sources-request@genrad.UUCP (04/12/85)
From: blyon@sun (Bob Lyon)
[I have taken selected parts of the rpc.spec document (after nroff'ing) and
included it here to describe this what this package is, and what it does.
This will be a ten part "shar" distribution - the moderator ]
Remote Procedure Call
Protocol Specification
1. Introduction
This document specifies a message protocol used in implementing
Sun's Remote Procedure Call (RPC) package. The protocol is
specified with the eXternal Data Representation (XDR) language.
This document assumes that the reader is familiar with both RPC
and XDR. It does not attempt to justify RPC or its uses. Also,
the casual user of RPC does not need to be familiar with the
information in this document.
1.1. Terminology
The document discusses servers, services, programs, procedures,
clients and versions. A server is a machine where some number of
network services are implemented. A service is a collection of
one or more remote programs. A remote program implements one or
more remote procedures; the procedures, their parameters and
results are documented in the specific program's protocol specif-
ication (see Appendix 3 for an example). Network clients are
pieces of software that initiate remote procedure calls to ser-
vices. A server may support more than one version of a remote
program in order to be forward compatible with changing proto-
cols.
For example, a network file service may be composed of two pro-
grams. One program may deal with high level applications such as
file system access control and locking. The other may deal with
low-level file I/O, and have procedures like ``read'' and
``write''. A client machine of the network file service would
call the procedures associated with the two programs of the ser-
vice on behalf of some user on the client machine.
1.2. The RPC Model
The remote procedure call model is similar to the local procedure
call model. In the local case, the caller places arguments to a
procedure in some well-specified location (such as a result
register). It then transfers control to the procedure, and
eventually gains back control. At that point, the results of the
procedure are extracted from the well-specified location, and the
caller continues execution.
The remote procedure call is similar, except that one thread of
control winds through two processes - one is the caller's pro-
cess, the other is a server's process. That is, the caller pro-
cess sends a call message to the server process and waits
(blocks) for a reply message. The call message contains the
procedure's parameters, among other things. The reply message
contains the procedure's results, among other things. Once the
reply message is received, the results of the procedure are
extracted, and caller's execution is resumed.
On the server side, a process is dormant awaiting the arrival of
a call message. When one arrives the server process extracts the
procedure's parameters, computes the results, sends a reply mes-
sage, and then awaits the next call message. Note that in this
model, only one of the two processes is active at any given time.
That is, the RPC protocol does not explicitly support multi-
threading of caller or server processes.
1.3. Transports and Semantics
The RPC protocol is independent of transport protocols. That is,
RPC does not care how a message is passed from one process to
another. The protocol only deals with the specification and
interpretation of messages.
Because of transport independence, the RPC protocol does not
attach specific semantics to the remote procedures or their exe-
cution. Some semantics can be inferred from (but should be
explicitly specified by) the underlying transport protocol. For
example, RPC message passing using UDP/IP is unreliable. Thus,
if the caller retransmits call messages after short time-outs,
the only thing he can infer from no reply message is that the
remote procedure was executed zero or more times (and from a
reply message, one or more times). On the other hand, RPC mes-
sage passing using TCP/IP is reliable. No reply message means
that the remote procedure was executed at most once, whereas a
reply message means that the remote procedure was exactly once.
(Note: At Sun, RPC is currently implemented on top of TCP/IP and
UDP/IP transports.)
1.4. Binding and Rendezvous Independence
The act of binding a client to a service is NOT part of the
remote procedure call specification. This important and neces-
sary function is left up to some higher level software. (The
software may use RPC itself; see Appendix 3.)
Implementors should think of the RPC protocol as the jump-
subroutine instruction (``JSR'') of a network; the loader
(binder) makes JSR useful, and the loader itself uses JSR to
accomplish its task. Likewise, the network makes RPC useful,
using RPC to accomplish this task.
.
.
.
3. Other Uses and Abuses of the RPC Protocol
The intended use of this protocol is for calling remote pro-
cedures. That is, each call message is matched with a response
message. However, the protocol itself is a message passing pro-
tocol with which other (non-RPC) protocols can be implemented.
Sun currently uses (abuses) the RPC message protocol for the fol-
lowing two (non-RPC) protocols: batching (or pipelining) and
broadcast RPC. These two protocols are discussed (but not
defined) below.
3.1. Batching
Batching allows a client to send an arbitrarily large sequence of
call messages to a server; batching uses reliable bytes stream
protocols (like TCP/IP) for their transport. In the case of
batching, the client never waits for a reply from the server and
the server does not send replies to batch requests. A sequence
of batch calls is usually terminated by a legitimate RPC in order
to flush the pipeline (with positive acknowledgement).
3.2. Broadcast RPC
In broadcast RPC based protocols, the client sends an a broadcast
packet to the network and waits for numerous replies. Broadcast
RPC uses unreliable, packet based protocols (like UDP/IP) as
their transports. Servers that support broadcast protocols only
respond when the request is successfully processed, and are
silent in the face of errors.
.
.
.
6. Appendix 2: Record Marking Standard (RM)
When RPC messages are passed on top of a byte stream protocol
(like TCP/IP), it is necessary, or at least desirable, to delimit
one message from another in order to detect and possibly recover
from user protocol errors. Sun uses this RM/TCP/IP transport for
passing RPC messages on TCP streams. One RPC message fits into
one RM record.
A record is composed of one or more record fragments. A record
fragment is a four-byte header followed by 0 to 2^31-1 bytes of
fragment data. The bytes encode an unsigned binary number; as
with XDR integers, the byte order is from highest to lowest. The
number encodes two values - a boolean which indicates whether the
fragment is the last fragment of the record (bit value 1 implies
the fragment is the last fragment) and a 31-bit unsigned binary
value which is the length in bytes of the fragment's data. The
boolean value is the highest-order bit of the header; the length
is the 31 low-order bits. (Note that this record specification
is not in XDR standard form!)
7. Appendix 3: Port Mapper Program Protocol
The port mapper program maps RPC program and version numbers to
UDP/IP or TCP/IP port numbers. This program makes dynamic bind-
ing of remote programs possible.
This is desirable because the range of reserved port numbers is
very small and the number of potential remote programs is very
large. By running only the port mapper on a reserved port, the
port numbers of other remote programs can be ascertained by
querying the port mapper.