Speed Overkill: How ISPs Sell You Bandwidth Instead of Performance (+ Bonus Free ISP Monitor Tool)

Поделиться
HTML-код
  • Опубликовано: 28 май 2024
  • In this video I demonstrate how ISPs sell you more bandwidth than you actually need, when what you're really looking for is a reliable network which carries your traffic with minimal latency and loss. I dive into the three main factors which make up connection quality (bandwidth, latency and loss), explain why speedtests are almost useless, and show you how to estimate how much bandwidth you actually need.
    Finally, I preview Netprobe, a simple and effective tool I wrote for monitoring your internet performance at home (and totally free and open source!)
    Useful links:
    How Speedtest.net Works - • Learn How Speedtest.ne...
    Bandwidth Estimator from OneRing - oneringnetworks.com/bandwidth...
    Netprobe - github.com/plaintextpackets/n...
    Netprobe Tutorial - • Free ISP Tracking in 5...
    Support the channel:
    buymeacoffee.com/plaintextpm
    Discord:
    / discord
  • НаукаНаука

Комментарии • 72

  • @thomasimmanuel4989
    @thomasimmanuel4989 Месяц назад +11

    Very informative. I work for the ISP in my country, and I usually just use speed tests when attending to customers' complaints. I learned something new; now I know how to approach customers' speed-related issues. keep it up

  • @Frozkadurrr
    @Frozkadurrr Месяц назад +17

    If u keep this up this channel will be one of the biggest networking/hacking channels in the future. Quality stuff mate!

  • @juancampos352
    @juancampos352 19 дней назад +2

    Certainly, your videos help me a lot because I'm studying networking but in my city there are not many ISP providers, there may be 2 or 3. Today, I have a 40 Mbps connection and cannot get the maximum download capacity when I try downloading a game such as GTA V, but at least I pay less money than the other ISP I had. With the first ISP I paid like 30 dollars, and they gave me 10 Mbps, now I pay 13-15 dollars for 40 Mbps.

  • @whatwhat-777
    @whatwhat-777 27 дней назад +2

    Man, why didn't I stumble upon this channel before. I am subscribed now.😊

  • @truckerallikatuk
    @truckerallikatuk Месяц назад +5

    I have way more down bandwidth than I need, and yet I still bought an upgrade... for more upload. Perils of having a lot of data that needs an offsite backup.

    • @plaintextpackets
      @plaintextpackets  Месяц назад +1

      I wish they would let you choose download and upload independently 🫤

  • @hesona9759
    @hesona9759 Месяц назад +4

    Best computer networking channel, keep it up 💯

  • @iko359
    @iko359 Месяц назад +1

    Thanks, wonderful video and job!

  • @MrT0mer
    @MrT0mer Месяц назад +2

    Love your content!!!

  • @TeslaMaxwell
    @TeslaMaxwell Месяц назад +2

    Great content and great tool sir!

  • @AmCanTech
    @AmCanTech Месяц назад +1

    Cool, love the tool!

  • @kevvyg04
    @kevvyg04 Месяц назад +3

    Quality not quantity….. its all we want 😀

  • @thevivariumforhalfmeasures7698
    @thevivariumforhalfmeasures7698 Месяц назад +1

    Latency is also a function of the number of hops and fibre optic cable issues.

  • @UserHandle0702
    @UserHandle0702 Месяц назад

    Great Content !!

  • @KyleWood1985
    @KyleWood1985 Месяц назад +2

    Seems like most of this would apply to a managed circuit that guarantees speed (COR) and includes a SAL. For home internet, don't most contracts state best effort?

    • @plaintextpackets
      @plaintextpackets  Месяц назад

      I believe they do but you should still be able to tell which ones have the best performance, if you were thinking of switching or trying others in parallel

  • @harryjohnson615
    @harryjohnson615 Месяц назад +2

    Speec vs bandwidth, here's how I expain it to customers.
    I sell you a lorry that does 60Mph and carries a million bricks and you're very happy until you find out I sold your neighbour a lorry that does 60Mph but carries 100 million bricks for the same price. Which lorry is faster? They both do the same speed so it must be a capacity difference. It's a very subtle but distinct difference.

  • @niksatan
    @niksatan Месяц назад +4

    not to self - been there when channel was small

  • @derickasamani5730
    @derickasamani5730 Месяц назад +3

    If you need guarantees, ask for a business or leased line

  • @chrisbarnard5741
    @chrisbarnard5741 Месяц назад +2

    Is there a windows version of Netprobe? I don't know anything about Dockers

  • @user-rf9nm8xd6e
    @user-rf9nm8xd6e Месяц назад +2

    Nice Video

  • @scotmaciver
    @scotmaciver Месяц назад +8

    You sir, are me. If I made a RUclips video. I've been arguing the bandwidth/speed deal but hard with all the 'experts'. 20 year sr network engineer and I have 50mbps. Would have 25 but Xfinity upped all their bottom feeders to 50. I can run production with web services, data pulls, tunnels, etc...over 3k servers on the backend serving up, all on 500mb but ole lady Betty Lou next door needs 1gig?
    Gonna carry this video around with me. Just need the first 30 seconds

    • @plaintextpackets
      @plaintextpackets  Месяц назад +4

      High praise! Thank you so much. I should turn the first 30s into a short

    • @TSPhotoAtlanta
      @TSPhotoAtlanta Месяц назад

      @@plaintextpackets if you do, put an +enticing+ overlay on the last 10 or 15 seconds to hype the full length feature. Nice discussion, BTW. Consider we have Gigabit fiber. It only goes down if our router goes down. We had a cable internet vendor, C****st - the big one, whose service delivered ~ 40 to 100Mb/s, once peaked @ 150. Cable cost more than Gigabit fiber. It shows peak delivery of 950Mb/s. Packet loss is typically approaching->0. We never saturate the service, usually not close. OTOH, it is so snappy! It comes w/ HBO-Max Premium. Subtract the cost of that and the price == 500Mb/s fiber service, so that we could saturate, but it’s a seeming deal.
      The cable-provider’s service was frequently affected by wind and especially by rain! We could lose any service for 3 to 5 days if it rained hard! Throw in an ice-storm (Atlanta area, SE United States where ice is unusual) and you’re stuch @Home (hey, the name of their service!) for up to a week with no internet, tv, or VOIP! There’s ya some packetloss, 100%.
      Holidays, esp. Thanksgiving in the states, we tend to see the most load balancing - I can’t quite say it’s wrong given the sudden volume of use, and that C****st would simply go down - stop providing any usable signal for internet or TV for most of the day! I suppose they balanced their over-provisioned load by serving priority customers (*higher end packages & businesses) with their available bandwidth while us suckers were left to jam the (holiday, remember) understaffed support lines: “have you tried rebooting the router?” and “try control + alt +delete” and best of all, “would you uninstall Netscape and try Explorer?”…”we can have a network engineer out to your address next week, but if they don’t find a fault with our equipment we’ll have to bill you at the service rate.”
      So, there are more than simply 3 technical factors to determine the quality of your internet service; yes, bandwidth, latency & packet loss, but also continuity of service, cost/month of the package, service responsiveness (*which ideally you’ll never need to find out!), even the wuality of the equipment provided can be an influence on the decision. Also consider downstream AND upstream bandwidt provision: fiber Gigabit is symmetrical-750+ down AND 750+ up. It’s basically ok to run a non-commercial server!

  • @magneticshrimp7429
    @magneticshrimp7429 26 дней назад

    "fun" fact. In the western world where wholesale bandwidth is cheap, on modern well managed ISP networks, averaging over X customers, delivering 50Mbps/customer costs about the same as delivering 1000Mbps/customer. Bandwidth is a very small part of operational costs for a ISP. So the upsells are almost pure extra profit. And once the base speeds are high enough people dont upgrade that often and now you (as an ISP) are probably spending more on managing different plans (backend systems/sales/support) than you make back on customers ugprading.
    So I guess just run everyone at link speed and call it a day!

  • @KavorkaDesigns
    @KavorkaDesigns Месяц назад +1

    I have a bridged router behind the isp modem, i've tried their dns, Google, CF, etc, at random times, dns errors, timeouts, any idea. I have new modem, i tried bridge and not, same affect, even with different modems, same errors, the cables are fine/tested and swapped, there seems to be no reason for the issues, I've done hard resets ip changes etc. any advice?

    • @plaintextpackets
      @plaintextpackets  Месяц назад +1

      Have you contacted the ISP? If they've swapped everything at your house sounds like it would either be outdoor cabling issues or perhaps something farther upstream at the ISP

    • @KavorkaDesigns
      @KavorkaDesigns 28 дней назад

      @@plaintextpackets I assumed the same after many weeks/months of drop outs and test. Last week I relocated the modem closer to the incoming coax source and the issue is apparently solved! :) I tried a signal booster but that didn't help, the cable its self must have internal damage. Thanks for your confirmed, it motivated me to move the modem, I wanted it inhouse, but we have to do what's needed sometimes, at-least til I have it fixed properly

  • @notaras1985
    @notaras1985 29 дней назад

    So how do we monitor the long-term effectiveness of our ISP?

  • @DJDumlao
    @DJDumlao Месяц назад +2

    will iperf do the same thing as Netprobe

    • @plaintextpackets
      @plaintextpackets  Месяц назад +1

      Iperf can pull similar stats but not sure you can run it 24x7, and it does not have a database to store the metrics natively

  • @virtualizeeverything
    @virtualizeeverything Месяц назад +1

    this is grate

  • @purrrfectnarrative5201
    @purrrfectnarrative5201 Месяц назад +1

    Good video codel or cake

    • @plaintextpackets
      @plaintextpackets  Месяц назад

      Both?

    • @purrrfectnarrative5201
      @purrrfectnarrative5201 Месяц назад

      Sure 2 routers inline.😸No I just tinker and have become obsessed lowering latency etc . Your videos are DOPE thank you

    • @TSPhotoAtlanta
      @TSPhotoAtlanta Месяц назад

      “datacentre | lan | metro | regional | internet* | oceanic | satellite | interplanetary ] [ besteffort | squash” - I have a maxim: no matter how esoteric or exotic the topic, there’s at least one corner of the internet where it’s being discussed .already. at an advanced level and there are at least two experts on the topic with at least one diametrically opposite viewpoint on at least one facet of the topic [see corollaries]
      1) there’s a ‘guide-model’ something, a very special, unimaginably \best/ piece of software, a router, canoe, sunglasses, single turbine, pico rasberry hat, silk scarf - you can’t compete with it, but you want it!
      2) the ‘guide model’ will soon be replaced by a much better one
      3) the correct answer is usually:

    • @purrrfectnarrative5201
      @purrrfectnarrative5201 Месяц назад

      @@TSPhotoAtlanta Send me some of your dope I already have a BIBLE

  • @carsonjamesiv2512
    @carsonjamesiv2512 Месяц назад +2

    GOOD INTEL!🎉😃👍

  • @necktwister666
    @necktwister666 Месяц назад +1

    in germany you get like 1gbit down with 50mbit up I need to buy more downspeed to get more upspeed but they dont sell more then the package I already have. latency is about 11ms and packetloss is basicly 0

    • @SalivatingSteve
      @SalivatingSteve Месяц назад

      That sounds typical of cable internet to have a max upload speed of around 50mbit. You’d need to switch to an ISP with fully symmetric fiber optic (like FiOS here in the USA) to get better upload speeds.

    • @necktwister666
      @necktwister666 Месяц назад +1

      @@SalivatingSteve Imagine having fiber in germany. I have a choice between 100mbit over phonelines and 1gbit over cable

  • @lazerusmfh
    @lazerusmfh Месяц назад +1

    I have 4 gigabit up/down my local isp is great

  • @goddavid8819
    @goddavid8819 Месяц назад +3

    Is there any solution to this?

    • @plaintextpackets
      @plaintextpackets  Месяц назад +2

      First step is to detect it’s happening. Then start with your existing ISP to complain about the connection. If that doesn’t go anywhere try another ISP

    • @goddavid8819
      @goddavid8819 Месяц назад +1

      @@plaintextpackets Thanks a ton sir/madam.

  • @W1LDGAMING.
    @W1LDGAMING. Месяц назад +3

    How to run GitHub repo without docker

    • @plaintextpackets
      @plaintextpackets  Месяц назад

      Unfortunately I don’t have a version that works without docker. But I would highly recommend installing it, it’s super easy

    • @W1LDGAMING.
      @W1LDGAMING. Месяц назад

      @@plaintextpacketsmy processor do not have a hyper visor can’t run docker

  • @aliencord8259
    @aliencord8259 Месяц назад

    me here casually saturating my 3gb link 💪💪💪

  • @OMNZero
    @OMNZero Месяц назад

    it nay es in Thee Band it es in The Baud
    Bandwidth
    BaudRate
    🤔

  • @lazzalicious6220
    @lazzalicious6220 Месяц назад +2

    Not me being first

  • @xVertigo101
    @xVertigo101 Месяц назад

    Fidium Fiber offers a max of 2Gbps, with my Ubiquiti equipment I get 2.5Gbps :o 500Mbps more than advertised it's crazy!
    Although Fidium is fast it all depends on the hops aka servers your connection needs to travel to, and from which can drastically reduce your internet speeds. thus, it's important to download anything from your closest server to you.
    My best speeds are from Boston I get around 1.5Gbps on game downloads sometimes it peaks to 2.2Gbps
    The west coast say California speeds sit around 600-800Mbps and ping is drastically increased due to the amount of hops and whether or not a hop in the region is slow. Sometimes the most optimal route is not the fastest either.

  • @drjankenstein
    @drjankenstein Месяц назад

    ok, but if you're going to use AI generated art, spend 5 minutes in photoshop to fix up the obvious gibberish words. "bandband, latency and loss" circa 0:45 seconds is not a great first look - if people dont get its AI they think you didnt proofread.....