REVIEW: On the scale and performance of cooperative Web proxy caching

REVIEW: On the scale and performance of cooperative Web proxy caching

The paper provides a study of effectivenesss of large scale cooperative
caching. It discusses what client population sizes of caches benefit from
cooperation. By analysing traces of web requests as well as modelling
requests, it was possible to show that although cache hit ratios increase
in general with the number of clients, the most benefit of cache
cooperation occurs with a relatively small number of clients. In seven-day
traces significant benefits were achieved for client populations of a
litle more than 2,500, and in real deployments only marginal benefits can
be achieved for organizations with more than 20K clients.

Based on these findings, the paper argues that there isn't much point in
expending effort in developing large scale cooperative caching schemes, as
any scheme will do well with such small client populations. It also points
out that intra-organization traffict has as little homogeneity as
inter-organization as well as clustering of clients by access does not
yield much benefit either. The performance of caching is highly limited by
document cacheability and thus this is the most promising area for caching
improvements.
The paper also infers that most popular web object are on average smaller
than unpopular ones, which in turn lead to higher cache-hit-rates for
objects than per byte.

It would be useful to see a further discussion on what constitues
inter-cache latency and bandwidth that justify cooperation, in comparison
to latencies to web servers, as well as how latencies affect the ideal
client population size.
Received on Thu Oct 20 2005 - 10:50:11 EDT