The general mechanism for client-server applications is provided by the Remote Procedure Call (RPC) package. RPC was developed by Sun Microsystems and is a collection of tools and library functions. Important applications built on top of RPC are NIS, the Network Information System (described in Chapter 13), and NFS, the Network File System (described in Chapter 14), which are both described in this book.

An RPC server consists of a collection of procedures that a client can call by sending an RPC request to the server along with the procedure parameters. The server will invoke the indicated procedure on behalf of the client, handing back the return value, if there is any. In order to be machine-independent, all data exchanged between client and server is converted to the External Data Representation format (XDR) by the sender, and converted back to the machine-local representation by the receiver. RPC relies on standard UDP and TCP sockets to transport the XDR formatted data to the remote host. Sun has graciously placed RPC in the public domain; it is described in a series of RFCs.

Sometimes improvements to an RPC application introduce incompatible changes in the procedure call interface. Of course, simply changing the server would crash all applications that still expect the original behavior. Therefore, RPC programs have version numbers assigned to them, usually starting with 1, and with each new version of the RPC interface, this counter will be bumped up. Often, a server may offer several versions simultaneously; clients then indicate by the version number in their requests which implementation of the service they want to use.

The communication between RPC servers and clients is somewhat peculiar. An RPC server offers one or more collections of procedures; each set is called a program and is uniquely identified by a program number. A list that maps service names to program numbers is usually kept in /etc/rpc, an excerpt of which is shown in Example 12.4.

#
# /etc/rpc - miscellaneous RPC-based services
#
portmapper      100000  portmap sunrpc
rstatd          100001  rstat rstat_svc rup perfmeter
rusersd         100002  rusers
nfs             100003  nfsprog
ypserv          100004  ypprog
mountd          100005  mount showmount
ypbind          100007
walld           100008  rwall shutdown
yppasswdd       100009  yppasswd
bootparam       100026
ypupdated       100028  ypupdate

In TCP/IP networks, the authors of RPC faced the problem of mapping program numbers to generic network services. They designed each server to provide both a TCP and a UDP port for each program and each version. Generally, RPC applications use UDP when sending data, and fall back to TCP only when the data to be transferred doesn’t fit into a single UDP datagram.

Of course, client programs need to find out to which port a program number maps. Using a configuration file for this would be too unflexible; since RPC applications don’t use reserved ports, there’s no guarantee that a port originally meant to be used by our database application hasn’t been taken by some other process. Therefore, RPC applications pick any port they can get and register it with a special program called the portmapper daemon. The portmapper acts as a service broker for all RPC servers running on its machine. A client that wishes to contact a service with a given program number first queries the portmapper on the server’s host, which returns the TCP and UDP port numbers the service can be reached at.

This method introduces a single point of failure, much like the inetd daemon does for the standard Berkeley services. However, this case is even a little worse because when the portmapper dies, all RPC port information is lost; this usually means you have to restart all RPC servers manually or reboot the entire machine.

On Linux, the portmapper is called /sbin/portmap, or sometimes /usr/sbin/rpc.portmap. Other than making sure it is started from your network boot scripts, the portmapper doesn’t require any configuration.