Minutes of the 2014-11-13 IAB Business Meeting
Honolulu, Hawaii, USA
- Mary Barnes
- Marc Blanchet
- Alissa Cooper (IESG Liaison)
- Lars Eggert (IRTF Chair)
- Mat Ford (ISOC Liaison)
- Joel Halpern
- Ted Hardie
- Joe Hildebrand
- Russ Housley (IAB Chair)
- Eliot Lear
- Xing Li (via WebEx)
- Cindy Morgan (IAB Executive Administrative Manager)
- Erik Nordmark
- Andrew Sullivan
- Dave Thaler
- Brian Adamson
- Victor Firoiu
- Jari Arkko (IETF Chair)
- Brian Trammell
1. IRTF Review: Network Coding Research Group (NWCRG)
Brian Adamson and Victor Firoiu, co-chairs of the IRTF’s Network Coding Research Group (NWCRG), gave an overview of the RG’s activities. The RG is pursuing network coding, taking techniques like forward error correction and applying it at the network layer. Network coding is a method of optimizing the flow of digital data in a network by transmitting digital evidence about messages. It potentially benefits capacity, robustness, security, and energy efficiency.
Brian Adamson noted that there have been some implementations where full network coding systems have been demonstrated. Packet erasure coding started popping up in IETF Working Groups, e.g., Reliable Multicast Transport (RMT). One motivation for forming the RG was that work on the remaining challenges could be applicable to other IETF work; it is also applicable to information-centric networking and software defined networking.
Brian Adamson noted that network coding will not be a panacea; there are some places where it will make sense to use it, and some places where it will not. Additionally, there are different ways to apply it at different layers.
The NWCRG is also looking at areas for future standardization, such as common encoding algorithms, protocols, service descriptions, and packet formats. Lars Eggert noted that any standards work would have to be done in an IETF WG rather than in a Research Group. Brian Adamson agreed, saying that that work would be post-IRTF. Right now, the first step is getting the interested parties together and raising awareness about areas that can apply network coding. One project is to develop a common taxonomy and a modular architecture so that people can re-use relevant components (e.g., the RMT building block approach).
Victor Firoiu added that the RG has collected a list of use cases, some of which have implementations behind them, and they are using those as a starting place for the building blocks.
The NWCRG was chartered in November 2013, and has had around 29 research presentations over the last 4 meetings. There are three individual drafts under consideration, as well as a proposal for a network coding architecture. The RG is also planning to do a tutorial, explaining the technical details on what is network coding-capable, and what it takes to use network coding.
Dave Thaler observed that some of the work is relevant to the Transport Services (TAPS) WG, and that TAPS is working on a taxonomy similar to what RMT did. He noted that the RMT work may be useful to present to TAPS. Brian Adamson agreed.
Victor Firoiu reported that participation on the mailing list has had bursts and lulls. Erik Nordmark asked if having the RG meetings during IETF was helping get more people involved. Brian Adamson replied that he thought so, but noted that things have been at an ebb as there are not currently any really hot topics; people are trying to cash in on the easy end-to-end stuff to solve the resiliency issues. Victor added that they have been seeing between 50-80 people at their meetings.
The slides presented during this discussion are available on the IAB website at:
2. RFC Format
The IAB discussed the upcoming RFC format changes.
One concern is how lawyers will respond to subpoenas if XML is the canonical format; the present plan is to send back a PDF that is rendered from, and contains, the XML.
That opens the question of what happens if there is a bug in the rendering; does the RFC simply get re-rendered, or is it re-released with a new RFC number? Joel Halpern pointed out that if it is simply a rendering error, we do not want an outcome where there are two different RFC numbers that have exactly the same source XML. Joe Hildebrand noted that the rendered date will be added to the RFC metadata to help alleviate some of that concern.
Joe Hildebrand noted that another concern is what will happen if, at a later date, it is discovered that the production of PDF/A-D RFC forms creates a security vulnerability; the RSE is actively considering this problem.
Joe Hildebrand concluded that there will need to be policies in place detailing how certain issues are handled. Ted Hardie added that part of those policies should include who is ultimately responsible for any decisions made: the RSE, or the stream manager.
3. Cleartext Statement
The IAB briefly discussed the proposed IAB statement on cleartext. Russ Housley noted that he believes the text has converged into something all parties can agree upon; he will start an e-vote for the IAB to approve the statement.