Doghouse – IoT Need For Openness

Recently, I upgraded my home Internet to fiber and 1Gbps symmetric service, so I theoretically get 1Gbps communication up and down from my provider.

For most people, that translates to a thought pattern of "I can move data at 1Gbps from my laptop/desktop/phone" which can be fairly far from the truth.

Because of overhead in communications, you can look for between five and 10 percent loss in your throughput, even in the best of all possible situations. Therefore you might see 995Mbps to 990Mbps as actual data throughput, even with a Cat 6 cable of four feet hard wired to the LAN ports of your WiFi router.

Recognizing that WiFi would probably be slower, and likewise, the server at the other end would help determine data speed, my partner and I used the fiber provider's speed test. We expected that it used a server (or generated the bits at their end) for the fastest test possible.

My partner and I measured 990Mbps as download speed through the router, but the upload speed was considerably worse, coming in at only 500Mbps, and that caused a temporary puzzle.

Fortunately, my provider offers an interesting application that keeps track of the bit traffic from every single hardware address associated with each device. Therefore I can see how many bytes are being uploaded and downloaded from each device – each day, each week, and each month.

From this, we could see that some devices were uploading, doing backups of data that were scheduled for a late-night time.

We could also see how the data was divided up among other sources that many people might not consider. In our case, it was a light bulb.

We use a lot of WiFi devices to control our home. WiFi-controlled speakers, Chromecast for video broadcasting, etc., but these were "off" at the time we were testing the speed. What we did not anticipate was the fact that a WiFi-controlled light bulb, turned on and off typically once in a 24-hour period, might download 14.41MB of data and upload 5.93MB of data each day. Wow. I might understand the 14.41MB down for a firmware update … but every day (on the average)? And why close to 6MB uploaded just to do what? Report back to some entity that the bulb had actually provided the light in the room … or not?

We have only one light bulb like this in our home, but we do have a series of WiFi electric outlets. Each one had similar upload and download traffic.

Our Internet provider noted that "Homes that have many of these types of devices (many being in the dozens of units) may experience less than optimum download and upload speeds." Duh.

All of this is affecting the speeds that the Internet user sees. This does not take into account the sharing of WiFi inside the house by all these packets that are simply handled by the router itself but that take away bandwidth from other WiFi uses.

All of this melded into a discussion among some friends of mine about the need for symmetrical Internet speeds, Internet of Things (IoT) discussions, and how these devices are using both electricity and Internet even when they are "off."

A lot of these devices are sending all of their communications back to the company that created them. It would have been better if these devices had a centralized "server" unit that could process a lot of this data, called "edge programming," boiling down the raw data and distributing firmware updates to all the devices using the same Internet sub-network.

It would be great if these devices also used different frequencies (perhaps the emerging 60GHz WiFi), or perhaps I need to spend more time setting up quality of service (QoS) for the router and Internet. After all, if it takes a fraction of a second more to turn a light on or off it probably will not make much difference.

One friend of mine said that they isolate all of these devices on a separate network, blocking them from "calling home." If the devices stop working, they get rid of the devices. Of course, they lose both the money paid for them as well as the service of the device.

Despite being a free and open source advocate for many years, I have so far taken a laid back position on these IoT devices, but in the next year I am going to refresh my home of 40 years, spending a considerable amount of money, and part of that will be to only purchase "free" IoT devices. This could also include purchases of significant appliances.

It is too bad that manufacturers of closed devices will not be able to receive my money, and the money of many more people who feel the same way.