Since the beginning of the new year, I upgraded our broadband from a copper wire service, specifically Time Warner Cable’s Road Runner, to a fiber optic connection from Verizon. Immediately, we more than doubled the downstream throughput: from an average 11 Mbps to more than 25 Mbps. The biggest difference, however, came in the upstream. Whereas our upstream was about 500 kbps with the copper wire cable, it is now about 25 Mbps with fiber optic. Theoretically, I can now upload files at about 50 times the speed. None of this, however, should come as a surprise. Not only is fiber more capacious than copper wire, the Verizon 25/25 service is also more expensive than Road Runner’s 10/0.5 service.
Independently, we also signed up for SamKnows broadband monitoring, which is how the FCC is measuring the nation’s broadband capabilities. Each month, SamKnows sends a summary of our broadband connection.
Here is our downstream…
…and our upstream.
As you can see, the connection is pretty stable. It is consistently around the advertised 25 Mbps, which is good. Uploading and downloading files always seems zippy, and when we stream video, it always looks very sharp. I never notice any buffering delays or any sudden pixelation in the picture.
However, two other very important metrics have been off. Our latency and our packet loss, which should be low, have been inordinately high.
Even with good throughput, latency can slow down broadband by delaying the connection with a remote host. Similarly, packet loss can cause errors that will result in either a “dropped” connection or requiring redundant, duplicate connection.
As you can see, it looks like those two metrics have dropped significantly around March 24. I wonder if there was an equipment upgrade on that date that might have improved the “fidelity” of our broadband connection.
From a purely observational perspective, I haven’t noticed any improvements in our connection. In fact, I thought our connection had been getting slower over the last week, but, of course, these kinds of observations can be deceiving.