On Tue, Aug 06, 2024 at 09:42:09PM -0700, Ted Mittelstaedt wrote: > How is it that you have so much faith in Ziply as to immediately suspect your > POS switch instead of blaming them? THAT is the real mystery! 😉
The piece-of-shit switch (which lasted 17 years, not too bad for shit) was the fifth and last thing I checked. I should have checked that sooner - testing cable wiring happened third, and that is far more difficult, even with spare cable runs to loop though. I pay Ziply $60/month for 200M/200M service - they actually provide 350M/350M (so far), though the fastest device I've got (a Chromebook) only does 300M/300M on a clear day with a tailwind. It would be nice to have a cheap cat6 dongle that tests Gbps, and emails a detailed report to me. Why does Ziply provide more than promised? Overprovision accommodates degradation and errors, costs them little, and earns friends at this early stage of Data World Domination. My fiber link is optical+filter with no optical amplifiers or electronics all the way between my fiber modem and Ziply's switching center in Tualatin, perhaps 10 optical fiber kilometers away. As a silicon chip "expert", what they do with optics and fiber welding and etched glass etalons kicks my feeble electron-pushing butt. Now I know what old vacuum tube engineers felt like when my integrated circuit designs kicked their butts. WAG calculation: assume Ziply uses an optical band between 1500nm and 1600nm infrared (vacuum wavelength, wavelengths are shorter at the same frequencies in glass fiber), with fractional dB loss between Here and There. BTW, fiber spans oceans with few repeaters; the repeaters are optical amplifiers, special nonlinear fibers "powered" by other optical wavelengths, no mid-ocean electricity involved. 1500 nanometers is 200 THz, 1600 nanometers is 187 THz, the bandwidth between is 13 THz, and a rule of thumb is that you can cram approximately 2 baud in a Hertz without complex multilevel coding. So, one $10/km graded index optical fiber (dollars per kilometer) can (theoretically) move 25 Tb/s. Optical data systems can accurately multiplex many wavelength subdivisions into that bandwidth with a dab of precision-etched glass called an etalon. The precision-etched glass etalon evolved from the same manufacturing processes as silicon chips, so I suppose I am not completely obsolete. Ziply can cram thousands of single mode graded index fibers (each feeding 10Gb/s to thousands of homes after splitting with etalons) into one armored fiber bundle, smaller than the diameter of my thumb. Compare this to Comcan't, which is slowly replacing their coax feed system with optical. However, the customer feed is still on coax, with the conversion powered by PGE (or your equivalent local power monopoly). When the local grid goes down, so does local Comcan't. BTW, I can keep my end of the optical data path running on laptops and UPS batteries and a generator, but without you clowns to swap blovating emails with, why bother? :-) Competitively, Comcan't has acres of lobbyists propping up mega-acres of obsolete technology, so they may still kill Ziply by legislative fiat. It's our job to learn the technology, the legislative landscape, and who our real friends are, and hack those into lobbyists into paralysis. Rinse and repeat, when Ziply and other upstarts are bought out by the forces of evil. It never ends, this shit. Keith L. P.S. Thanks to others for "hijacking" part of this thread into informed discussions of alternate firewalls. Useful stuff. Please hijack the titles to match, simplifying my email archive search process. -- Keith Lofstrom [email protected]
