While using YSLOW for performance testing, I have noticed that YSLOW gives different statistics when site is accessed differently. I mean when I accessed our local intranet site through its IP address, YSLOW gave me minimum page weight. Then I used local host name and run YSLOW on it, this time it showed slightly increased page weight. When I accessed the site through full qualified domain name, YSLOW showed maximum page weight, almost doubled the weight this time. On expanding Gzip components I found that on one of the javascript component gzip reduced weight from 450KB to 111KB when local host name was used but the same 450KB js component was not gzipped when Full Quality Domain Name was used. Can anyone help me understand this difference?

Ambreen Ali