I’ve been reading the SL4 RC docs and noticed that aspects of security have changed since the beta and since I made these screencasts on networking.I think these are positive changes and, from what I’ve seen so far both TCP and UDP sockets drop their security limitations for an elevated application. That is – a non-trusted application (whether in the browser or out of browser) has restrictions imposed on it; 1. TCP sockets can only be opened to ports 4502 to 4534.
2. TCP sockets can only be opened once a security policy allowing the opening has been downloaded via either;
1. TCP over port 943 on the target server
2. HTTP from port 80 on the target server ( this is new in the RC )
3. UDP multicast sockets can only be opened to ports above 1024.
4. UDP multicast groups can only be joined once a security policy allowing the joining has been downloaded via either;
1. UDP unicast to port 9430 on the target server ( for a single source multicast group )
2. UDP multicast to port 9430 on the multicast group ( for an any source multicast group )and all those restrictions go away if you’re running trusted. Read more: Mike Taulty's Blog
2. TCP sockets can only be opened once a security policy allowing the opening has been downloaded via either;
1. TCP over port 943 on the target server
2. HTTP from port 80 on the target server ( this is new in the RC )
3. UDP multicast sockets can only be opened to ports above 1024.
4. UDP multicast groups can only be joined once a security policy allowing the joining has been downloaded via either;
1. UDP unicast to port 9430 on the target server ( for a single source multicast group )
2. UDP multicast to port 9430 on the multicast group ( for an any source multicast group )and all those restrictions go away if you’re running trusted. Read more: Mike Taulty's Blog