I have a client that wants to publish a 1 Mbps live stream from a location in the Pacific Northwest, to a max of 10,000 simultaneous connections. I was thinking of trying out one or two loadbalanced 4xlrg instance(s) in AWS Eastern region to handle the subscriber load.
I'm wondering if it would be optimal to have an WMS EC2 small instance running in the new Oregon region as the ingest point for the publishing live stream and then have that instance push the stream to 4xlrg instance in the Eastern region. My thinking is that the AWS backbone is better suitable to getting the 1Mbps stream over to the Eastern region faster than the publisher going through more hops to go to the Eastern server directly. Is this just nonsense on my part?