On Fri, Jan 6, 2023 at 11:29 AM Stewart Bryant <stewart.bryant@xxxxxxxxx> wrote:
I suppose that you have to question whether IP is the ideal base for multicast?
Our networks are no longer mono-protocol and multicast tends to be domain specific.
Many of the original uses for multicast are now dominated by unicast packet duplication with edge computing making this less of a bandwidth hog, so it is not clear what the long term future of multicast is.
- Stewart
What protocols are in widespread use besides IPv4 and IPv6?
Are Akamai and the like really rolling their own network protocols to squeeze the nth degree of performance out of their networks?
I suspect the biggest barrier for deployment of multicast is the steep learning curve for implementers. I have no idea how I would go about getting a multicast address group assigned, no idea how I would set up a test stand, etc. etc. And the information is certainly not easily accessible. Then as mentioned earlier, multicast is only giving me a limited UDP functionality that is essentially unidirectional. I have to do my own work at transport and above and on top of that, the security model is changed.
And after all that, the return on multicast is only going to be there if I have a broadcast application with at least a hundred people watching a stream on a regular basis.
Anything that does multicast is going to end up involving some sort of voodoo at the point where one packet goes in and multiple packets go out. That voodoo is not going to be a function of the router core, it is going to be something higher level poking that core.
So given modern hardware, does it really make such a big difference if that voodoo is happening in a processor that is in the router chassis itself or in another box in the same rack connected by a nice fat pipe?
