From: Tarik Nesh-Nash (tarikn_at_microsoft.com)
Date: Wed Jan 28 2004 - 18:04:08 PST
This paper is a good technical introduction to RPC. It seems it was the
first production implementation to this mechanism and many of its ideas
still apply today.
RPC extends the idea of local procedure calls to network communication.
Semantics, efficiency and generality are taken into consideration as key
to success.
The paper goes into details on the used structure at the level of the
client (user, user-stub and RPCRuntime) and server (server, server stub
and RPCRuntime). The mechanism is simple and well described.
Next, the paper discusses the binding and naming conventions that allow
communication. A new protocol is designed to fully support RPC,
existing network protocol give unsatisfactory performance results.
The performance results was acceptable but still need improvement,
today's implementations seem to give a better results. An interesting
discussion topic would be the performance improvements progress.
I would like to get a link to the Nelson's thesis. It will help me
understand the many decisions taken on this document based on the paper.
This archive was generated by hypermail 2.1.6 : Wed Jan 28 2004 - 18:04:34 PST