Re: We don't seem to be following our processes (Re: Network Programming - Penultimate Segment Popping)

[Mark Smith <[e-mail]>]

On Sat, 7 Dec 2019, 22:43 , <[e-mail]> wrote:

> Mark,
>
> > Do you agree in principle that this approach of "header insertion in a
> controlled domain" can be made to work,
> > and should be indiscernible from A inserting the extension header itself?
> >
> > Why can't they use Link-Local Addresses to tunnel between each hop,
> i.e., between A and B, and then B and C?
>
> There are of course many alternative approaches, depending on use case.
> But that isn't an answer to the question I asked.
>
> In the example I gave, "do you in principle see any reason why it could
> not be made to work"?
>


Sure, it could be made to work. Anything that can be implemented in
software or hardware can be made to work.

However, that's a low success threshold.

It is ignoring other factors like if the mechanism is efficient (doing
things right) and effective (doing the right things).

It is ignoring whether the mechanism is the best performing for the least
cost (efficient restated).

It is ignoring how the mechanism fits within an existing architecture.

It is ignoring how obvious the mechanism's operation is, and how well it
utilises and aligns with existing, proven and well known mechanisms.

It is ignoring how easy and quickly it is to rectify faults when they
inevitably occur, a critical factor in production commercial networks.

It is ignoring how consistent the mechanism is with existing specifications
and existing knowledge.

Showing something works doesn't really prove anything other than it can be
made to work with software or hardware.

Something working well expects all these other factors are used as
judgement criteria.

In my experience there is a big gap between working* and working well.

These other "works well" factors are ones that matter the more when the
mechanism is deployed operationally, because operators can automatically
assume it works when they buy it as it has become a commercial product.

Regards,
Mark.




* This works, but I wouldn't say it works well by the above criteria.
Notice the recommendation at the end.

"A dirty trick to save a couple of IPv4
addresses on a LAN link"
https://www.ausnog.net/sites/default/files/ausnog-2018/presentations/2.10.2_Mark_Smith_AusNOG2018_Lightning.pdf


> Best regards,
> Ole
>
>
>