Tag Archives: IoT

My experience with the Lifx no-hub smart light

LIFX

Note: Review follows video

Disclaimer: I have no connection to this company, in fact never heard of them until about four months ago. They didn’t ask me for this review, they didn’t pay me, didn’t send me the bulbs, I bought them on my own, this is all my own opinion, as it always is.

I recently needed to get a smart light, one I could control remotely. You know the kind of thing; turn it off and on once in a while, make it look like someone’s home, give the illusion of life. The thing was, I had never looked into these kinds of lights too deeply, and I still haven’t so this isn’t a condoning or condemnation of other brands, but rather my impressions of the one I ended up getting.

I went to the local Best Buy, a store I rarely visit, because I knew they had a lot of them and I was going to be shooting into the wind. My need for this light happened rather suddenly, so I didn’t have time to research and just went in blind. I looked around at the options on display, from well known brands like Philips Hue and lesser known brands like Sengled. At least I think the latter is lesser known..I had never heard of them, but they have a lot of options, and to be fair I’d never heard of the brand I ended up deciding on either.

The issue I have with most of these lighting systems is that they require a hub, or as Philips refers to it, a bridge. I explicitly did not want that, as I was only intending to get a single bulb; no need to complicate it. That’s why I ended up taking a chance on another brand with which I was unfamiliar: Lifx.

Let’s just get this out of the way right now: I had no idea how to pronounce that. “Life-x?” “Liff-x?” No clue. After some Wikipedia-ing, I discovered it’s “Life-x,” and it’s a company that grew out of a successful Kickstarter campaign. I chose them because they do not require a hub and offered the basic functionality that I needed on short notice, and it turned out to be a good choice, with some important caveats.

The specific model of Lifx light I decided on was the Mini Color, which advertised a light output of 800 lumens. That’s pretty good; many LED lights advertise themselves as the equivalent of some wattage, but their lumens are absurdly low, sometimes rating at 300 or 450, which is very dim. 800 lumens is the actual equivalent of a 60-watt bulb, and that’s exactly what I was looking for.

I also liked the fact there was no hub. I wasn’t looking to set up some centrally-controlled network of devices, I just wanted a light bulb, and Lifx fit that bill. In fact, if I recall correctly, it was the only one that did; all the others required some kind of central device to which they would connect, yet since IoT devices like these form a mesh network anyway, the need for a hub when simply using light bulbs just isn’t necessary.

I took it home, screwed it in, and was guided through a relatively simple process (on Android; I can’t speak to IoS) to connect my phone to the light and the light to the network. Once that was complete, I was up and running.

I only wanted to get a light for a garage, but once it was set up I couldn’t help but go through the options available to me in the app. There are four separate screens for normal operation: ‘Colors,’ ‘Create,’ ‘Effects,’ and ‘Day & Dusk.’ There is also a main screen from where you can access groups, Nest integration, IFTTT features, as well as some other integrations and even buy more lights.

I discovered there was much to like about this bulb, and the more I experimented with the app the more pleased I became, even though I had no real need for any of the more esoteric features. I’ll start with the app’s individual feature screens and come back to the main page at the end, because one of the major issues I have manifests there.

The first screen, ‘Colors,’ actually gives you two choices: You can control the white temperature of the bulb, measured in Kelvin, which ranges from a very cold, blue 9000K, like the blue LEDs you see in icicle lights around the holidays, to a very warm 2500K that represents the more amber tone of an incandescent bulb, or even a candle. Incidentally, I know that the higher temperature is referred to as cold and the lower warm, but that’s just the convention here. Also, the ranges this light bulb offer are way past the choices you normally have, giving a much wider set of options for temperature.

Color temperatures

Color temperatures

You can select the color temperature by spinning a wheel of temperatures, as it were, to select the temp you’d like. It’s very easy, however it’s also discreet selections, so choosing along a continuum isn’t available; you’d have to do that on the actual color screen discussed next. Not only that, when switching between the coldest blue and warmest warm, a sort of amberish, yellowish hue, the bulb flashed a BRIGHT yellow, which was curious.

Even with the choices it provides for white temperature, I still didn’t find the warm setting to be terribly accurate in terms of its similarity to a warm incandescent, or standard bulb. It isn’t bad, and it’s bright, but it still comes off as artificial. The cold temps, though, the blues, holy cow: They are BLUE. If that’s what you’re looking for, this bulb delivers in spades.

If you are feeling more festive, a Tinder-esque thumb-swipe to the right and you can select from a range of actual colors. It works the same as the white temp screen, by rotating a hue wheel and determining the saturation of the color you select by adjusting a slider on the color wheel. It’s ingenious, really, in its simplicity: Rotate the ring to the color you want, slide the slider to select intensity, and that’s it. It’s very responsive, easy to make adjustments, and easy to use.

Color selection ring

Color selection ring

Another nice thing about these screens is that you can dim the light from them as well, using a simple slider. It dims quite far, something not all LEDs can do. Normally, a light bulb dims by reducing power to it, but LEDs can often only dim to about 10% before being cut completely. The Lifx app claims the dimming goes to 1 percent, but it doesn’t look like that to me, although it does seem to dim further than most other LEDs.

The next option, ‘Create,’ I don’t use that much. It’s misnamed, as you don’t actually create anything here but select from pre-designed themes such as ‘Relaxing,’ ‘Energizing,’ ‘Peaceful,’ and ‘Powerful,’ the image for which is a tropical sunset for some reason. Doesn’t quite match up with Powerful but it’s a minor issue. There are others, and selecting one just changes the color of the light to match up to what I suppose will enhance the chosen mood, and you can change these in a different screen.  The second screen under ‘Create’ is just a matrix of colored circles in a ROYGBIV arrangement that you can choose, although I don’t know why you would use this as opposed to the color wheel. Perhaps if you just wanted, say, pure orange and didn’t want to have to make fine adjustments to get it.

Themes / Colors

Themes / Colors

The next screen, ‘Effects,’ is my favorite because of one in particular. There are eight effects to choose from: ‘Animate theme,’ which allows you to modify themes from the previous screen, ‘Color cycle,’ ‘Flicker,’ ‘Music Visualizer,’ ‘Pastels,’ which I would never, ever use, ‘Random,’ ‘Spooky,’ and ‘Strobe.’ Because of the potential health issues that come with strobe lights, you have to hold that one down to use it.

Effects

Effects

They’re all self-explanatory, but I have to highlight one and call out another. I loved the ‘Spooky’ effect: Being a fan of horror movies, this one emulates the horror movie trope of the abandoned hospital or car park that has the flickering, randomly flashing light. When selected, it flashes the light randomly for 60 seconds (the minimum, which I REALLY wish could be shortened), then go bright red, then turn off. Beautiful. The big problem with it is that when controlling a grouping of two lights, the ‘Spooky’ effect only worked with one light, even when controlling them as a group. That needs to be fixed.

‘Music Visualizer’ is the one I have to call out. It flashes the lights all over the place when it hears noise, monitored, as it claims, through the phone’s mic. However, when I shut off all music and all sounds, and covered the phones mic, they still flashed randomly. Also, while testing using songs with a heavy rhythmic component from AC/DC and Metallica, the lights flashed randomly, not rhythmically. Therefore, I can only surmise that it is not actually monitoring the sound and just making random color changes. That’s a shame if true, because a visualizer would be a neat feature.

Finally, there is Day & Dusk, which gives options to have the lights come on and go off at certain times of day, intended as a wake up and sleep thing. You can also set timers to change color or temperature at certain times of day, which may be necessary based on ambient light, working conditions, or other factors. I don’t use this feature, but I like it, and feel it would be very useful to many people.

Day & Dusk

Day & Dusk

Finally comes the main screen. This shows you your lights, your groups (you can group lights together so any changes affect all lights in that group; it’s a fantastic feature and one that really elevates the usefulness of the app, as does being able to switch between multiple locations using a dropdown), create IFTTT (If This Then That) rules, and buy new bulbs. The IFTTT integration makes me worry for humanity: You can set the lights to react to various events, some of which are very useful like blink lights when your Uber arrives, or turn off the lights when you leave home. You can even blink the lights if it starts snowing! But blink when you’re tagged on Facebook? Or mentioned on Twitter? You might have a social media addiction if…

IFTTT Integration

IFTTT Integration

What I really liked, though, and what turned out to be the biggest flaw in this whole setup, is the Nest integration that happens through a program called ‘Works with Nest.’ Lifx lights can be set to flash when a paired Nest smoke / CO2 alarm detects something amiss. As someone who barely survived a high-rise fire, I loved this feature, think it’s incredibly useful and potentially lifesaving, and as I have Nest smoke detectors in both locations where I use these Lifx bulbs, I couldn’t have been happier. HOWEVER: It appears you can only set up this connection at one location. Once you’ve done that, you can’t connect any others. So I connected the light to the detector in my office, but when I use the app to switch my location to home, there’s no option to connect the lights to the Nest detector there: Only to disconnect the other connection already made. I experimented and tested and connected and disconnected and reconnected and switched locations in the app and really tried to find a solution, but was unable, and I find this implementation to be absolutely baffling. So if this is a circumstance that applies to you, choose wisely: You can only connect these lights to a smoke detector at one location regardless of how many you have. This is a major oversight and desperately needs to be addressed.

Nest Integration

Nest Integration

Other than the fake visualizer, the single-light spooky thing, and the grossly limiting Nest integration, these are good lights and I’m happy with them. Even bought more after using the first one for a while. Bright, easily grouped and controlled, usable app, not overly expensive, and ultimately it’s the app and its features and functionality that needs some fixing, while the bulbs themselves are great.

Recommended.

IoT skills, deployment lagging behind expectations

Your data is here

Because I am teaching a project class in ubiquitous computing this quarter, I was struck by a post on the BPI (Business Performance Innovation) network discussing the results of a survey they conducted along with Nerdery and the Internet of Things Institute that states while industry does genuinely wish to adopt IoT strategies and deployments, they’re not happening as quickly as their enthusiasm might make it seem.

I can’t say I’m surprised by this. The most striking statistic as one reads through the report is that “…just 1.5 percent of executives at large companies say they have a clear vision with implementation well underway, while another 57 percent are either beginning implementation, have pilots underway or are committed and in the planning stages.” It’s the ‘clear vision’ aspect of this that is truly telling, especially when paired with the rest of the sentence. Immediately the question that presents itself is ‘If only 1.5 percent of executives have a clear IoT vision, how is it that 57 percent can be at various stages of design and/or implementation if they don’t?’

One clarification: In my class, and for the purposes of simplicity, we use Ubiquitous Computing and Internet of Things (IoT) interchangeably, however they’re not quite the same and I do let my students know that. The IoT is a subset of the concept of Ubiquitous Computing, the platform on which it’s enabled, similar to how the Internet is the technical foundation on which the Web operates. With that out of the way, the one issue I run into more than any other in this course is the belief of a small contingent of students that it should be a programming course, that they should be learning how to program sensors and the IoT protocols and whatnot. What I explain to them is that anyone can buy a book on how to program for the IoT; what they can’t buy is a book on how to think about the IoT. Like all technologies, IoT technologies can’t, as I tell them ad nauseum, be developed in a metaphorical vacuum. The technical issues of networking, the pragmatic issues of security and privacy, the enterprise issues of data collection and management, as well as many others, all must be considered as well when developing and deploying IoT strategies. After all, they’re not called strategies for nothing.

The second most common issue that comes up in this course is the question of what exactly the Internet of Things is. What defines it? What separates it from the regular Internet? Is that not an Internet of Things as well? How are they delineated? And where does ‘The Cloud’ fit in to all of this, if at all? Is that part of it? As is the case with all definitions, it can often depend on who you talk to. generally, in my class we define it terms of its low-power devices (sensors being the big ones) and lightweight protocols. However even in that case there can be disagreement and confusion. Do we need to make those delineations? The Internet is the Internet, yes? Do we need to divide it into TWO Internets, one for regular electronic devices and one for low-power sensors? What about our phones? Our TVs? They may be using Bluetooth but they certainly are not low-power, lightweight devices. And while this post isn’t meant to be about the technology per se, there are no fewer than twenty different protocols that can call themselves IoT protocols which has almost immediately led to standards overload. The image below gives and idea of this, although it contains both lightweight IoT as well as heavyweight regular networking protocols and where they fall in the TCP/IP stack.

Some IoT and non-IoT protocols

Some IoT and non-IoT protocols

And here’s this, for good measure, since we’re long past this moment in terms of the IoT. I rail against this all the time, but it never ceases to happen (and you should all read XKCD anyway):

Yep

Yep

True to this idea, The Technology Partnership in the UK states it should be called the Internet of Sensors, not the Internet of Things, since the Internet has always been an Internet of Things. A valid point, in my opinion, but then there’s also this. Let’s make up our minds, people!

And that brings me back to my introductory statistic. The linked article above has the following quote: “In my view, far too much attention is focused on getting the ‘Things” connected and not enough time is spent understanding the data insights that will actually drive the business forward.” That’s exactly right. Everyone is trying to figure out how they’ll get ‘things’ connected and get sensors out there and have an IoT strategy, but to what end? They’re so busy thinking about doing it that that they’re not thinking about why they’re doing it, what it all means, or if they need it, or how it will impact their business, or how it might impact their customers not just in terms of, say, improved customer service but also in terms of privacy and data collection and retention / distribution.

It’s important to take a step back and consider the gestalt of the IoT, and all the concerns and considerations that go along with it. It’s not a programming exercise; that’s just one low-hanging leaf on a vast tree of issues. I would hope that some people would be willing to take a step back for a moment, away from the headlong rush to have an ill-defined IoT strategy or deployment, and simply consider what it all means. We can all rush forward after that.