Web Caching and Replication

6.3. Object Types and Cacheability

In the "good old days" (which were only a few years ago, come to think about it), there were only a few kinds of Web objects: static HTML pages and images and documents generated dynamically by CGI scripts. Today, a plethora of object types can be accessed via a standard Web browser.

At the present, an overwhelming majority of all downloads, in terms of both the number of accessed objects and bytes downloaded, involve HTML pages and images, with images outnumbering HTML pages. Together, they account for around 90 percent of all Web accesses and, typically, over 70 percent of downloaded bytes, although the reported byte fraction varies more widely, from 60 percent to over 80 percent [Douglis et al. 1997a; Williams ...