Open Research Issues in Internet Congestion Control
| draft-irtf-iccrg-welzl-congestion-control-open-research-04 | |
| Michael Welzl, Dimitri Papadimitriou (ed.) | |
| Michael Scharf, Bob Briscoe |
| Intention: show researchers that finding the ideal TCP++ is not the only open issue for congestion control | ||
| This works: I received a couple of emails from people (e.g. Ph.D. student) saying that the draft helps them | ||
| Effort initiated during 1H’07 (v00 in Jul.07 at IETF 69 – now v04 at IETF 75) | ||
| Document starts with “Global challenges“ | ||
| General observations | ||
| … followed by “Detailed challenges“ | ||
Recent document changes (v03®v04)
| Outcome (as outlined in the slides distributed during the IETF 73 ICCRG meeting) of an in-depth re-work of several sections | |||
| Global challenges: Stability (2.2), Fairness (2.3) | |||
| Specific challenges: | |||
| Addition of dedicated challenge on flow startup (3.4) | |||
| Reworked multi-domain congestion control (3.5) | |||
| Addition of dedicated challenges for | |||
| Multi-layer (6.8.4) | |||
| Multi-path (6.8.5) | |||
| ALG & middleboxes (6.8.6) | |||
Additional Comments
(since Tokyo)
Example
| Section 3.6: Precedence for Elastic Traffic | |
| For instance, low precedence traffic should experience lower average throughput than higher precedence traffic. Several questions arise here: what is the meaning of "relative"? What is the role of the Transport Layer? | |
| The preferential treatment of higher precedence traffic with appropriate congestion control mechanisms is still an open issue that may, depending on the proposed solution, impact both the host and the network precedence awareness, and thereby congestion control. [RFC2990] points out that the interactions between congestion control and DiffServ [RFC2475] have yet to be addressed, and this statement is still valid at the time of writing. | |
| Comment: This situation is addressed by gTFRC that distinguishes between the guaranteed part from the best-effort part (green and red traffics) resulting from an AF conditioning (specified and tested such congestion control inside DCCP/CCID3 for DiffServ/AF networks) | |
| Reference: "Guaranteed TCP Friendly Rate Control (gTFRC) for DiffServ/AF Network" <http://manu.lochin.net/publications/draft-lochin-tsvwg-gtfrc-02.txt> |
| Extended comment/review period until August 15th | |
| Version 05 will be released after that period (for IRSG review) | |
| Heterogeneity | ||
| Various network environments, various protocols interacting, etc. | ||
| Stability | ||
| Fairness | ||
| flow rate vs. “cost“ fairness | ||
| Network support | ||
| Performance and robustness | ||
| Granularity of network component functions | ||
| Information acquisition | ||
| Feedback signaling | ||
| Corruption loss | ||
| Packet Size | ||
| Flow startup | ||
| Multi-domain CC | ||
| Transport of congestion signals | ||
| Information exchange | ||
| Pseudowires | ||
| Precedence for elastic traffic | ||
| Misbehaving senders and receivers | ||
| Others | ||
| RTT estimation | ||
| Malfunctioning devices | ||
| Dependence on RTT | ||
| CC in multi-layered networks | ||
| Multipath e2e CC and TE | ||
| ALGs and middleboxes | ||