World wide web has been more and more connected to people's daily
life. Internet explorer and Netscape are two main commercial web browsers.
In both cases, browser cache is used to accelerate the browsing process.
But how does browser cache work? Does it help the user browsing experience?
In our project, we did a study on web browser cache and related user browsing
behavior. All the study is on IE v5.0. Finally, we implemented a cache
monitor tool to help user make more intelligient usage for cache. We found
that the cache replacement policy is based on LRU with the minimum cache
stay of 10 minutes. In addition, file visiting frequency and embedded graphic
file is taken into consideration. When cache is full, old files are flushed
out until the cache size reduced to ~85% full. In evaluating the efficiency
of the cache, cache hit rate and byte hit rate are used. Cache hit rate
is more biased to small files while byte hit rate represent more accurate
cache usage information. The trace file shows that most accessed files
are small. The LRU policy will be more favor to small file since file size
isn't taken into consideration. However, large file miss hurt the performance
more than small file, we think it is a good idea for cache management
policy indcluding the file size information. Finally, the survey
shows that users are relied more on the system intelligence rather than
themselves. The cache function in IE is usually not fully utilized by users.
Microsoft Powerpoint version
Postscript
Source Code
Contact information:
Yong Lu: lu@cs.wisc.edu
Bo Zhan bo@cs.wisc.edu