Hash routing and related approaches for clusters of web servers
Hash routing seems well suited for distributing URLs among
a neighboring set of web caches in the same administrative domain.
How to make distributed proxy servers by URL hashing.
Cache Array Routing Protocol (CARP).
The Cache Array Routing Protocol (CARP) uses a hash function to divide
the URL space among loosely coupled proxies.
Keith W. Ross,
Hash Routing for Collections of Shared Web Caches,
IEEE Network, November/December 1997.
This paper discusses hash-routing schemes for mapping requests
to a collection of shared web caches residing on a LAN.
- Valloppillil & Cohen ,
Hierarchical HTTP Routing Protocol
21 April 1997.
This paper proposes that the administrator of an array of proxies
use hashing to distribute the URL space
among several web caches.
Peter Danzig, NetCache Architecture and Deployment,
3rd International WWW Caching Workshop,
NetCache uses a hashing scheme to uniformly partition
cache-to-cache traffic to that
peer caches are responsible for non-overlapping URLs.
Other approaches for clusters of web servers:
David Thaler and Chinya Ravishankar,
Using Name-Based Mappings to Increase Hit Rates,
IEEE/ACM Transactions on Networking, Vol. 6, No. 1,
This paper proposes a name-based mapping algorithm for use in
a cluster of web caches
with equivalent functionality and capacity. This mapping
would be based on both the URL and on server addresses,
for robustness when individual servers in a cluster go up or down.
[Adaptive Web Caching at LBNL]
Last modified: September 1998