- From: Mike Schwartz <schwartz@latour.cs.colorado.edu>
- Date: Tue, 4 Apr 1995 18:30:13 -0600
- To: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
Hi. I'm posting this message to the HTTP Working Group mailing list because I think you may be interested in our generic approach to improving HTTPD performance. - Mike Schwartz Principal Investigator, Harvest project ------------------------------------------------------------------------------- The Harvest "httpd accelerator" is a specially configured Harvest object cache that intercepts incoming HTTP requests, quickly returning those requests that have already been cached. The accelerator resolves cache misses and dynamically evaluated queries by contacting your real HTTP server to evaluate the incoming request. The httpd accelerator is compatible with both CERN httpd and NCSA httpd (it may work with other httpds as well). It is easily installed, and is available for SunOS 4.1.x, Solaris 2.x, OSF/1 2.0 and 3.0, Linux, HP-UX, AIX, and IRIX (note that at present we only support the first 3 of these platforms, however). The httpd accelerator services cache hits blazingly fast because it: - runs as a single (threaded) process that never forks, - is implemented with non-blocking I/O, - keeps meta data and especially hot objects cached in RAM, - caches DNS lookups, - supports non-blocking DNS lookups, and - implements negative caching both of objects and of DNS lookups. Our measurements indicate that the NCSA httpd (running in standalone mode) can handle 3 requests per second, while cache hits to the Harvest httpd accelerator are serviced at 200 requests per second. For more information about the httpd accelerator, see http://harvest.cs.colorado.edu/harvest/httpd_accel.html For information about Harvest (including demos, papers, software, and documentation) see http://harvest.cs.colorado.edu/
Received on Tuesday, 4 April 1995 17:41:11 UTC