Re: [arch-d] possible new IAB programme on Internet resilience

[Brian E Carpenter <[e-mail]>]

On 30-Dec-19 12:40, Toerless Eckert wrote:
> On Mon, Dec 30, 2019 at 08:38:09AM +1300, Brian E Carpenter wrote:
>> However, resiliency doesn't seem to me to require a discussion of values. It's a very clear technical requirement on its own. Surviving human error, equipment failure, natural disasters, political interference, and acts of war might each generate different resiliency considerations.
> 
> Germany seems to think that its good for resilience and cost reduction
> to be able to source oil from all places it can, including russia (nordstream 2).
> The USA claims values against that and is hence fighting it.
> 
> To make a more technical _networking_ resilience argument: In design of
> resilent network paths, one key ask is to have non-single point of failure
> in forwarding/control-plane code-paths, aka: diverse router vendors paths.
> But there are also government regulations based on "values" requiring
> network paths not to pass through specific vendors equipment or certain countries.
> 
> Aka: You are grasphing for straws if you try to declare something to be
> beyond being abused as an argument about values. And remember that
> (plastic) straws too are outlawed in certain countries because of values.

True. But that doesn't prevent us doing the analysis of resilience issues
without making political assumptions. If it comes out that the best form
of resilient routing would be some form of Byzantine routing with no fixed
paths between end points, that would be a valid technical conclusion.
(Whether that particular approach has been seriously analyzed since about
1962, I don't know.) If we could design a resilient unambiguous namespace
without a central authority, that would also be a valid technical conclusion
that might prove troubling to ICANN.

   Brian