Web pages are made up of multiple objects. This first object fetched is
an HTML document, which tells your browser which additional objects it
should load and how to display them. Additional objects could include
scripts, style sheets, and images.
For complex pages, these objects are often served from different
on one server and its images on another. Sometimes, sites serve pages
differently over HTTP and HTTP/2 (e.g., serving CSS in a separate file
in HTTP/2 that was previously inlined directly in the HTML):
These results compare the HTTP and HTTP/2 versions of the sites we
crawled in terms of:
Number of Objects: How many objects did each version of the page load?
Number of TCP Connections: How many connections did our browser make when it loaded each version of the page?
Number of Domains: How many servers did our browser contact when it loaded each version of the page?
Reading a CDF
The plots on this page are cumulative
distribution functions (CDFs). A CDF describes a set of
measurements by showing the probability that a measurement will take on
a value less than or equal to x. For example, the CDF below
shows that 50% of HTTP 1.1 sites contain 64 objects or fewer:
You can also read it “in reverse”—for example,
30% of HTTP 1.1 sites contain more than 100 objects: