Difference between live, liverepeater-origin and liverepeater-edge
So as it is described in numerous guides/threads, if we want to have a kind of CDN to be built using Wowza, we configure servers as origin and edges, and use packetizers for original input, and repeaters on edges.
We had a test setup, where two Wowza servers were both just "live" and both had packetizers, and no repeaters. In this setup Wowzas were able to pick up streams from each other over RTMP and successfully rebroadcast it to users.
So question is - what is the difference/negative outcome of using just "live" and packetizers on all servers, while still repeating streams? Does it have any impact on broadcast, or maybe CPU/RAM usage? As if it works in this simplified setup (and it does, just we don't know consequences) - we would rather avoid complications and stick to easy "one same app per server" setup?
Wowza will try to drop frames for low bandwith clients, this Property can be used to disable that behavior. I only see it set to true for origin stream types, which makes sense as you should have plenty of bandwidth between the origin and the edge, where the edge is the playback client to the origin.
Yes, as you can see this is where StreamTypes are defined and can be altered. You can define a new StreamType if you want. Experiment. It is difficult answer every hypothetical question, and there is a limit to how much we can or will go deeper than the
published documentation. The Property names try to be self-explanatory.
The source code of Wowza is not available, if that is what you are asking.
I would look at the documentation, examples and tutorials first. Then search around for discussions as necessary. Charlie's answers are always authoritative, though features and fixes may have made some comments obsolete.