Measure your end users’ bandwidth

Use bandwidth measurement to learn the differences in your end users' experience and adjust your delivery to meet their needs. Knowing your end users' bandwidth can help control data center costs and resource consumption as well.

Bandwidth measurement is a option that’s enabled for a short period of time. Usually a few days to test the impact of a new release or feature, and then it’s disabled. It’s also common to run a sample of your end users rather than running it against all your end user.

Set up a bandwidth test to filter page load times and other mPulse data by bandwidth range. This can help you to better recognize how your end users’ experience might vary.

A bandwidth test setup includes test images of varying sizes that you place on either the origin or a CDN. Each time a bandwidth test runs, mPulse downloads the images from smallest to largest. mPulse measures page load times until all of the images load. Or until an image exceeds a predetermined time, at which point the test stops. In either case, mPulse calculates bandwidth from the accumulated data.

Where you place the mPulse test images depends on what you want to measure.

Consider the following:
  • When placing the test images on your origin, the images are not cached on the client. mPulse uses cache-busting query string parameters to force an image reload. This ensures that mPulse tests the bandwidth to your origin and not the bandwidth to a local cache. Also, browsers ignore Cache-Control, Expires, and Last-Modified headers sent with the image. To make sure that the test is accurate, it's best to send as few headers as possible when serving these images.
  • When placing the test images on a CDN, the CDN sends a call back to your origin server each time an end user makes a requests. This may not be the best use of your CDN. Configuring your CDN to ignore query string parameters for these images can be an option (see Set up caching for bandwidth images in Property Manager).