Collaboratory - Network Problems/Solutions

Under certain network conditions, attempts to join a Collaboratory session yield the message "Hub cannot establish back-connection to Node..." Two potential reasons, not mutually exclusive, are discussed:

1. The Collaboratory uses an implementation of the Common Object Request Broker Architecture (CORBA) specification to communicate between instances of Chimera. By default, CORBA randomly allocates network ports for communication. Clients behind firewalls may see this network traffic as a security threat and block the connections.
   
   
2. Some routers and firewalls use a process called Network Address Translation (NAT) that enables several computers on an intranet to be connected to the internet via a single IP address. A computer on a network using NAT is assigned a "local" IP address, often 192.168.1.102, that is only meaningful within the intranet (cannot be used for messages from elsewhere on the internet). A computer on a network using NAT may supply its local address to the Collaboratory hub rather than the address needed for internet communications.

These problems can be solved by designating a port number and IP address using the Collaboratory preferences. Regardless of which of the back-connection problems apply, a specific port number should be designated for communications. The port number must be greater than 1024, as ports up to 1024 are reserved for use by the system. If a firewall is blocking connections, it should be configured to allow incoming TCP traffic on the designated port. If NAT is known to be used, or the back-connection error message still appears after a port number has been designated, the proper IP address should also be supplied. The proper address is that of the router or computer supplying the internet connection.