an update on me

[Francis Dupont]

> On Thu, Feb 16, 2012 at 8:30 AM, Francis Dupont <fdupont at isc.org> wrote:
> >> To expand on what Francis said, as it relates to the CPE:
> >> What we have is a modified miniupnpd. This proxies UPnP and NAT-PMP to
> >> PCP, but only over DS-Lite (it doesn't make a local NAT mapping, but
> >> just passes the PCP request to AFTR).
> 
> Where is the tree or patches? I note that I presently do *everything*
> in public, and have standardized on github as my default git
> repository.

=> we have a (private) git and we do private distribs, weekly for
Juniper and upon requests. For many silly reasons there is no uptodate
public distribs, IMHO mainly because the specs so the code are unstable.
The URL is ssh://fdupont@repo.isc.org/var/repo/git_prod/aftr/usermode
(replace fdupont by your kerberos username, of course here I assume
you have a kerberos account at isc.org). There are miniupnpc+,
a miniupnpc client with more IGDv2 support, and transmission,
a bittorrent client extended to support PCP.
BTW if you have no account just ask Paul (who is in the list)
so he will add you to the recipients for the next distrib.

> The only things that aren't public are merely because I haven't had
> time to do so.

=> yes but only the AFTR code is officially supported. I expect
this will change as soon as the PCP base spec will be published
as an RFC (in fact when it will be in the RFC editor queue).

> While I'm at this, is there a more modern AFTR than 1.1? I'd built it,
> last year, recoiled at the default nat timeout, and not played with it
> since.

=> there are a very few code changes so it is better to use the
last version.

> "b4iwf" needs some googlejuice.

=> not very bad for an IWF code designed to run on a B4, isn't it?

> >> We've got a contract with Juniper (separate from the SD-NAT work) to
> >> work on PCP, which includes a deliverable to make the CPE work in a
> >> NAT444 scenario (i.e. make a local NAT mapping, and request a
> >> corresponding mapping from the CGN). This is my next task.
> 
> To what extent must this work be 'private'?

=> in fact it is not 'private', it is more 'not public', i.e., we give
it to anyone who asks for it (and we proposed it as soon as someone
manifested any kind of interest).

> the wndr3700v1 has no 'v1' on the version string.
> 
> This has pictures of the box and back of a wndr3700v2.
> 
> http://www.bufferbloat.net/projects/bismark/wiki/Wndr3700v2

=> so the wndr3700 I have at home is really a v1 (I have no doubt
as it is old, but this is an useful info if I have to get a new one).

Regards

Francis Dupont <fdupont at isc.org>