[Cerowrt-devel] Fwd: [homenet] Routing protocol comparison document

Dave Taht dave.taht at gmail.com
Mon Mar 2 14:45:35 EST 2015

I currently plan to enable some form of ipv6 translation by default in
the next version of cerowrt - and make direct access optional - (or
the reverse! I'm easy ) if somehow we get it together enough to
actually have a way to do a cerowrt-scale effort again.

Any objections here? Suggestions for how to make one of the ipv6
translation techniques work right?

---------- Forwarded message ----------
From: Dave Taht <dave.taht at gmail.com>
Date: Mon, Mar 2, 2015 at 11:32 AM
Subject: Re: [homenet] Routing protocol comparison document
To: Juliusz Chroboczek <jch at pps.univ-paris-diderot.fr>
Cc: "<curtis at ipv6.occnc.com>" <curtis at ipv6.occnc.com>, Ray Hunter
<v6ops at globis.net>, "homenet at ietf.org Group" <homenet at ietf.org>, Teco
Boot <teco at inf-net.nl>, Mikael Abrahamsson <swmike at swm.pp.se>

On Mon, Mar 2, 2015 at 10:59 AM, Juliusz Chroboczek
<jch at pps.univ-paris-diderot.fr> wrote:
>>> If we carry NAT over to IPV6, then shame on us.
>> I am sorry, I no longer share this opinion [...]  The next version of
>> cerowrt will do translation from the external IPv6 address range to
>> a static internal one (or ones, in the case of multiple egress
>> gateways),
> (Insert strong expression of disagreement here.  Use any means available
> to convince Dave otherwise, including flattery, threats, demagoguery, ad
> hominem attacks and photographs of cute animals.)

Hahaha. Thanks juliusz! I have laughed far too little in the past few
weeks. ( just one example:

Let me make clear: CeroWrt is (or was, ENOFUNDING) a *RESEARCH PROJECT*,
IMHO *the best - and nearly the only - one - that exists*, one that
has poked into many of the very real problems billions of home
networks have.


Through it, we identified *and fixed* multiple epidemic problems so
far, including fixing bufferbloat, making dnssec deployable, and -
along with the wonderful devs over at openwrt - helped make ipv6 work
a zillion times better along the edge than it ever has before.

We have made available the code, and firmware, to a large, dedicated,
brilliant group of testers, who have all done testing, providing their
feedback on each idea, each rfc, and the good and bad ideas in both
the code and rfcs better sorted out.  And a metric ton of bugs were
fixed along the way in both the kernel and the userspace stacks.

Some of these things have fed back as requirements into this wg,
notably the need to do mdns proxying, and prefix distribution, and to
some extent, source specific routing, and (finally) you are beginning
to recognize the real problems and complexities that real home
networks have, and are beginning to grok wifi.

Regrettably, progress on multiple other fronts for CeroWrt have been
slow (no funding, not enough devs), and of all the problems I have run
into in comcast's (otherwise pretty darn good) ipv6 deployment,
getting renumbered has been the biggest PITA, followed by a couple
borked dhcpv6 implementations, and then by trying to get hnetd to work
at all - at the moment it feels like the "systemd for home routers" -
and that is *not* a complement.

To make it clear - after actually *testing* some form of ipv6
translation technology - maybe all of them - in the next version of
CeroWrt[1] - we plan to find the bugs, document the problems, and do
whatever we can to fix them - and... if those are more severe than the
problems that hnetd introduces - try to fix hnetd - with enough people
helping, at the same time - and incrementally evolve whatever we works
best into something your mom can use and home, and small business -
into something that can be more widely deployed.

I am rather upset at the whole wg for the pathetic level of actual
testing and dogfooding of what code now exists - only *1* person was
willing to to do a plugfest on my previous "modest proposal", and he -
like me - is too broke to attend the ietf, but not too broke to spend
60 bucks, 5 minutes reflashing an off-the-shelf router, and do a
little testing on it's behalf. Arguably hnetd now has the most amazing
ratio of ported code (1000s of platforms, 36 cpu architectures) to
actual users (4?) that has ever existed.

I - and my userbase - are not going to let y'all inflict the broken
ideas on my mom, or for that matter the home router industry, without
actually testing running code. Sorting out the good ideas from the bad
is why we actually build things, and test, before committing to final
versions of specifications. At least, that's how I have always

> -- Juliusz

[1] And the main reason for the next version of cerowrt has little to
do with the products of this working group - but to further develop,
and test, some MAJOR improvements to wifi we've come up with.

Dave Täht
Let's make wifi fast, less jittery and reliable again!


Dave Täht
Let's make wifi fast, less jittery and reliable again!


More information about the Cerowrt-devel mailing list