Re: [tsvwg] Comments on draft-ietf-tsvwg-l4s-arch-03

["Scheffenegger, Richard" <[e-mail]>]

Hi Bob,


I finally found the time to closely read the l4s-arch draft.

The architecture document can be read to concern itself more with 
streaming-type loads, as those examples are very prominently mentioned.

For the uninitiated, a short paragraph discussing the interaction of 
streaming flows with burstly flows might be good.

There is some discussion around this later in the document (sections 
6.2, 8.2) though. Also, I consider the explicit mentioning of 
heterogenous data centers in the applicability section to be very valuable.



Editorial: I believe the canonical name for DCTCP is Data Center TCP 
(sec 2, point 3) although that may run against your british english 
instinct :) (Centre is used consistently throughout the document; I 
assume people are more likely to search for data center though)

Link in Mathis09 is unreachable; found a copy at 
https://www.gdt.id.au/~gdt/presentations/2010-07-06-questnet-tcp/reference-materials/papers/mathis-relentless-congestion-control.pdf

Top of page 14, two nested opening brackets in the first paragraph

Para 8.1, comma instead of dot (before ", In turn,")

s/best efforts/best effort/ (traffic)


Best regards,
   Richard