Re: [tcpm] Review of draft-zimmermann-tcpm-reordering-detection-02

[Pasi Sarolahti <[e-mail]>]

Hi Michael,

On 24 Feb 2015, at 23:54, Michael Welzl <[e-mail]> wrote:

> I just read https://tools.ietf.org/html/draft-zimmermann-tcpm-reordering-detection-02
> 
> Since I don’t fiddle with the TCP recovery code on a daily basis, I am a member of a large group of people who, in my opinion, can never make solid claims about the correctness of a mechanism like this one. All people like me can do is read and comment and hope for the best  :-)    because fully understanding this from simply reading, not looking at code, is REALLY hard. I don’t blame it on the authors: I appreciate that making this document easier to read must be very, very difficult.

Thanks for the comments! I have had bit similar thoughts than what you say above. TCP recovery engine with all its performance optimizations starts to be quite complex already, so it is hard to review various corner cases or interactions with other algorithms. I’m sure the authors have done thorough work in implementing and analysing the algorithm, but can the rest of us understand it sufficiently well? There are quite a few steps in the algorithm to be digested.

- Pasi