[prev in list] [next in list] [prev in thread] [next in thread] 

List:       mpls
Subject:    Re: [mpls] How many flags could be set in an NAI Flag field
From:       Tony Li <tony.li () tony ! li>
Date:       2022-04-27 15:14:32
Message-ID: 36745678-EAF3-4040-9D5E-85BDAD5D3CB0 () tony ! li
[Download RAW message or body]


Hi Loa,


> I've been thinking about "how many flags" will be set?


Predicting the future is easier once you've lived it.

> -)


> Assume that we have the possibility to set about 20 Network Action Indicators in a \
> flag field carried in a new bSPL (more than what can be carried in the repurposed \
> TTL and TC fields.


In fact, depending on the encoding, a solution could carry many more NAI.


> I assume the most packet will not carry the bSPL.
> 
> Of those that does, most will have just one flag.


This is not clear to me. It's pretty clear that ECMP is popular and is likely to \
become more popular, based on the trends that we see within data centers.  The claims \
of 5G network slicing seem dubious to me, but on the off chance that I'm wrong, we \
could easily end up carrying EL and GISS frequently.  

And this says nothing about flags that are still to be proposed.

My gut says that we could easily see a day where slicing is almost ubiquitous and \
ECMP is equally common.  So most packets would have two NAI.


> Very few will more than three flags set.
> 
> Is this an accurate assumption?


No one can answer that.


> What will be the balance between "implicit" and "explicit" ancillary data?


I'm unclear on the distinction you're making here.

Tony

_______________________________________________
mpls mailing list
mpls@ietf.org
https://www.ietf.org/mailman/listinfo/mpls


[prev in list] [next in list] [prev in thread] [next in thread]