Communications and Technology Industry Research

A second annual survey of wired network Internet speeds show that the United States, sponsored by the Communications Workers of America, suggests that speeds have increased in both downstream and upstream directions, but perhaps not much. The median download speed for the nation was

 

2.3 Mbps, while the median upload speed was 435 kbps.

 

In the 2007 survey, downstream speeds average1.9 megabits per second and the median upload speed was 371 kbps. On average, downstream speeds increased by 900 kbps while upstream speeds increased 64 kbps.

 

The 15 percent of U.S. Internet users who connect at dial-up speeds were not included in the survey. The CWA also notes that 33 percent of Americans do not use the Internet.

 

One might question the findings, though. I took the speed test on the speedmatters.org site, using at T1 line, and repeatedly was told my download speed was  between 14 kbps and 18 kbps, with an upstream speed of 685 kbps or so, during repeated tests about 5:30 p.m. local time. Using other speed tests, I got dramatically different results.

 

Using bandwidthplace.com, I got a downstream reading of 717.97 kbps. Using the Los Angeles server on Speakeasy (News - Alert).net, I got a downstream reading of 1.225 Mbps and an upstream speed of 1.343 Mbps.

 

A sequential test using dslreports.com and a Los Angeles server showed a downstream speed of 1.195 Mbps and an upstream speed of 1.310 Mbps.

 

The cnet.com server says my connection was running at a low of 28 Mbps upstream and downstream, while another test suggested my downspeed and upstream speeds were in excess of 33 Mbps.

 

A test using myspeed.visualware.com showed a download speed of 743 kbps and an upstream speed of 1.4 Mbps.

 

A test using www.performance.toast.net showed downstream speeds of 919 kbps or 936 kbps.

 

The www.att.net test showed downstream speeds of 782 kbps and 993 kbps, with uploads of 1.432 Mbps and 1.435 Mbps.

 

All the tests were conducted between 5:30 p.m. and 5:55 p.m. on Wednesday, Aug. 14, 2008.

 

The test results using the speedmatters.org site clearly were the most random of all the tests. The downstream tests were wrong by two orders of magnitude, the upstream test off only by an order of magnitude. If that is the source of the CWA report, I take little stock in it.