC for encoding? s Digitale Medien und Netze 26
Signaling compression u Usefulness in the presence of encryption? u How application independent can we get? u Relationship to TCP filter proposals? (end 2 end) s End 2 end does not work with existing terminals u Relationship to e. g. PPP CCP standards? Digitale Medien und Netze 27
0 -byte – way forward? u Lots of confusion on what we are doing here Distinguishing element: use synchronous, fixed frame channel s Allow for buffering in the compressor s u Architecture (End) system “IP Stack” architecture s Protocol architecture s u Does it work in mid-path? u Document limititations E. g. , non-transparent solution may not work with payload compression that uses SN/TS as initialization vector s ECN bits, IP-ID, … on downlink side… RTCP… s u It seems we need a requirements delta document Digitale Medien und Netze 28
Bakeoff? u PPP u WCDMA? u EDGE? u CDMA 2000? u Host u Test sequences s Negotiation, mode transitions, state transitions, packet formats u Infrastructure, reference points Digitale Medien und Netze 29
Rechartering? u Develop EPIC as a base technology u RTP ROHC extensions (e. g. , 0 -byte) u Need new timeframe for TCP u SCTP – how urgent? Digitale Medien und Netze 30
- Slides: 29