Performance Improvement Studies in Accessing Web Documents
As the World Wide Web has now become the standard interface for interactive information services over the Internet, the perceived latency in WWW interaction is becoming an important and crucial issue. Currently, Web users often experience response delay of several seconds or even longer to non-loc...
Saved in:
Main Author: | |
---|---|
Format: | Thesis |
Language: | English English |
Published: |
1998
|
Subjects: | |
Online Access: | http://psasir.upm.edu.my/id/eprint/9442/1/FSAS_1998_30_A.pdf http://psasir.upm.edu.my/id/eprint/9442/ |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Summary: | As the World Wide Web has now become the standard interface for interactive information services over the Internet, the perceived latency in WWW
interaction is becoming an important and crucial issue. Currently, Web users often
experience response delay of several seconds or even longer to non-local Web sites
especially when the pages they attempt to access are very popular. For WWW to be
acceptable for general daily use, the response delay must be reduced.
The potential solutions to the problem lie in the extensive use of caching (disk
based) and prefetching in WWW. Both caching and prefetching explore the patterns
and knowledge in the Web accesses. This thesis describes and tests the efficiency of a batch prefetching update
(refreshing) in accessing HTTP and FTP documents on the global Internet. The
update is scheduled to run at idle time when the traffic is less congested and the
server activity is low. The batch refreshing effort would be fruitful when the
refreshed documents are really requested before they tum stale again. The
effectiveness of the batch refreshing is verified by running a statistical analysis of the
access log files.
In the first part of the study, a Proxy Server at the LAN of FTMSK, ITM was
set-up, configured and monitored for the use of 400 users. Access log files are
collected and analysed for a period of six months. The analysis result would be a
benchmark for the caching proxy with batch refreshing in the second part of the
work. |
---|