1044
The Free Market Lie: Why Switzerland Has 25 Gbit Internet and America Doesn't (sschueller.github.io)
submitted 6 months ago by to c/technology@lemmy.world
175 comments fedilink hide all child comments
you are viewing a single comment's thread
view the rest of the comments
[–] 3 points 6 months ago (4 children)

It's only around $100 for a pcie 25gb Ethernet card. Not particularly a huge hurdle.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 4 child comments
    •
  • [–] 6 points 6 months ago (3 children)

    Networking researcher here. Your bottleneck wouldn't be the NIC, but memory bandwidth, CPU compute (for TCP), PCIe bandwidth, and Storage bandwidth, also the bandwidth of the server you're connecting to. You'll also need some sort of fiber SFP connection for your entire house, and those have firmware that usually makes them vendor locked. Most networking issues are also latency related, so increasing your throughput to 25Gbps wouldn't help.

    So yeah, not a good time for home use.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 3 child comments
    •
  • [–] 2 points 6 months ago (2 children)

    I'm pretty sure that nic cards for those speeds would really need more hardware offloading and dma to stand a chance of those speeds. With those it should be possible. With the right hardware handling there shouldn't be a problem, ssds connected to pci manage a lot more.

    In real terms, right now who needs it aside from to post speed test results?

    I have gigabit symmetric and can upgrade to 2.5. But, I cannot imagine we'd need 2.5 let alone 10 or 25. And I'm a fairly heavy user.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 3 points 6 months ago* (1 child)

    1. All NICs already work off of DMA to access/copy packets into/from memory. Yes, even your $10 ones. So "would need DMA to stand a chance" doesn't have any technical meaning other than putting a bunch of words together.

    2. The bottleneck for TCP is sequence number processing, which must be done on a single core (for each flow) and cannot be parallelized. You also cannot offload sequence number processing without making major sacrifices that result in corrupted data in several edge cases (see TCP chimney offload, which cannot handle the required TCP extensions needed to run TCP at 1Gbps). So no, "more offloading" is easy to say but not feasible.

    3. Who needs it: data centers trying to scale legacy software, or dealing with multi region data replication (rocev2 is terrible for long distance links). But no, no home user would need it

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 1 point 6 months ago

    Well I was thinking more home users since this was what the post was about. Pretty sure data centres have solutions for this already with price tags that would make us cry.

    At home I can see only edge cases where even going to 2.5 would be useful for us here anyway. Let alone more.

    Now I'm sure as time passes demands will continue to increase and we'll need more speed. But for now running 2.5/10 internally and 1gbit to the Internet is more than enough.

  • permalink
  • fedilink
  • source
  • parent