Re: [tcpm] [Technical Errata Reported] RFC7413 (5373)

[Yuchung Cheng <[e-mail]>]

On Thu, May 31, 2018 at 7:23 AM, RFC Errata System
<[e-mail]> wrote:
>
> The following errata report has been submitted for RFC7413,
> "TCP Fast Open".
>
> --------------------------------------
> You may review the report below and at:
> http://www.rfc-editor.org/errata/eid5373
>
> --------------------------------------
> Type: Technical
> Reported by: Vladimir Nicolici <[e-mail]>
>
> Section: 4.1.3.1.
>
> Original Text
> -------------
>    For any negative responses, the client SHOULD disable Fast Open on
>    the specific path (the source and destination IP addresses and ports)
>    at least temporarily.
>
>
> Corrected Text
> --------------
>    For any negative responses, the client SHOULD disable Fast Open on
>    the specific path (the source and destination IP addresses
>    and the destination port) at least temporarily.
Looks good to me. Thanks for catching this!

>
>
> Notes
> -----
> The original language seems to imply that the cached negative response should only affect connections if they are initiated from the same source port and source IP.
>
> Since the client source port can change for subsequent TCP connections, and it's unlikely that just changing the source port would result in a successful TCP FO connection when a previous connection from a different source port failed, associating the cached negative response with the source port is probably not very useful, and could actually be detrimental to performance and reliability, depending on the implementation.
>
> If the implementation would decide to check the source port when matching negative cached responses to a new connection, it would negatively impact performance when the source port changes, because the implementation wouldn't find a matching negative response in the cache.
>
> Furthermore, if each connection retry is made from a different source port, checking the source port when matching the cached negative responses would make the client unable to connect to the server, until all possible source ports are included in cached negative responses.
>
> This means it's much better not recommending to associate the source port to the cached negative responses, to prevent any confusion and possible implementation issues.
>
> Either that, or add additional clarification, describing exactly how a negative cached response should be matched to a subsequent connection attempt.
>
> Instructions:
> -------------
> This erratum is currently posted as "Reported". If necessary, please
> use "Reply All" to discuss whether it should be verified or
> rejected. When a decision is reached, the verifying party
> can log in to change the status and edit the report, if necessary.
>
> --------------------------------------
> RFC7413 (draft-ietf-tcpm-fastopen-10)
> --------------------------------------
> Title               : TCP Fast Open
> Publication Date    : December 2014
> Author(s)           : Y. Cheng, J. Chu, S. Radhakrishnan, A. Jain
> Category            : EXPERIMENTAL
> Source              : TCP Maintenance and Minor Extensions
> Area                : Transport
> Stream              : IETF
> Verifying Party     : IESG