Re: [arch-d] On the (f)utility of prescriptive architecture

[S Moonesamy <[e-mail]>]

Hi Brian,
At 08:21 AM 21-03-2020, Brian Trammell (IETF) wrote:
>True, there is no overarching prescriptive 
>"architecture for the Internet", because IMO 
>developing such a thing would be a colossal 
>waste of time. In the absence of a lever with 
>which to impose architectural will, in the best 
>case the broad and diverse community that builds 
>and runs the Internet would look at such a 
>thing, maybe read the abstract, say "yeah that's 
>nice" then go back to the business of keeping 
>things running. What is more useful is 
>descriptive architecture in the small: looking 
>at the way things are and drawing insights 
>toward evolutionary change; the RFCs cited above 
>are all examples of this sort of thing.

I read Stewart's message differently.  If I 
understood it correctly, it was about thinking 
about the long-term evolution of the Internet.  A 
person can look at the details (what is referred 
to as "in the small") if he/she has the time to 
do that.  The person might still have to look at 
the broad picture if he/she is interested in the 
evolutionary/architectural aspects.

>Most of the problems the long history of IP 
>replacements (including the ones I've worked on, 
>and continue to work on) over the years have 
>identified aren't really technical, or at least 
>are not technical in tractably addressable ways 
>(the speed of light and the fact that all 
>operations one can do in a network have strictly 
>nonnegative latency are problems as old and hard 
>as the non existence of perpetual motion 
>machines). Rather, they are issues of inter 
>organizational relationships and business 
>models: rethinking inter domain routing or end 
>to end QoS require an alignment of incentives 
>that seldom emerges organically in a multi 
>operator network. If you'd like the IAB to 
>address these authoritatively, you can talk to 
>your friendly local NomCom about appointing some 
>economists, but it's not clear to me that that's 
>the best business for the IAB to be in.

Some points (please see above) could be doable 
without replacing IAB members with economists.

There is a message from the IAB at 
https://mailarchive.ietf.org/arch/msg/ietf/AWEWI7rCmbOzifT5xQRpyciPLE8/ 
It is understandable that it has to see to the 
other aspects of its Charter.  I could not find 
anything [1] in there which takes a "let's think 
ahead" approach or a "let's rethink".

The recent events could produce a significant shift in the existing models.

Regards,
S. Moonesamy

1. There may be an exception to that.  It may 
take a few years to see whether that is actually the case.