1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Virgin Media Discussion Thread

Discussion in 'Networks & Internet Connectivity' started by macca40, 7 Mar 2007.

  1. Rainmaker

    Sgarrista

    Joined: 18 Aug 2007

    Posts: 9,070

    Location: Liverpool

    I'm running a flent tuning session as we speak. The results so far are... probably worthy of their own thread. :eek: Maybe you can pop back later in the week when I have some real data.
     
  2. uvarvu

    Mobster

    Joined: 29 Apr 2004

    Posts: 4,355

    Location: Bath

    Just ran the tool on a VM 1gig service. A medium responsiveness is not ideal right?
    @ChrisD. Can you run on yours out of curiosity?

    EDIT: Cleaned up output and @Rainmaker Can we see your output too please as you’ve set configuration on your router to combat bufferbloat?

    Last login: Mon Nov 8 18:42:38 on ttys000
    [email protected] ~ % /usr/bin/networkQuality -v

    ==== SUMMARY ====
    Upload capacity: 44.792 Mbps
    Download capacity: 894.068 Mbps
    Upload flows: 20
    Download flows: 16
    Responsiveness: Medium (942 RPM)
    Base RTT: 21
    Start: 08/11/2021, 18:43:27
    End: 08/11/2021, 18:43:42
    OS Version: Version 12.1 (Build 21C5021h)
     
    Last edited: 8 Nov 2021
  3. Rainmaker

    Sgarrista

    Joined: 18 Aug 2007

    Posts: 9,070

    Location: Liverpool

    While most of my devices run Linux or BSD, I do have a MacBook Pro. Unfortunately, it's old (mid-2012) and due replacement. I'm stuck on Catalina, and as far as I can see the networkQuality tool is only present in Montery. I'm searching for how to grab the binary as we speak, and I'll update if/when I find it. For now, here's my dslreports speedtest with fq_codel enabled (OpenBSD), with cake/bbr running on the desktop (Linux) end:

    [​IMG]

    That's a huge improvement over what I saw before, latency wise. I can't get rid of the sawtooth on the upload, but I assume that's something to do with VM/DOCSIS 3 on the upstream. I'll keep playing/testing/tweaking. I'm going to move the edge router back to Linux for a while at some point (eg openwrt as Dave suggested) and see how that plays out.
     
  4. ChrisD.

    Caporegime

    Joined: 20 Sep 2006

    Posts: 27,428

    Code:
    ==== SUMMARY ====                                                                                         
    Upload capacity: 72.064 Mbps
    Download capacity: 845.084 Mbps
    Upload flows: 12
    Download flows: 12
    Responsiveness: High (4507 RPM)
    Base RTT: 8
    Start: 08/11/2021, 19:40:40
    End: 08/11/2021, 19:40:50
    OS Version: Version 12.0.1 (Build 21A559)
    
    I'm not long home so I don't know if anything is downloading or uploading, my Veeam weekly backups run on a Sunday from memory so they might still be offloading to B2. I can check again tomorrow when I start work as I know that the connection should be quiet.
     
  5. uvarvu

    Mobster

    Joined: 29 Apr 2004

    Posts: 4,355

    Location: Bath

    Wow, the difference is pretty stark. I don’t have fibre here unfortunately and it’s not even on the roadmap where I live :(
     
  6. Rainmaker

    Sgarrista

    Joined: 18 Aug 2007

    Posts: 9,070

    Location: Liverpool

    I've fixed the upstream sawtooth and latency, got a solid >50Mbps upstream with basically 0 bufferbloat now. :D Still tweaking.
     
  7. ChrisD.

    Caporegime

    Joined: 20 Sep 2006

    Posts: 27,428

    Responsiveness explained on the Apple site:

    https://support.apple.com/en-gb/HT212313

    Handy little tool (if you have a Mac!).
     
  8. uvarvu

    Mobster

    Joined: 29 Apr 2004

    Posts: 4,355

    Location: Bath

    Just had a read and there are links to profiles for iDevices too. I installed a profile on my iPad and get a result of 994 on there.
     
  9. ChrisD.

    Caporegime

    Joined: 20 Sep 2006

    Posts: 27,428

    Wifi is something I need to sort in my house, I have the nanoHD and the performance seems to have dropped quite a bit in recent weeks. It's just awkward to arrange a planned outage to do some testing.:cry:
     
  10. Rainmaker

    Sgarrista

    Joined: 18 Aug 2007

    Posts: 9,070

    Location: Liverpool

    Tell me about it! :D When I switched from Unifi (AC Pro) to Ruckus (R710), I had to issue notice in advance, plan the outage for outside 'business hours' (once the kids were asleep) and still pre-warn the wife again before it went down temporarily. It's worse than being at work! The WiFi complaints disappeared after that, though, so it was very worth it. Our eldest's room was a dead zone with the Unifi, but with the Ruckus he can watch anything in UHD and not skip a beat, so mission accomplished.

    As for bufferbloat... I'm finding it's like overclocking a CPU. I can get the upstream to 0 bufferbloat easily, with good upstream speeds too. But when you adjust the upstream, downstream changes and vice versa. It's like balancing a seesaw (i.e. like overclocking a CPU). So far I've progressed from A rated bufferbloat (on DSLReports) to an A+. I even got idle bloat down to almost nothing, but download then goes up (not ridiculously). If I get it so up and down stream have almost no bloat under load, idle shoots up to about 160ms... Maybe that's better, as it's under usage where it counts. I'm still playing, got the bug now... I'll be up all night at this rate lol. No joy on the Apple networkQuality for Catalina, though. :( I might have to patch it and upgrade (again)...
     
  11. Dave Taht

    Associate

    Joined: 6 Nov 2021

    Posts: 33

    Location: Bristol

    You are on pfsense?

    0 bufferbloat really isn't a goal, you need some "Good queue": https://queue.acm.org/detail.cfm?id=2209336

    secondly ack traffic on a very asymmetric connection such as yours can be really problematic, which is why cake has an ack-filter: https://blog.cerowrt.org/post/ack_filtering/

    thirdly, cable has framing issues, not so bad as DSL, but it's one reason why it's very hard to get good rate close to the actually subscribed rate. cake has a "docsis" mode which lets you (on egress) get to within a small fraction (.01%) of the cablemodem rate. Otherwise we recommend about 95% below the configured rate on the up, 85% on the down - and then to get to bed!

    Lastly the test that (after we've settled on nearly good values with tcp_nup and tcp_ndown) we aim for good results on is the rrul test, which is a bittorrent-like but even nastier test that tests both up and down and various forms of ping at the same time. You might have thought you've tuned it up properly, but after running the rrul you will see some sideffects both of framing and ack traffic and have to bump both down and up down.

    I am very interested in your "before" and "after". Also most likely, all your down bufferbloat is now on your wifi, and I have no idea if ruckis is shipping fq_codel or not. Get 50+ feet from the AP, through a couple walls, and repeat your tests from a laptop, but not outdoors in the middle of the night.

    To try and comment on the gigbit down cable results, the responsiveness test result circa 1000 pretty much proves there isn't pie or any form of smart queue management there, and if you are inspired by rainmaker's efforts, slam an opewrt/dd-wrt/etc x86 box in front of it, and enable cake. A tcpdump packet capture can be more revealing.

    The fiber test... not sure... 4000 is a pretty good result, partially due to the lower baseline RTT of fiber, partially due to having smaller buffers by default than most cable modems, again a tcpdump and rtt plot from this test would show what the underlying latency under working conditions was. It is possible to do much better than 4000.

    Whenever you are next at a friend's house, or a shared wifi space like a cybercafe, try this test there. When I first got out of the house this year and went to a busy coffee shop (not starbucks which at least in these parts has google wifi with all these fixes in it), I was seeing rpm values of 25-45. That's not thousands, or hundreds.... I gave that coffee shop one of my routers in exchange for a few pastries.

    The idle bloat is usually an artifact of powersave. Not much you can do about that, and do you care what your car does when it's parked?

    I'd like (someone else!) to produce an open source version of that client with some more features like attempting to isolate where the bloat is coming from - the wifi, the isp, or deeper in the network. Not everything wrong with the network is the ISP's fault. The current version of the apple network quality specification is here: https://github.com/network-quality/.../master/draft-cpaasch-ippm-responsiveness.txt

    And apple's server side code is here: https://github.com/network-quality/server which will let more people test their backend services. There's a ton of bloat in the cloud, also. A lot of users of microservices have thus far totally missed the benefits of sch_fq in linux, BBR, and for that matter, something as basic as using TCP_NOTSENT_LOWAT.

    I note that I did get banned for spam earlier, and I am now, kind of doing SEO on these new efforts by apple, as they have only been released for a week or two, and ultimately we have a whole internet to fix. I hope the moderators don't mind.
     
    Last edited by a moderator: 9 Nov 2021
  12. Rainmaker

    Sgarrista

    Joined: 18 Aug 2007

    Posts: 9,070

    Location: Liverpool

    I'm on Ethernet Dave, not WiFi. That comment was a throwaway about the wife/kids, who use it for their i-things. All my testing is gigabit Ethernet through an OpenBSD router. About the best I managed to far was 23ms idle 24ms load on upload (DSLReports) with downstream being closer to 25ms idle and 35ms load. My flent graphs are instructive, as they gave me an idea where to start. Something for tomorrow, as now I really should sleep(!). Thanks again - I'll be putting openwrt on a box and testing that with cake asap, and I'll be happy to send the flent/ruul and other stuff you want to see as and when I can get them. If you want to send a short list of graphs you're interested in seeing, I'll do my best.

    For now, bufferbloat has gone from horrendous (grade C on DSLR) to A+ so it'll do for tonight. :p
     
  13. Dave Taht

    Associate

    Joined: 6 Nov 2021

    Posts: 33

    Location: Bristol

    heh. Here's another tool for you all from the bufferbloat effort: irtt (it's available via repo from many sites src is at: https://github.com/heistp/irtt.git ) and is written in go so it works on nearly everything. I've been using it lately to construct really high resolution "maps" of starlink and lte services, at a 3ms resolution via the json output. a 20ms resolution is a decent simulation of voip.

    flent will use it instead of ping for a variety of tests if it's in the path.

    Since I've heard complaints of y'all thinking your isp may be optimizing for ping (ICMP), irtt being udp based can tell a better truth. Me being me, I plot the output of every udp ping on a cdf, but the summary report is good. I have a flent server in london I can share but it's very old and I just noticed it wasn't running irtt.
     
  14. Dave Taht

    Associate

    Joined: 6 Nov 2021

    Posts: 33

    Location: Bristol

    I updated the london server for flent and irtt. you can run flent or irtt sessions against london.starlink.taht.net now, and I'll add the networkQuality backend in when I get time.

    In testing it did 4GB/sec inside the DC, YMMV.
     
  15. Ayahuasca

    Capodecina

    Joined: 23 Apr 2014

    Posts: 24,522

    Location: Huem

  16. Avalon

    Soldato

    Joined: 29 Dec 2002

    Posts: 6,769

    For anyone else going the volt route, it seems to be an automated process now. It’s stated as taking ‘up to 14 days’, but I ordered late on Sat 30/10 and have just had emails confirming the upgrade to Gig1 from M500, in my case I need to wait 7 more days for a new Super Hub, so hopefully a shiny new SH5 lands on the doorstep :)
     
  17. Rainmaker

    Sgarrista

    Joined: 18 Aug 2007

    Posts: 9,070

    Location: Liverpool

    For anyone considering it, I'd highly recommend OpenWRT. I made the switch from OpenBSD based on Dave's recommendation and the difference is night and day. I'd forgotten just how snappy websites could be. :eek: (Edit: That's because of setting up SQM with cake, not because OpenBSD is rubbish at routing, to be clear!).

    My bufferbloat score is A+ (up from D) and I've lost hardly any downstream or upstream bandwidth. My flent graphs look way better now as well, with a nice uniform ramp up from 20ms idle (thanks to @Dave Taht for use of the London server!) to mid 20s under load and no more spikes past >70ms as bandwidth progressed. Graceful, and a nice smooth, fast ram up and down in speed.

    Swapping from integrated Intel NICs (I211 and I219-V iirc) to dedicated pci-e Intel I350-T2v2 probably helped a bit as well, but OpenWRT really is a fantastic bit of kit. Easy to work with in the GUI and full Linux underneath, so easy for us terminal heads to configure by hand where preferred. I just have to get Unbound set up to run DNS over TLS to upstream now (I did, but it broke dnsmasq's DHCP component... doh) and I'm golden. If anyone's not sure how to set it up then please feel free to drop me a PM.
     
  18. shiver

    Wise Guy

    Joined: 13 Jun 2007

    Posts: 1,172

    Location: London

    Good to hear it - I am hoping the wait for them to allow for o2 account in same household but not same name will also make the chance of SH5 higher
     
  19. zia

    Hitman

    Joined: 29 Jan 2007

    Posts: 667

    So Got 1 gig installed last week currently using sh4.in router mode

    [​IMG]

    [​IMG][/url]

    will switch asus ac88u at some point,wish it supported openwrt,

    Speed increase is fantastic although latency hasn't improved

    Zia
     
  20. iMacMart

    Capodecina

    Joined: 8 Jun 2006

    Posts: 11,759

    Location: Hertfordshire

    Can I ask what you paid please