RE: I-D ACTION:draft-kompella-tewg-bw-acct-00.txt

[Vadim Suraev <Vadim_Suraev@KereniX.com>]

Hi, Kireeti.
I have few comments/questions to the draft:

1). It seems that a case of link bundling is not yet addressed in the draft.
Should maximum normalized BW of bundled link be the max of component links'
normalized maximum BWs ?

2). My assumption is that normalized maximum BW of bundled link is max of
normalized maximum BWs of component links.
If it is correct, I think that a node, which is setting up a flow, cannot
predict a new value of the maximum normalized BW.
Further more, even if the new value of unreserved BW can be predicted, the
node cannot deterministically calculate a path for another flow
before a link state update was received.

3). Regarding to the paragraph 5.2 Analysis, I think it is not clear, to
which values of  unser[0] & unres[7] the following example is applied:
"Another drawback is that Z cannot predict whether setting up a given flow
will cause preemption or not. In the above example, say that Z wants to set
up a 300Mbps flow. In case 1, this will require that the priority 7 flow be
preempted; in case 2, this flow fits without preemption. If Z doesn't want
to cause preemption, Z cannot tell whether it should use link L or not." 
Is it applied to the case 1 when unser[0] == 400 & unser[7] == 200 ?

Best regards,
  Vadim.

KereniX                           	    Data-Aware Optical NetworksTM
----------------------------------------------------------------------------
------
Vadim Suraev 			http://www.kerenix.com     
Phone: +972-3-9050830	                Mobile: 972-53-568711
Fax:  +972-3-9347720		vadim_suraev@kerenix.com 
70 Pinsker St., POB 10053,   Petach-Tikva 49221 ISRAEL
----------------------------------------------------------------------------
------