Coda File System

Re: Coda with weak networks

From: LEE, Yui-wah <clement_at_cse.cuhk.edu.hk>
Date: Mon, 10 Aug 1998 20:27:58 +0800 (HKT)
Peter J. Braam wrote:
> 
> 
> Venus _could_ do this with some modifications.  In fact there are plans to
> implement several such "side effects" as well as a more elaborate
> mechanism using a "proxy client" to actually repeat work (ask Yui-Wah Lee
> (Clement) <clement+@cs.cmu.edu>) about this.
> 

Yes, I am extending Coda in a rather interesting direction (as
part of my PhD-thesis work).  

One problem of of using Coda (and any mobile file systems in general)
with weak network (phoneline, wireless modem etc) is the bottleneck of
shipping back updates made on clients back to servers.  Modern files
become larger and larger and it is a pain to ship them across weak
networks.  To solve this problem, we may ship only the difference (or
delta) so as to reduce the network traffic, this is not yet done in
Coda and should be an interesting project.

At the same time, however, I am exploring another possible technique.
Observing that many modern huge files are generated by user operations
which are rather compact in size, I am proposing to ship those user
operations rather than the files that they produce, by so doing, the
network traffics is subtantially reduced.  On the opposite end, there
is a mechanism to re-execute those operations to re-produce the
results.  Of course there are many issues we need to take care:
correctness, binary compatibility etc.  Work is in progress but the
initial result is very encouraging (I gave a talk to Coda group about
it this Summer).

Yui-wah LEE (Clement)

======================================================================
Department of Computer Science and Engineering,    Fax: (852)-26035024 
The Chinese University of Hong Kong     Email: clement_at_cse.cuhk.edu.hk
======================================================================
Received on 1998-08-10 08:29:20