[e2e] current congestion control specs?

[Joseph Ishac]

In response to your requests:

(1)  RFC 2001 has been obsoleted by RFC 2581 (which, from what I 
understand, is geared more towards an implementor's POV)

(2)  SACK is a fairly good loss recovery scheme, and it's implementation is 
growing [See M. Allman, "A Web Server's View of the Transport Layer", ACM 
Computer Communication Review, 30(5), October 2000] so it would be good to 
include it in any implementation of a network stack.  SACK is defined in 
RFC 2018 and extended in RFC 2883.

(3)  I have not read RFC 2914 yet either, so I cannot provide feedback on 
that, but perhaps someone else can?

Hope this helps,

-Joseph
----------
Joseph Ishac
NASA Glenn Research Center
jishac at grc.nasa.gov
http://roland.grc.nasa.gov/~jishac/

At 01:03 PM 4/17/01, Mark Boolootian wrote:

>With regards to congestion control mechanisms, can folks tell me what specs
>current TCP stacks should be implementing?  I assume RFC 2001 is the
>baseline set of mechanisms.  What else could/should a protocol stack
>implement?  SACK?  Does RFC 2914 encapsulate current thinking regarding
>congestion control requirements?  (I haven't yet read it).
>
>thanks,
>mb