Update on good Ethernet switch

@mc1969 , the best ethernet cables I've had are made by Audioquest. 

LHY now has the SW-6. Despite the lower product numbering this sits between the SW-8 and SW-10 and I believe it will even be superior to the later because its not repurposing cheap commercial switch hardware - it's a custom board design.

To my knowledge the only other switch that is custom built for audiophile considerations is by Nordost and its very expensive. Even those crazy expensive Ansuz switches use commercial boards inside.

Because the SW-6 is a custom made board the clock is now located right where it needs to be rather than on a daughter board. Also one output has dedicated isolation from the other 4. It has a SPF connection as well if you want to do optical.

Here's a new, lower priced option, from LHY:

https://www.beatechnik.com/fmc

Uptone is working on a new Etherregen; meanwhile the first one is available again, isloation mode and 10m clock compatible. 

For my home network, I use a managed tp link switch which is more expensive than unmanaged, as well as a mikrotik 24 port manged switch router.

A couple of things about ethernet. The ethernet standard is that cat 5e cable will support gigabit up to 100 meters. A switch will tend to clean up incoming frames by re-generating them.

There are two main TCP/IP protocols: TCP and UDP. TCP sits on top of the lower levels of the stack and will monitor sequence numbers of incoming packets/frames and will request re-transmission if a missing packet is detected. In this manner it is a lossless protocol. The UDP protocol is a broadcast protocol as it simply passes frames along with best efforts, and has no mechanism for detecting and correcting errors.

As mentioned above, at the switching layer 2--frames that fail the crc check are discarded with no notification to the layers above it. UDP has lower overhead than TCP and is often used for media output that is generally thought to be tolerant of missing data... O:

One can’t know the protocol used unless the vendor of the source data machine puts it in the specs, or you can run a sniffer like wireshark and see what is going on.

All this is leading to the idea that if music is sent over the network using a broadcast protocol like udp, there could be some variability in data quality depending on the hardware and quality of network connections and so forth.

These concepts apply to wireless as well except that layer 2 for that is using radio waves instead of a wire. I don’t have enough knowledge about wireless to opine on the reliability and frequency of lost packets--but if UDP is the higher level protocol, still not going to ensure complete transmission of data.

* I don’t want to get into the deets of the tcp/stack, so I am using frames, packets,and segments interchangeably to get the concepts across, though if one digs into the network stack, these names have specific meanings and must be kept accurate.