“How to determine Internet speed”
By Scott Hamilton
I have noticed a lot of advertising recently for new “high-speed” rural Internet coming in the mail lately in Edgar Springs and surrounding areas, so I thought I would give you some pointers if you are looking at changing providers. I will give you some questions to ask the provider as well as some tests to run if you want to see how your Internet performs. There is one classic mistake in Internet marketing and it has to do with the metrics used to measure the speed of the Internet.
There are two factors that impact the performance of websites when you are surfing the web, and they are not equally important. The really surprising thing I am about to share is that the Internet marketing materials all promote themselves using the metric that is far less important for most people. The metric you read in nearly every Internet service provider’s advertising is bandwidth, so we will start with that one.
Bandwidth is the measurement of how much data can be transferred across the wire, or air if you are wireless, from the server on the Internet back to your computer. I like to think of this like the number of lanes on a highway. I am sure you have all experienced a traffic jam from an accident on I-44 in Rolla, Mo.; if you have not you are extremely lucky. Internet bandwidth works in much the same way. Internet service providers have wires that are like super-highways with thousands of lanes for Internet traffic. The bandwidth on these super-highways are in the 10s of terabytes per second range. A terabyte is 1,000,000 megabytes so we will use the megabyte as a lane for simplicity. The connection coming into your home is like a multiple lane exit from the super-highway containing only a relatively few lanes, most of the providers in our area offer less than 20 megabytes per second download and around five megabytes per second upload. This means we have five lanes to enter the highway and 20 lanes to exit the highway and there are 10,000,000 lanes on the highway that we share with all our neighbors.
As you can see, if we try and upload too much data, or download too much data we wind up waiting, just like a traffic jam on the highway. If you are streaming videos you need a minimum of three lanes for each video, and for ultra-high definition videos you need closer to 10. This is a good measure for knowing how well the connection will serve your family when it comes to streaming movies and videos, but what about reading e-mails, online gaming, checking Facebook, Twitter, and other social media platforms? These types of activities require a second metric called latency.
Latency is a lot harder to describe as it has to do with round trip traffic, and is really a factor of how far away from the source server your computer is located. It can be described like the speed limit on the super-highway, or at least that is a part of latency, the distance the traffic needs to travel is another and probably a more important part. Think of latency like delivering a message by truck and waiting for a response. If the truck needs to drive from Licking to Edgar Springs and back again (15 miles) at an average speed of 60 miles per hour (mph), the round trip will take 30 minutes. If the speed limit is increased to 120 mph the round trip will take 15 minutes, but if the message only has to go from the south end of Licking to the north end of Licking (five miles) it will only take five minutes at 60 mph and 2.5 minutes at 120 mph. The latency is the time needed to make the trip. Internet data travels at around 67 million mph, so it takes a really long distance to make a major impact. This is one of the primary problems with satellite based internet; though they can provide 100 gigabyte per second bandwidth (think 100 lanes) the data has to travel to space and back, making for high latency (think driving round trip to Chicago instead of Rolla).
Having these facts the question to ask any new Internet provider is about their latency. A paper written by two researchers from MIT shows that for every 200 milliseconds of latency we can save, we cut the page load time by over one second. You can find out more about their research at https://blog.cloudflare.com/making-home-internet-faster/. You can have all the lanes you want, but if the round trip latency is too high you will still have to wait a longer time for web pages to load, and online gaming will be nearly impossible on high latency networks like satellite.
If you want to know how your Internet connection performs on both metrics, you can get the measurements from http://www.testmy.net. For me, on UScellular home internet I get an average latency of 697 milliseconds (ms) to New York, NY; 524.9 ms to San Francisco, Calif.; and 20 ms to Rolla, Mo. On HughesNet satellite internet, the latency to Rolla, Mo., is 600 ms, which is 30 times longer, meaning web pages will take about three seconds longer to load on HughesNet, according to the MIT study mentioned earlier. As you can see, latency is an important metric that most providers will not share with their customers and usually has a large impact on the perceived speed of the internet. Until next week, stay safe and learn something new.
Scott Hamilton is an Expert in Emerging Technologies at ATOS and can be reached with questions and comments via email to sh*******@te**********.org or through his website at https://www.techshepherd.org.