On Dec 31, 2015, at 12:13 PM, Leif Johansson <leifj(_at_)mnt(_dot_)se> wrote:
Skickat från min iPhone
31 dec. 2015 kl. 18:04 skrev Jared Mauch
<jared(_at_)puck(_dot_)nether(_dot_)net>:
On Dec 31, 2015, at 5:00 AM, Leif Johansson <leifj(_at_)mnt(_dot_)se> wrote:
On 2015-12-31 06:16, Patrik Fältström wrote:
Is this connected to the fact that not even people developing standards
use very same standards?
The problem is that we don't have enough dogs on our tasting panels: we
need to get more ops folks directly & actively involved in the IETF.
Examining some of our success stories I suspect we'd find extensive
involvement from operations in every case.
Sadly that’s not the case here.
The reason we (as an operator) can’t use BCP-38 is the vendor hardware can’t
do it at line-rate and the performance hit is too much to sustain.
Sorry to disappoint. We even dropped it as a requirement in 2015 because it
was clear the gap was getting wider not narrower.
Happy to discuss in person in BA or anywhere else we end up at the same time.
- Jared
I dunno if that is true or not but in either case you just proved my point.
That’s not to say I’m not in favor of BCP-38, it’s just that we are often the
wrong place to do the drops as well, spoofed packets should get addressed.
It’s often hard to describe topologies and while RPSL might be a technique to
do it, and it’s a standards based way to describe these things, it’s certainly
not commonly used.
- Jared