[NNagain] FCC - delete, delete, delete

[Robert McMahon]

> so it should be proprietary code instead?

Mostly RTL that can't be modified by a black hat. ASICs for ethernet
switches come to mind. 802.11 can be done this way too. A Sun
workstation isn't required.

https://www.youtube.com/watch?v=IujKrI3BlOM

"In the context of semiconductor manufacturing, Tapeout is a
significant milestone in the production of integrated circuits (ICs).
The term originates from the era of reel-to-reel magnetic tape, which
was used to store design data. During the tapeout process, the final
design data of the IC is sent to a semiconductor foundry for
fabrication. This data is stored on a magnetic tape, hence the term
'tapeout'.

The tapeout process is a critical phase because it signifies the
transition from the design phase to the production phase. Any errors
or issues in the design must be identified and corrected before
tapeout, as changes after this point can be costly and time-consuming.
So, it naturally involves rigorous checks and validations to ensure
the design is error-free and ready for production.

The importance of the tapeout process in semiconductor manufacturing
cannot be overstated. It is the culmination of months, or even years,
of design work. The quality of the tapeout process directly impacts
the success of the IC in the market."

Bob

On Thu, Mar 13, 2025 at 4:59 PM David Lang <david at lang.hm> wrote:
>
> On Thu, 13 Mar 2025, Robert McMahon via Nnagain wrote:
>
> > My opinions:
> >
> > There should be no more linux kernels in the customer premise with
> > Fi-Wi. 30M lines of code and 11,000 config options is a form of sw
> > bloat that's impossible to secure. Particularly since most noone is
> > getting paid for this work.
>
> so it should be proprietary codde instead?
>
> David Lang
>
> > Reducing the radio head/client (STA) density to near 1/1 and shrinking
> > the cell size will minimize the media access latency. Packet latency
> > can use non queue building techniques so there will be no substantial
> > packet queueing delays. All delay will be distance and speed of
> > photons related per physics & spacetime.
> >
> > Our issue isn't regulators - it's that white collar workers and our
> > leadership haven't engaged the blue collar workers, and we haven't
> > kept advancing our engineering. We need to teach fiber installer
> > businesses how to build these Fi-Wi networks so that our kids get life
> > support and productivity capable networks that can be depended upon.
> >
> > And everyone that adds value needs to be paid somehow. Best done
> > through markets. Fi-Wi creates high paying jobs in the trades for in
> > premise fiber installers.
> >
> > I think we lack vision and leadership, followed by execution. It's not
> > a cult thing like Musk's failed prophecies - it's the real deal that
> > impacts our lives. Low latency will become ubiquitous if we act to our
> > abilities. Waiting on regulators is like Waiting for Godot.
> >
> > Bob
> >
> >
> > On Thu, Mar 13, 2025 at 10:00 AM Frantisek Borsik
> > <frantisek.borsik at gmail.com> wrote:
> >>
> >> Hey Bob,
> >>
> >> I don't think that improving latency is about mandating of a specific algorithm - it's about an improvements to broadband definition.
> >> Broadband that servers the needs of us all today, goes beyond 100/20, it's should include a low latency, low consistent jitter.
> >> Now, what are the right numbers, that's another discussion. But it's a discussion we need to have.
> >> I would certainly let market to decide on the tools/algorithms that will achieve those numbers - be it a Quality of Experience middle box (like LibreQoS, Preseem, Bequand/Cambium Networks QoE, Paraqum or Sandvine), L4S etc.
> >>
> >> As for the other issues that need some love - for example, making vendors to update kernel and provide updates to routers they sold, that's a good thing.
> >>
> >> All the best,
> >>
> >> Frank
> >>
> >> Frantisek (Frank) Borsik
> >>
> >>
> >>
> >> https://www.linkedin.com/in/frantisekborsik
> >>
> >> Signal, Telegram, WhatsApp: +421919416714
> >>
> >> iMessage, mobile: +420775230885
> >>
> >> Skype: casioa5302ca
> >>
> >> frantisek.borsik at gmail.com
> >>
> >>
> >>
> >> On Thu, Mar 13, 2025 at 5:33 PM Robert McMahon via Nnagain <nnagain at lists.bufferbloat.net> wrote:
> >>>
> >>> >
> >>> > As for "what the FCC can do", "dissolve itself" comes to mind. AFAIK, it's been over a decade since they have done anything helpful or useful for any American citizen who isn't the head of a major corporation. If you delete the entire organization, there will be no one around to enforce whatever regs are still on the books so who cares? ... and you'll save another few 10's of millions of dollars annually which will fit nicely in the pockets of the "good folks", aka FODT. 😊😊😊
> >>> >
> >>> I worked with a CA state regulator in a tech support role prior to
> >>> so-called broadband (actually, internet access beyond dial-up MODEMs)
> >>> This was post 1996 telco act, just prior to the dot com bubble. The
> >>> lobbyists at the time disliked having 50 States regulating things.
> >>> They pushed made it so the 5 commissioners on the FCC became the
> >>> primary regulators. Many call this regulatory capture.
> >>>
> >>> Unfortunately, I don't think we can get rid of the FCC. Our utility
> >>> poles are mostly regulated by them as one example.
> >>>
> >>> I also don't think the FCC can mandage any specific AQM algorithm.
> >>> That's a long term disaster in the making for sure. Let network
> >>> engineers and the market battle that out.
> >>>
> >>> Bob
> >>>
> >>> PS. Good to hear from you RR - i hope all is well. I've got a Fi-Wi
> >>> project you may be interested in - not sure.
> >>> _______________________________________________
> >>> Nnagain mailing list
> >>> Nnagain at lists.bufferbloat.net
> >>> https://lists.bufferbloat.net/listinfo/nnagain
> > _______________________________________________
> > Nnagain mailing list
> > Nnagain at lists.bufferbloat.net
> > https://lists.bufferbloat.net/listinfo/nnagain