Yeah I see your point. I guess I’ll be a little bit smarter about it. If the users’ bandwidth is medium or low, default to 1935,443,80, etc. But if they’re on a high speed connection, use rtmpt because it will always go through. Really what I’m trying to accomplish is getting a connection as soon as possible to the stream. Which I think happens altering the order this way. And I guess overall it’s not something that’s “critical”, but I’d rather know that port 80 is always open, and should connect the first attempt.

Do you have any insights into how much less efficient it is?

Thanks man.

hey no worries. I’ll reply to your post on the mailing list later today.

-A

In fact, the default order is pretty logic in my sense. It first attempts an RTMP connection on port 1935. This makes sense since port 1935 is the default port assigned to FMS and since RTMP is more efficient than RTMPT.

Second, an attempt is made on port 443 in case port 1935 is blocked. This will sometimes happen in companies (though this is less and less frequent all the time). So an RTMP conenction that can’t go through port 1935 might go through port 443 if RTMP traffic is allowed on that port. And that port is open for sure.

Then, if all else fails, we fallback on port 80 (which we know is open) using RTMPT streaming. This is less efficient than RTMP but, when used that way, will almost always go through unless there is a packet sniffer that detects that this is in fact RTMP packets wrapped inside regular HTTP packets. This will almost never happen. To do this, a network administrator means he REALLY WANTS to block all RTMP traffic on his network.

Just my 2 cents…