What’s your broadband speed? There has been much said lately about measuring the speed of broadband connections, especially as relates to the FCC’s National Broadband Plan. The FCC has even placed two different tools on their Broadband.gov website in an attempt to provide consumers with information regarding the speed and quality of their broadband connection. Both of these tools, one from Ookla, and the other from Measurement Lab (founded by the New America Foundation’s Open Technology Institute, the PlanetLab Consortium, and Google) have their methodologies published on the FCC website. FCC Chairman Julius Genachowski has said that “Transparency empowers consumers, promotes innovation and investment, and encourages competition, and it seems that placing these test on the Broadband.gov website is an attempt to foster that idea.
Business customers need to be aware, though, many factors come into play that can affect the results of these tools when trying to test the speed of a broadband connection. Starting with the computer that is actually performing the test, there is the possibility that it is resource constrained. For example, if the computer has other software running during the test, such as a screen saver, email application, antivirus application or business applications, these other programs could limit the resources available to the testing application and the required Java environment, which can affect the outcome of the test. And let’s be honest, that’s a very common occurrence.
Moving away from the focus on a single computer, there could be additional business traffic on the network in your office or store. For example, if the broadband speed test is being performed while other users are active on the network — for example, downloading inventory reports, watching training videos, or receiving email — again, the results of the test can be affected.
The type of network the computer is on can also affect the test. For example, if the computer is connected to the Internet via Wi-Fi, delay can be introduced into the test results and affect the test, as the wireless path can be degraded by various types of interference.
And then there are many issues that come into play with the architecture of the Internet, and the technology of the connection you are using. There are very different physical characteristics between DSL, Cable, T1, Ethernet, Fiber, Satellite and 3G Wireless networks. Each of these network technologies, by definition, can affect the results of the test. Also, as you can see from the methodology descriptions, the test makes use of TCP and DNS, which can affect the results. The tests make a best guess to determine the location closest to your networks physical location, based on IP address look ups; these are educated guesses, but they can still be wrong, and you may find yourself testing against a site that the test thinks is logically close, but due to the complexity of Internet traffic routing could actually be far from your network.
These speed tests are trying to measure a changing, dynamic environment, and the two endpoints of the test (your computer and the test server) may not reflect the path your business traffic actually travels. There are many points along the path the test packets take, starting with your computer and your local network, and including your service provider’s network, the other networks your service provider peers with, content delivery networks that can sit as on overlay on these networks, the path the DNS server look up takes, the access network the test server sits on, and the test server itself, which could be resource constrained by the number of simultaneous tests being performed.
The net-net of bandwidth and business broadband speeds
Business users need to know that they are getting the bandwidth they require to communicate with their customers, partners and employees. Performing these tests on your broadband connection will give you one data point when you are making a decision on upgrading bandwidth, but there are many other factors that you need to take into consideration. What is the nature of the traffic you need to move across your network? What business applications are you running? How many users will be sharing the same connection? At what time of day do you need the best network performance? Is your network traffic secure? Does your network provide Quality of Service to make sure that your voice traffic has priority over the emails and web traffic moving across the network?
For consumers, the results of multiple speed tests over time, and performed at different times of day, will give them some information regarding the performance of their home network connection. When performing a test, they should make sure to be the only user of the network (i.e., the kids can’t be watching YouTube upstairs), and be sure to close all other applications on their computer. Businesses, though, should work with their Managed Service Provider to make sure that broadband speed test results are one data point taken in context with all of the other complexities involved in operating a next generation business network.
Posted on the MegaPath Connectivity Blog.