The rpcbind utility replaces the portmap utility available in previous releases of the Oracle Solaris environment. This appendix is included to help you understand the history of port and network address resolution using the portmap utility.
Oracle Solaris RPC-based services use portmap as a system registration service. It manages a table of correspondences between ports (logical communications channels) and the services registered at them. It provides a standard way for a client to look up the TCP/IP or UDP/IP port number of an RPC program supported by the server.
For client programs to find distributed services on a network, they need a way to look up the network addresses of server programs. Network transport (protocol) services do not provide this function. Their task is to provide process-to-process message transfer across a network, that is, a message is sent to a transport-specific network address. A network address is a logical communications channel. By listening on a specific network address, a process receives messages from the network.
The way a process waits on a network address varies from one operating system to the next, but all provide mechanisms by which a process can synchronize its activity with arriving messages. Messages are not sent across networks to receiving processes, but rather to the network address at which receiving processes pick them up.
Network addresses are valuable because they allow message receivers to be specified in a way that is independent of the conventions of the receiving operating system. TI-RPC, being transport independent, makes no assumptions about the structure of a network address. It uses a universal address. This universal address is specified as a null-terminated string of characters. Such a universal address is translated into a local transport address by a routine specific to the transport provider.
The rpcbind protocol defines a network service that provides a standard way for clients to look up the network address of any remote program supported by a server. Because this protocol can be implemented on any transport, it provides a single solution to a general problem that works for all clients, all servers, and all networks.