Re: comments on draft-ietf-itrace-03

It wasn't processing performance I was worried about: it was establishing the 
applicable key correctly.  The key could change between capture and processing.

Leech, Marcus [CAR:W669:EXCH] wrote:
> "Taylor, Tom-PT [CAR:5N00:EXCH]" wrote:
> 
>>I'm sorry -- I should have responded to the original note.  Most of my responses
>>would have been: "Agreed.  I'll make the necessary changes."
>>
>>I'm glad the second timestamp issue has been brought out.  I am the one who made the
>>change on my own initiative, because I foresaw the possibility of a significant time
>>lapse between when the traced packet was captured and when the the trace packet was
>>formatted and sent out (e.g. if the traced packets got sent to an off-board computer
>>for processing).  I pointed out this change to Marcus and Steve at the time, but I
>>should have sent a note to the general list.  If people don't think this is needed,
>>I'll take it out again.
>>
> 
> I think that the "time captured" is the most important data, while the second
>   provides some statistics on processing performance, it's not really useful
>   for actual ITRACE processing.  I'd say remove it.
>