Search results

  1. Dave Taht

    Virgin Media Discussion Thread

    No, crusader crashed. This is still beta code, I'm afraid. Tho very attractively written in rust, I am thus far too incompetent to debug it. Restarted now. I'll put a wrapper around it to restart it automatically in the morning. Error from client :eledid:3:: Expected object Serving...
  2. Dave Taht

    Virgin Media Discussion Thread

    well, by most conventional standards, the 60ms on the upload is "good enough" for most purposes, including gaming and voip, and I hope, is a common outcome for basic fiber installs everywhere. My goal was 20ms in general, but compared to everything out there, 60ms under load is amazing. You can...
  3. Dave Taht

    Virgin Media Discussion Thread

    @Rainmaker, judging from your crusader result, your upload is slightly too high. peaks at 15ms and slowly gets down to 5ms...
  4. Dave Taht

    Virgin Media Discussion Thread

    Lovely result for fiber. I like especially that smokeping is reporting less latency than the smokeping average, sometimes. Some forms of fiber allow you to get inside the request/grant loop and can actually offer less latency under load for sparse packets than otherwise. Rainmaker and a few...
  5. Dave Taht

    Virgin Media Discussion Thread

    I have now seen more than a few awful implementions of fiber, everything from "femtotrenching", to oversubscription of the gpon, to overly agressive "policing", to total lack of ipv6, etc. So I have hope at least some of your isps are doing the right things for CPE and their headends, if not...
  6. Dave Taht

    Virgin Media Discussion Thread

    @Rainmaker et al. I am kind of curious if the bufferbloat has got any better in your country? I see samknows is all over the subject ( ) , I've heard of cake on new CPE, docsis 4 stuff... but don't have an field data. I've busied myself since finishing up the mikrotik cake thing on a few other...
  7. Dave Taht

    Virgin Media Discussion Thread

    Meh. Runs out of cpu, too. Buggy. It can be improved w/cake: https://github.com/fabianishere/udm-kernel I have in general given up on ubnt, since they laid off their california tech crew and went public, and abandoned edgeos. YMMV.
  8. Dave Taht

    Virgin Media Discussion Thread

    Sometimes, it is the network's ability to carry interactive loads *at all* that is worth testing, even without an induced load. We've developed a new test leveraging irtt over here ( https://forum.openwrt.org/t/cake-w-adaptive-bandwidth/108848/3238 ) to actively measure baseline and loaded...
  9. Dave Taht

    Virgin Media Discussion Thread

    Both these results were very good. My principal kvetch is that speedtest and waveform simply do not run long enough to accurately measure bufferbloat at gig speeds.
  10. Dave Taht

    Virgin Media Discussion Thread

    great analogies @Rainmaker - is it any comfort that after you are a bufferbloat-beating-advocate for a year or two more it does get easier to hold your temper? The new speedtest.net app (and website) are finally measuring loaded latency: https://www.ookla.com/articles/introducing-loaded-latency...
  11. Dave Taht

    Virgin Media Discussion Thread

    corrected thank you!
  12. Dave Taht

    Virgin Media Discussion Thread

    i keep hoping more will measure their bufferbloat and page load times on gige fiber. It does take work, preferably with the flent tool, rather than a web browser, to stress it out (as per: https://forums.overclockers.co.uk/posts/35236152/ ) Anyway, @Rainmaker - happy new year, for starters!! -...
  13. Dave Taht

    Virgin Media Discussion Thread

    While I just tremendously enjoyed your enthusiasm, and I look forward to more data @croniclard ? Flent's comparison plots can be easier on the eyes. (Data->add other files), and you can also scale the plot to the same scale as another.
  14. Dave Taht

    Virgin Media Discussion Thread

    @Rainmaker I know that having perfectly low latency under load ( https://forums.overclockers.co.uk/posts/35236152/ ) internet is terribly addictive (I am patient zero) and an actual increase of your gaming activities might explain not seeing you here for a while. Could you email me your before...
  15. Dave Taht

    Virgin Media Discussion Thread

    I looked over boringtun. I think it would benefit from being threaded, but my rust is non-existent. I am encouraged by the quality of the "crossbeam" library I've been looking over on that front, but going from "hello world", to threaded programming in that language, over a weekend, is not...
  16. Dave Taht

    Virgin Media Discussion Thread

    I really miss netnews. What does it take to have your own server setup? leafnode? cnews?
  17. Dave Taht

    Virgin Media Discussion Thread

    I have been trying to escape my bubble and understand real problems real people have to direct my future research. So I'm glad I came, though i'd rather like to pop over to the Fleece for a cold one and a live band this evening, y'all are 6? 7000 miles away.. "If I drop the rates down again (eg...
  18. Dave Taht

    Virgin Media Discussion Thread

    I wouldn't use the term "flooding" as torrent's underlying transport has the LEDBAT congestion controller. LOTS of UDP encapsulated entering into a single AQM, perhaps. I've said elsewhere we'd made site-2-site wireguard work pretty well with fq-codel, but was unclear about saying if that was...
  19. Dave Taht

    Virgin Media Discussion Thread

    Win!!!! If you want DSCP markings to work for you on egress through the router: option squash_dscp '0' If you want to try ecn on your client, see https://www.bufferbloat.net/projects/cerowrt/wiki/Enable_ECN/ and on most OSX implementations in addition to the ones documented there: sudo sysctl...
  20. Dave Taht

    Virgin Media Discussion Thread

    De nada. But I'm not so much into feelings as measurements. :)
Back
Top Bottom