3 ms·
Latency. Sadly, Ethernet over powerline (IEEE 1901) made the same unfortunate choice as WiFi and the (analog phone line) modem V.4x standards in including erro
by guenthert 7y ago
Latency. Sadly, Ethernet over powerline (IEEE 1901) made the same unfortunate choice as WiFi and the (analog phone line) modem V.4x standards in including error correction. This is not what you want or need for TCP/IP. It's better to drop a packet than to deliver it late. IEEE 1901 tries to mitigate this by using forward error correction, but succeeds in that only to a degree.
Bandwidth is alright-ish at about estimated 2MiB/s here (single pair, about 50ft apart), but RTT of ~4ms with occasional spikes in excess of 1s (at a different location, I've seen spikes in excess of 5s). This is only barely noticeable for accessing the WWW, but unpleasant for interactive GUI work (e.g. via NX or VNC) and a deal-breaker for cluster communication protocols.
- aidenn0 7y ago1. There are still some devices designed for lan-only use that assume packet loss will be extremely rare. 2. TCP throughput suffers when non-congestion related packet-loss exceeds a fairly low value
- gitgudnubs 7y agoOnly an idiot would design a device to assume no packet loss, unless they were _also_ allowed to design the network. If you buy such a device and put it on a janky power network, then you have only yourself to blame. Standards decisions should never be made to accommodate designers that want to assume there's effectively no packet loss. Some degree of error correction is totally reasonable, but your first point is irrelevant to discussion.
- Dylan16807 7y agoTo a TCP implementation, what's the difference between a dropped packet and one that's delayed by hundreds of round trip times? What problems are caused by not dropping that packet? (If all the packets are failing to go through for that long, then you have much bigger problems than the error correction.)