Re: [multipathtcp] Comments on "draft-hoang-mptcp-sub-rate-limit-00.txt" draft

[Viet Hoang Tran <[e-mail]>]

Hi Nagesh,

Thanks for the comments, please find my responses inline below (sorry since outlook does not have a proper inline mode).

By the way, two draft repos are publicly available at:
https://github.com/hoang-tranviet/mptcp-sub-rate-limit-opt
https://github.com/hoang-tranviet/ietf-mptcp-idle-timeout

________________________________________
From: multipathtcp <[e-mail]> on behalf of Nagesh shamnur <[e-mail]>
Sent: Thursday, 11 July 2019 15:07
To: [e-mail]
Cc: Ashutosh prakash
Subject: [multipathtcp] Comments on "draft-hoang-mptcp-sub-rate-limit-00.txt" draft

Dear Hoang & Olivier,
        Thanks for taking up the topic and realizing the same as a draft.

1) section 3 mentions that, " Hosts SHOULD resend is several times, but not more frequently than once per second.". Instead more clarity can be provided on the strategy that needs to be adapted, or can explicitly mention that it is implementation dependent or to be decided later.

[Hoang] Yes, it should be discussed more. An alternative is to send at most 3 times per RTT, or attach it to a DSN. We just started with something sensible.


2) Section 3 has editorial issue, " Hosts SHOULD resend is several times, but not more frequently than once per second." needs to be changed to " Hosts SHOULD resend it several times, but not more frequently than once per second."

[Hoang] Thanks, should be fixed.


    Also do we need to consider sequence number for the SRL option, since if host sends SRL options with different rate, receiving the same out of order, would result in sending rate which is different from the expected value.

[Hoang] It seems more complex, but I am not sure this is really necessary in the real world. 
Sending too frequently different values does not look like a good behavior for me and should be restricted as above.


3) As per "draft-ietf-mptcp-rfc6824bis-18" section 3, then length field is specified in bytes and includes 2 bytes of kind of length. But in this draft 3.1, says that the length of the option is 5, it needs to be changed to 7.

[Hoang] Thanks, will fix it.


4) This draft doesn't address how this option needs to be handled on receiving side which is running older version of MPTCP stacks and doesn't implement this draft.

[Hoang] Old stacks should ignore unknown options, as stated in RFC6824(bis). This is a general requirement to deal with  new options.



Regards,
Nagesh S

Date: Tue, 9 Jul 2019 09:04:48 +0000
From: Viet Hoang Tran <[e-mail]>
To: multipathtcp <[e-mail]>
Subject: [multipathtcp] Fw: New Version Notification for
        draft-hoang-mptcp-sub-rate-limit-00.txt
Message-ID:
        <[e-mail]>

Content-Type: text/plain; charset="us-ascii"

We submitted two drafts on  Subflow Rate Limit Option and  Inactivity Time Option (https://tools.ietf.org/html/draft-hoang-mptcp-inactivity-time-00)

Feedbacks and suggestions are more than welcomed.
Thanks,

Hoang & Olivier

________________________________________
From: [e-mail] <[e-mail]>
Sent: Monday, 8 July 2019 16:28
To: Viet Hoang Tran; Olivier Bonaventure; Olivier Bonaventure
Subject: New Version Notification for draft-hoang-mptcp-sub-rate-limit-00.txt

A new version of I-D, draft-hoang-mptcp-sub-rate-limit-00.txt
has been successfully submitted by Viet-Hoang Tran and posted to the IETF repository.

Name:           draft-hoang-mptcp-sub-rate-limit
Revision:       00
Title:          Multipath TCP Subflow Rate Limit Option
Document date:  2019-07-08
Group:          Individual Submission
Pages:          6
URL:            https://www.ietf.org/internet-drafts/draft-hoang-mptcp-sub-rate-limit-00.txt
Status:         https://datatracker.ietf.org/doc/draft-hoang-mptcp-sub-rate-limit/
Htmlized:       https://tools.ietf.org/html/draft-hoang-mptcp-sub-rate-limit-00
Htmlized:       https://datatracker.ietf.org/doc/html/draft-hoang-mptcp-sub-rate-limit


Abstract:
   This document defines a new MPTCP Option that enables hosts to
   request their peers to limit the maximum transfer rate on a per-
   subflow basis.

_______________________________________________
multipathtcp mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/multipathtcp