> On Feb 1, 2024, at 12:33 PM, Les Ginsberg (ginsberg) 
> <[email protected]> wrote:
> 
> Acee -
> 
>>> ---
>>> //Acee
>>> 
>>> A bigger issue from that historical artifact is that some text refers to 
>>> "LSP
>> Burst Window" when they should refer to the "new" (from -01) "Receive
>> Window sub-TLV". That's a bigger issue as "in letter" this may be seen as
>> technical change (even if "in spirit" the Flow Control Receive Window 
>> logically
>> refers to the Receive Window sub-TLV)
>>> This requires the following changes in the 6.2.1 "Flow control" section:
>> 
>> I see that the changes are already in place in -06. Note that I think you 
>> should
>> rename “Receive Window” to “LSP Burst Window” for consistency.  In reading
>> the guidance in section 6.2.1, It seems to me that this should have been “LSP
>> Receive Window” all along. We are not changing the semantics of "LSP Burst
>> Window" or "LSP Receive Window”.
>> 
>> I’d be interested in what the other co-authors think (especially Les 😎).
> 
> [LES:] It seems to me you are confusing "Burst" and "Receive".
> 
> When we discuss fast-flooding we tend to focus on the "major incidents" e.g., 
> a node with 1000 neighbors goes down - we are likely to need to flood 1000 
> LSPs.
> "Burst" isn't very useful here because the Burst Size (was Burst Window) is 
> small compared to the total number of LSPs that need to be flooded as quickly 
> as possible.
> 
> But far more common are the "minor incidents" where a single link goes down - 
> where a small number (typically 2) LSPs need to be flooded - or a single node 
> with a modest number of neighbors goes down where the number of LSPs which 
> will be flooded is still modest (5 or 10 or 20). In these cases, being able 
> to flood a "burst" is worthwhile because it means we can flood all the LSPs 
> associated with a topology change more quickly - meaning fewer SPFs need to 
> be executed. This notion was first highlighted 20 years ago when 
> "fast-convergence" work was done.
> 
> But the "Burst Size" is certainly expected to be significantly smaller than 
> the "Receive Window" - and the latter logically includes the LSPs received as 
> part of a Burst.
> 
> So I don’t see that what you are suggesting makes sense.

I’m talking about -06 changes to section 6.2.1.1 and 6.2.1.2 to reference 
“Receive Window” rather than “LSP Burst”. Please look at these. 


Thanks,
Acee



> 
> Let me know if I have misunderstood your point.
> 
>   Les
> 
>> 
>> Thanks,
>> Acee
>> 
> _______________________________________________
> Lsr mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/lsr

_______________________________________________
Lsr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/lsr

Reply via email to