[Cerowrt-devel] [Make-wifi-fast] my arin NRO board candidacy

[Dave Taht]

On Sun, Oct 27, 2019 at 3:18 AM Toke Høiland-Jørgensen <toke at redhat.com> wrote:
>
> Dave Taht <dave.taht at gmail.com> writes:
>
> > A while back I decided to run for ARIN's (the american registry of
> > internet numbers) NRO board, and attend their conference and election
> > next week in dallas texas.

It was in austin, actually... good music town.

> >
> > While I decided to run to discuss the ipv4 extensions project, I
> > certainly intend to raise issues of direct concern here (bufferbloat,
> > binary blobs, wif, 5g, ipv6, middlebox problems) at a pretty high
> > level and in a place I've not done so before, in front of people that
> > have never heard of them.
>
> Woohoo, go get 'em! If you ever run for the RIPE board, I'll definitely
> vote for you! ;)

Well, despite repeated attempts at emigrating to europe, I've not
succeeded. So if ever *you* run,
I'll root for you.

>
> -Toke
>

Anyway, the voting period for the ARIN election runs until nov 8th. If
anyone here is an ARIN
member and wants to saddle me with this job, the voting instructions are at:

https://www.arin.net/announcements/20191031_election/

There are three candidates up for the NRO with only one slot
available. These were my intended remarks... I have no idea what I
actually said in my speech! I ended up veering from the text
significantly, and although
it was filmed and transcribed I've not actually seen it yet.

...

Hi, I'm dave taht and I'm running for the NRO.

A bit about me: I run a little company that does wifi (anyone here use
wifi?) and embedded mostly-middlebox and home router software R&D.

I just came here from the third world... california. I was without
power for 5 days - comcast went down immediately - my t-mobile
fallback only lasted a day, only verizon stayed up, and in no case,
would my ipv6 out of my campus, have kept working.

I've been working to make the internet faster, more reliable and
resilient for a very long time, and I'd have liked it if last week's
experience had been less stressful.

I am primarily a technologist. Running for the NRO is my first
excursion into the policy making arena. In addition to the stuff on
the slides behind me, I also sit on the board of the commons
conservancy.

Anyway, perhaps the thing I'm most well known for, is for helping fix
the bufferbloat problem and the IETF AQM working group - anyone here
heard of those?

Our core bufferbloat-beating algorithm (from my perspective) -
fq_codel (rfc8290) - is now the default in most of apple's products,
and the default of nearly all the linux distributions, in BSD, and in
a ton of home routers - it's in well over a billion boxes so far, from
a starting point of zero, in 2012.

Except it's mostly the wrong billion. Trying to get vendors and ISPs,
to implement and deploy rfc8290 along their edge... and vastly improve
their latencies under load... :sigh: It's like pushing jello uphill.

But, for a change, I'm not here to talk about bufferbloat today.

The principal reason why I decided to show up for an ARIN meeting and
run for the NRO is because for the last year, John Gilmore and Paul
Wouters and Toke Hoiland-Jorgensen and I and multiple others have been
pushing out into the open source world "the ipv4 extensions project" -
converting the former class-e space - 240/4, and 0/8, and portions of
the overallocated multicast address spaces to unicast use.

The end result of that may be as many as 420 million more, new ipv4 addresses.

... and while some technical barriers remain, making new ipv4
addresses, like fixing bufferbloat worldwide, has got more than a few
policy implications that need sorting out, and it's within a scope
that crosses all the orgs with responsibility for the internet.

Before I'm pilloried for this project by the ipv6 crowd...

I've also had a longstanding interest in getting ipv6 deployed, again
working from the edge out, on consumer routers and middleboxes and
that road's been hard. I've worked on code and RFCs designed to make
the ipv6 deployment easier, notably on cerowrt, and in the IETF
homenet working group and on things like source address dependent
routing - which has seen next to no deployment as yet.

So I thought getting in here and seeing what the real problems were in
the RIRs and ARIN community, trying to understand the policies and
deployment problems actually were and to attempt to apply my
technological experience to them, might be worthwhile.

Overall... In getting involved with the NRO and ARIN, I'd really like
to help develop a more reliable, resilient, internet.

And I'd appreciate your vote.


--

Dave Täht
CTO, TekLibre, LLC
http://www.teklibre.com
Tel: 1-831-205-9740