Throughput 101: Don't Guess At File Sizes

I enjoy calculating real throughput on real networks with real devices. In many cases, the easiest and most realistic test is a file transfer using SMB, FTP, HTTP, etc.

The main point of this video is to introduce you to basic testing best practices. For example, you should be using an ftp script or wget to keep the test consistent, take multiple measurements (in some cases at different times of the day), and approach the test with an idea of how long it should take. That way you immediately have an idea if you should investigate further.

The intent isn't to compare or debate if file copies are "better" than generating a stream like iperf. I'm simply showing you how to do the math and calculate either the file size or duration of a test.

I make a point early in the video that you should take my saying, “You are only as fast as your slowest link” literally. As I mentioned, in some cases the slowest link is not a network connection, but an interface in the computer. Examples include USB, Firewire, Inifiband, Fibre Channel, and SATA.

Enjoy your testing.