• Midnight Wolf@lemmy.world
    link
    fedilink
    English
    arrow-up
    14
    arrow-down
    2
    ·
    10 hours ago

    I see this so often and nobody ever seems to realize that local/home VPNs use upload bandwidth, which for some is in dire low supply. I can’t have 4 full-time users using my upload connection routing through wireguard, when all 4 stream videos throughout the day. And that’s just 3rd party services like YouTube and Twitch, not plex. Then you add in two additional, off-site users who want to watch something with me on plex, and we are all given ~1.5 megabits a piece of a 10meg upload pipe over here. Mmmm, crispy pixels. ‘you can just use some IPs in wg so you don’t need to tunnel all data, just what you need’, they say, and I rebuke by showing them my dynamic IP address. ‘ask for a static one’ and they haven’t offered that for years besides enterprise customers.

    And that’s before I ask everyone ‘so everyone download wireguard and scan your individual qr code, or I will send you the config file’ and everyone but a single user just hears the ocean. Then I need to teach them about VPNs, why we use it, why plex doesn’t work when the little lock isn’t showing on their phones, why ‘I had the lock in the corner but I couldn’t make a call or get online, so we are all getting [thing you don’t like] for dinner since I couldn’t ask’. Then I have to troubleshoot and tell them to toggle it off and on again…

    The we get to the bit where they try to cast to the TV, and the chromecast is like ‘lol wtf is a VPN’ and we are back at square one, everyone hates me, I hate everyone right back, all changes from this experiment get reverted, and I lose credibility.

    VPNs are useful, but I rage at people who assume they are a blanket solution for all situations and use-cases. And often, the people suggesting them are smug, like they have found something that nobody knows about and are superior because their situation doesn’t color outside of the lines.

    Damn that was nice to vent. Been bothering me for way too damn long.

    • the_crotch@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      4
      ·
      7 hours ago

      I had the lock in the corner but I couldn’t make a call or get online, so we are all getting [thing you don’t like] for dinner since I couldn’t ask’. Then I have to troubleshoot and tell them to toggle it off and on again…

      “I’m sorry I made my collection of movies available for you to watch for free, I’ll make sure to never do anything like that again”

    • Saik0@lemmy.saik0.com
      link
      fedilink
      English
      arrow-up
      9
      arrow-down
      2
      ·
      10 hours ago

      and everyone but a single user just hears the ocean.

      I’m sorry, but this made me bust laughing. This is dead accurate for a few people in my life.

      Then I have to troubleshoot and tell them to toggle it off and on again…

      And this is exactly the type of support a lot of people just don’t want to do (including me). And the options really boil down to settle for supporting all this, or the risk of public access to unauthenticated endpoints.

      They could just fix the endpoints and it’ll be a non-issue. But they won’t because “backwards compatibility”.

      There are even other options that I can pre-emptively offer… but they all SUCK.

      You can whitelist ip access… ISP ips rotate and are dynamic.
      You can setup crowdsec and/or fail2ban… until a user fails to login a few times in a row because users are users and get themselves banned, now you’re back to support role.
      VPNs already covered ad nauseam.

      There are options… they all suck, especially when the answer of JUST FIX THE ENDPOINT is sitting right there.

    • catloaf@lemm.ee
      link
      fedilink
      English
      arrow-up
      3
      arrow-down
      10
      ·
      9 hours ago

      Upload is upload. It doesn’t matter if it’s over the plain Internet or over a tunnel, you’re still uploading roughly the same number of bytes per second.

      • Saik0@lemmy.saik0.com
        link
        fedilink
        English
        arrow-up
        2
        arrow-down
        2
        ·
        edit-2
        8 hours ago

        Tunnels have overhead. MTU overhead itself can cut 5% of your total bandwidth as a default (1500 -> 1420). Forget all the side-channel control stuff.

        MTU itself is an interesting issue for wireguard. It defaults to 1420, which should be fine in most cases as the default is 1500 for most ISP connections. But there are interesting cases where you need to go less… If you try to cram a 1420 MTU packet down a 1440 MTU ISP connection (you need 28Bytes overhead minimum, so would need 1412 in Wireguard in this case)… you’re rewriting a fuckton of packets and splitting tons of data that can ruin your connection speed (halving immediately).

        I have seen some people recommend 1384 MTU before… The lower you tune this for compatibility the less speed you get.

        Once again though… this is way over a normal users head. And likely even over yours since you don’t seem to recognize that this is happening and that it isn’t byte per byte the same.

        You should expect wireguard to lose you 5% speed minimum… with other issues potentially making it worse.

        Edit: clarification on a sentence cause the wording was bad.