Skip to content

Latest commit

 

History

History
64 lines (54 loc) · 3.4 KB

ietf106-nwcrg-loops.md

File metadata and controls

64 lines (54 loc) · 3.4 KB

IETF106 NWCRG Meeting, LOOPS slot

Time: Thursday 2019-11-21, approximately 15:10–15:30

07- About LOOPS (Local Optimizations on Path Segments) (Carsten Bormann)

https://tools.ietf.org/html/draft-welzl-loops-gen-info

Notes distilled from NWCRG meeting minutes and video.

Carsten Bormann asked who had been in a previous LOOPS side meeting, about half of the people in the room (approximately 10 of 20) showed their hands. Carsten then gave a brief introduction to the LOOPS activity.

Dave Oran asked: Is reordering one of the things you are not doing? Carsten: The network might do reordering and we might have to do resequencing. Dave: Is it a requirement to take packets misordered inside the tunnel and fix the misordering at the tunnel exit? Carsten: We want to be robust against reordering in the tunnel; whether reordering needs to be fixed at the exit [resequencing] we do not know yet (Carsten mentioned that he would like to avoid having to do it).

Marie-Jose Montpetit: With or without chair hat: how can we collaborate? Carsten: First, we look at outputs from the research group; also, based on a number of assumptions, we might have specific questions where we would look for help from the RG. Marie-Jose: We indeed have strong opinions. Carsten: I guess the WG asks specific questions, that may end up being research questions, where we need research and simulations to answer them. (Carsten then gave a question about good choices for the FEC symbol size "E" that he had asked earlier in the meeting as an example.)

Emmanuel Lochin: There are many research issues here. With tunnels, you may aggregate flows and the protection may be per-flow or on the aggregate. I do not have the answer for that. You may have different types of flows and different coding schemes for them. You may end up having head of line blocking. Carsten: This is about different traffic classes? Emmanuel: ... and microflows, so you may have packets from different flows, and a packet may be missing from one of the flows slowing down another flow. Carsten: We will have a hard time identifying the different flows and so may not encounter this issue. Can you send the results you have to the list? Emmanuel: Sure; we have some results with PIE and Codel. Marie-Jose: If you do that, please copy both NWCRG and LOOPS lists.

Colin Perkins (without any hats on): You list a bunch of FEC schemes with very different characteristics. When we worked on video and audio for AVT, we started with a simple scheme and went into more complexity. Do you expect the same? You may end up having different FEC solutions for different levels of protection, etc. Carsten: We may want to dynamically change the FEC scheme that is used. Marie-Jose: Add a specific signaling to detail the code that you want to use? Carsten: Yes. Marie-Jose: For QUIC, we added some signaling. Colin: And we did for real-time media. Carsten: Feedback information might include what packets were lost or maybe just how many. Colin: You may also have ACK mechanisms that report different things depending on the FEC mechanism that is used.

Carsten then gave his view on the next steps for LOOPS: Explore the design space, until a WG is formed — hopefully in time for IETF 107. Marie-Jose: You may want the same type of interaction between NWCRG and LOOPS as there is between T2TRG and CORE.