Paper Evaluation

Bryan Wang (
Tue, 16 May 2000 00:58:30 PDT

Get Your Private, Free E-mail from MSN Hotmail at
Bryan Wang, 5/15/00

Implementing Remote Procedure Calls

Birrell and Nelson's paper address the need to make RPC appear as
closely as possible to a local procedure call, in both syntax and
performance. Traditionally RPC has not experienced widespread use
because its semantics and other things are too complicated.

Their approach was to introduce the "user stub", "server stub", and
"RPCRuntime" as levels of abstraction to the RPC process. These
procedures emulate local procedure calls as closely as possible. If
you move the user, server, and the three RPC concepts above on to a
single machine, the programs will still work.

This paper is my first exposure to remote procedure calls so I learned
what they are and how they worked. I didn't learn what they are for
or why we use them, but I guess they work sort of like the
procedure-based dual of a distributed message-based system.

I thought the paper was good and well-written. There was a lot of
vocab and concepts, but the diagrams helped illustrate their purposes
and interactions. The paper also left several ideas for further
research, including a set of unused optimizations that would yield an
additional 2x performance increase.