Is coax networking slower than Ethernet?

Here at Solid Signal, we’re all about helping you find ways to make life better. For example, we’ve showed you how to network your home using just your satellite cables, and several months ago we did a feature piece on using spare coax lines to run computers or connected equipment. That piece got a few comments, and this was one:

(…) while I know DECA is limited to 100 mbps (vs. the 1 gb I would get with a straight CAT6 Ethernet cable connection to my router), the results from Speedtest are perplexing. Using the WiFi-N connection (with the status displaying a 300mbps wireless connection), my Speedtest.net results with my TWC Ultimate 50/5 service are 113 mbps down, 5 up (and a ping of around 10 if I remember). The 113 speed may be due to TWC upgrading LA service for free from 50 to 100. However, doing the test with the DECA hardwire connection (and WiFi off), the results are about 14 down, 5 up and ping around 40. The Windows Ethernet adapter status shows a 100 mbps connection. Why is there such a tremendous speed degradation using DECA?

That’s a great topic for a test… is coax networking slower than plain Ethernet? Let’s find out!

Test parameters

To create a fair environment for this test, we created a complete home network with DIRECTV equipment and wiring. Because some of this equipment was borrowed from Solid Signal staff, we won’t post exact model numbers, but if there is a question they can be provided privately upon request.

Here’s what was on the network:

HR44 Genie DVR
HR24 DVR
DECA II Broadband Cinema Connection Kit

PC Laptop with Broadcom 1000BaseT Ethernet and Broadcom 802.11n networking
Cisco Linksys router
Seagate NAS

The satellite system was setup and both DVRs were powered up and functioning. The system was connected to Verizon FiOS 50/25 service with no other devices on the network.

The goal was to compare performance of the laptop when it was connected via wired Ethernet, Wi-Fi, and Coax networking, both by connecting to the internet and to the NAS.

Hypothesis

Here’s what we believe will happen:

We believe that the NAS (which is basically a hard drive with a network port) would not be able to deliver Gigabit Ethernet performance, so we believe that the numbers we get from the wired ethernet test to the drive will represent the drive’s fastest speed.

Because the computer is in the same room as the router, we expect to get close to the max speed of 300Mbps when connecting through Wi-fi.

Coax networking has a maximum speed of 320Mbps but the adapters themselves are limited to 100BaseT speeds of 100Mbps, and we expect that we’ll get something similar to that performance when connecting through them. We’re going to test whether it makes a difference if just the computer is on coax networking, or whether adding both devices to the coax network improves things. We expect a modest improvement when both devices are on the coax network.

Test #1: Speedtest.net

Speedtest.net is the most commonly used speed test on the internet. It uses media files of different sizes to get a good sense of the user’s internet speed, and it does not store any information on the user’s computer so the test really measures internet speed and not disk speed.

What we found was not surprising. Whether connected by Ethernet, Wi-Fi or coax, the results were about the same. They surpassed the rated speed of 50 down, 25 up.

We weren’t able to duplicate the claim that internet speeds were slower over coax than they were over Ethernet.

Test #2: Speedof.me

We did a second internet speed test using speedof.me. Speedof.me uses a different system than speedtest.net — speedof.me sends multiple files of different sizes and stores them on the local hard drive. We chose this test because if there was a possibility that using the hard drive would cause the speed to drop using coax networking, we wanted to see it.

There was a slight difference in speed between Ethernet, Wi-Fi and Coax using this test, about 10Mbps or 16% of the total speed. The drop was there on Wi-Fi and Coax equally, and it’s possible that the speed difference was due to read and write speeds on the hard drive. While 10Mbps was significant, it’s nowhere near the 80% reported drop in speed from our commenter.

Test #3: Local file transfer

If there is a speed difference to be found, we figured it would show up with the local test. For this test, we copied an 807MB file from place to place and timed the actual transfer time. This probably has a margin of error of +/- 1 second.

Here’s where the results get interesting. It’s clear we found the speed limit on the drive, since it wasn’t able to get all the way up to 1000Mbps on the Ethernet test, topping out at 589Mbps when receiving a file. The big surprise here was how poor Wi-fi was — with no other computers on the network and a clear line-of-sight to the router we expected a lot better.

We connected the computer via a DECA and then tested the results with the NAS connected via Ethernet and then via DECA. What we found was there was no difference in how the NAS was connected; either way we got within 10% of the rated 100Mbps speed of the DECAs which is what you would expect from any 100BaseT network.

Conclusions

We reached some “throwaway” conclusions that weren’t terribly important to this test:

  1. The NAS we tested is a lot better at receiving files than sending them. This may be due to a limit on the NAS’s hard drive which may be faster at reading than writing.
  2. Wi-Fi performance is nowhere near as good as it should be in this scenario.

We also concluded that in this scenario, there isn’t a really meaningful difference between coax networking and wired Ethernet. In either case it’s possible to exceed the rated speed of the line and it’s possible that any fluctuations are due to factors that have nothing at all to do with the coax network.

It’s also reasonable to say that for everyday video streaming, which tops out at 10Mbps at its fastest, either coax or Ethernet will work fine, and since it’s possible to do a 4KTV stream in well under 60Mbps, the 90Mbps top speed of a DECA is a total non-issue.

For our commenter who got much worse performance, it’s hard to say why. Perhaps the coax network was congested due to showing programs across receivers or perhaps one of the lines was degraded. Leave a comment here and we’ll try to help you figure it out.