Re: [spring] TSV-ART review of draft-ietf-spring-segment-routing-msdc-08

[Gaurav Dawra <[e-mail]>]

Hi Martin,

Apologies from my end we had multiple discussions amongst authors.

For below comment - I will remove the citation of the really open paper. Will update in next rev:

Section will read like this now:

A flowlet is defined as a burst of packets from the same flow
followed by an idle interval.

With some ability to choose paths on the host, one may go from per-
flow load-sharing in the network to per-packet or per-flowlet.  The
host may select different segment routing instructions either per
packet, or per flowlet, and route them over different paths.  This
allows for solving the "elephant flow" problem in the data-center and
avoiding link imbalances.
 
Cheers,

Gaurav

> On Apr 9, 2018, at 8:02 AM, Martin Stiemerling <[e-mail]> wrote:
> 
> Hi Gaurav,
> 
> Am 01.03.18 um 16:01 schrieb Gaurav Dawra:
>> Hey Martin,
>> Seeking your feedback here. Regarding Section 7.1, The intent of the document is to describe one suggestive way and do not attempt to standardize any suggested mechanisms based considering the Informational nature. Authors do not attempt to fully solve – but to indicate how SR helps towards this.The details of how the hosts figure out what paths to take through the network such that the TCP and application are not affected is outside the scope of this document.
>> Would you prefer to add perhaps - “details outside the scope” or “this is one of the ways” or something else?
> 
> 
> Since the cited paper is really old and there is no up-to-date proof of any working implementation with real Internet traffic: Please remove this part.
> 
> This will require a lot of theoretical work, discussions the transport community, discussions with operating system vendors and testing in the wild open Internet.
> 
> Kind regards,
> 
>  Martin