OPEN ISSUE: Standards Track

[Charles E. Perkins]

Hello Keith,

Keith Moore wrote:

> > _nobody_ claims that AH and ESP were insecure because they
> > were standardized without key distribution.
> 
> not insecure, just useless.  and the fallback to bare IP _is_ insecure.

They weren't useless, because as Steve B. said, you could
use preconfigured keys.  Do you really claim that's useless?

> I'd also claim that IPsec implementation in hosts is useless without an
> API that allows apps to determine the validity of their peers'
> credentials...

We can't boil the ocean all the time.

First the basic headers.

Then the key distribution.

Then the APIs.

If the basic headers were never published in a stable
document, I'll bet the whole process would stall and
_nothing_ would have gotten done.

> maybe this is another illustration that we don't really try to
> understand the full scope of the problem until it gets to IESG.
> and it's hardly surprising if IESG does a poor job of fixing that.

It's equally well an illustration that we can't boil
the ocean over and over again -- especially with new
people every time.  We need to boil a few gallons first,
and then a few truckloads.  Failure to publish is a
distinct impediment in the process, and I think that
timely publication of a stable Proposed Standard should
be a good initial step.

Regards,
Charlie P.