ethernet VS kcp-go

After creating the ARP Request packet we need to now create an Ethernet Frame. The Ethernet frame contains fields such as Source and Destination Hardware (MAC) among others. Now, as the communication inside a network is carried out using the MAC Address, we can set the value of destination hardware address field to theMAC Address to which we want to communicate. Learn more about Ethernet Frame here.

You can send raw packets with something like this [1] [2].

I imagine you want to test the approach is software before implementing it in an FPGA.

You can then construct Ethernet frames to be as simple as you like [3].

This question seems like what you might be trying to do [4].

[1] https://stackoverflow.com/questions/12229155/how-do-i-send-a...

[2] https://stackoverflow.com/questions/57133295/how-can-i-liste...

[3] https://en.wikipedia.org/wiki/Ethernet_frame

[4] https://stackoverflow.com/questions/815758/simple-serial-poi...

Ethernet frame structure (https://en.wikipedia.org/wiki/Ethernet_frame)

Some communication does happen at the MAC level, like ARP, but these days it's all done with the expectation that your machine is going to be making use of TCP/IP. So this layer 2(-ish) protocol really just exists to establish the addressing. So the point of a private IP is mainly because all of the typical communication protocols we use expect one. Could you make something work without one? MAC-to-MAC? Yes, that's ethernet traffic, and the Wikipedia page shows the source and destination MACs as part of the structure of the header of a frame.

I find looking at the header structure directly helps me to wrap my head around it. At layer 2 you are dealing exclusively with Ethernet frames which have a number of fields in their header that indicate their source and destination MAC addresses, among other things. These frames are what the switch builds it's CAM table from and it uses that table to forward traffic, and for a layer 2 switch that is the entire picture.

The preamble and SFD (and IPG) are not part of the frame, they are part of the Layer 1 Packet. This is one of the most common Ethernet misconceptions. See Wikpedia's Ethernet Frame article.

Wrong layer. :) This is later two, or the Data Link layer. https://en.wikipedia.org/wiki/Ethernet_frame

I guess I should look into something like this before I get much further. It sounds exactly like what you were recommending. I do enjoy reinventing the wheel sometimes, but it is usually not successful lol.

I used kcp-go. The server includes two parts, Mothership and child server. After the mothership server matchmaking system finds a game it will pass those users to the child server.