3 Comments

Summary:

Broadband speeds as advertised by service providers are only part of the equation when it comes to assessing overall broadband quality, according to a study out today by the Oxford University’s Said Business School and the University of Oviedo’s Department of Applied Economics. The study, which […]

digital data flow through optical wireBroadband speeds as advertised by service providers are only part of the equation when it comes to assessing overall broadband quality, according to a study out today by the Oxford University’s Said Business School and the University of Oviedo’s Department of Applied Economics. The study, which was sponsored by Cisco (a company that would love to boost broadband speeds so it can sell more gear), ranks the U.S. 15th in terms of broadband quality, which it assesses based on download speed, upload speed and latency.

We write a lot about both download and upload speeds but latency isn’t something we bring up that often. The speeds we typically focus on are a function of how fat a pipe is — the number of bytes that can be sent per second — while latency is a measure of how long it takes for what you asked for to get to your computer. It’s affected by download speeds (it takes big data a long time to squeeze through skinny pipes), as well as by how quickly the servers respond.

But as we consume more streaming services, try to engage in video conferencing, and expect a more real-time web, latency is increasingly becoming an issue. Imagine if you’re having heart problems and need to have your heart rate monitored remotely and sent to your doctor over your broadband connection for review. It’s a small amount of data, so having a fat pipe isn’t important, but you care desperately that the latency is as low as possible so that it gets there quickly in case your heart stops first.

The study notes that both latency and upstream speeds are going to become more important when it comes to measuring the quality of broadband connections over the next 3-5 years, and have subsequently changed its formulas to give both more weight. A focus on speed needs to also be a consideration in the debate that’s currently playing out in Washington, D.C., as the FCC tries to figure out how to define broadband as part of its National Broadband Plan.

latent1

Chart provided by FirstMile.US, which filed it with the FCC after surveying its members as to which aspects of broadband performance should be measured.

  1. Those factors are very important to the broadband experience. One I would put up there with those in importance is packet loss, which could be considered a form of latency–infinite latency that is.

    Share
  2. Somewhat frustrating that this issue has been known about for some time:

    http://portal.acm.org/citation.cfm?id=1022596

    Share
  3. [...] see this ACM article, Latency Lags Bandwidth (registration required), and this GigaOM article, Pain at the Pipe: Latency Matters.If you enjoyed this article, and want more like it, follow FTTxtra with an RSS feed or [...]

    Share

Comments have been disabled for this post