Robin Laurén <[EMAIL PROTECTED]> writes: > My question is about the number on one of the last lines of the > logged output, the reported download speed. What exactly does > wget's download speed report? Is this the speed of just the data > downloaded, or does the value include the lag time between the > issued request and when the data starts to arrive? In other terms, > does the clock start ticking when i press enter or when the first > chunk of data arrives?
Neither. The clock starts ticking when Wget starts waiting for the first chunk of data, which happens after the HTTP response has been read. Therefore "download speed" refers only to payload (body data). Additionally, for HTTP, initial part of the data can come in the packet carrying the response headers, which is why Wget sometimes reports ridiculously large download speeds when downloading small files. For example: $ wget http://www.apache.org/style/style.css ... 00:59:40 (86.9 MB/s) - `style.css' saved [1002/1002] I don't have a gigabit connection to www.apache.org; it's just that the whole 1K file was delivered along with the HTTP response header, fooling Wget into thinking the data was downloaded in a *really* short time. Fortunately this problem only appears with really small files.
