Broadlink products – no APIs and other surprises

If you surf on Chinese shopping web sites you might have seen devices from Broadlink like the A1. They are quite cheap and seem to be powerful. However, have a second look before buying.

First of all, some sellers claim that the A1 has a PM2.5 particle sensor “coming soon”. Looks like this isn’t implemented yet in this software – right? Wrong! The device doesn’t have the necessary hardware included. There might be an add-on module coming in the future, but it will be a hardware-add-on, not just a software upgrade.

a1Another thing that Broadlink doesn’t publicly communicate is the fact that there are no open interfaces available. It seems that there is an SDK for the device, but Broadlink only allows to use it under NDA and it is only available as library for IOS and Android. That makes it hard to impossible to use this hardware as a part of a larger integration.

Therefore I can’t recommend any of the Broadlink products as long as Broadlink doesn’t offer any documented APIs.

Blocking sensors in Home Assistant

I really like the open source software Home Assistant. However, today I noticed that the whole system wasn’t working anymore. Ok, sometimes a Reboot helps, but not this time. Looking further into this I found the cause of this problem. One of the WiFi enabled switches wasn’t responding anymore. And Home Assistant completely blocked.

This is a major issue in a home automation software as you can’t be sure that in a complex environment every component will work as expected. An issue has been opened for this problem. I hope that there is a easy way to fix this kind of problem.

If your Home Assistant isn’t working anymore, start it from command line and check the output. You might find a subsystem that blocks the startup.

Personal assistants

Two weeks ago, Google announced their Google Home product which is similar to Amazon’s Echo product. While both look like a voice-controlled speakers, in my opinion they are much more. Both are the first products of a new segment that I would call “personal assistants”. What do I mean with this?

A personal assistant is a system that

  • allows communication in natural language,
  • has a deep knowledge about the world,
  • knows your personal preferences,
  • that can interact with physical objects (like a light switch, your stereo equipment or your washing machine)

In my opinion the last point is quite important. To call something an assistant, it really has to DO something. It is also not a butler, that just does what you tell him. When you go to bed an say “turn off the light”, a butler would just do this. An assistant will do more. Imagine the following dialogue:

You: “Turn off the light”
Assistant: “I see that you have an early appointment tomorrow. Should I wake you at 6 instead of the normal 6:45?”
You: “Yes, please”
Assistant: “Have a good night” (turns lights off)

The digital assistant should also know that it’s not a good idea to talk to you in the sleeping room, when your husband or wife went to bed early and is already sleeping.

Google Home is just an interface to the assistant service. It will be able not only to manage your appointments, play back music but I would expect that it will be also an interface to your home. If your digital assistant knows the time of your first meeting at work, it can not only plan the wakeup call, but also power on your coffee machine. It will know when you leave your house and automatically lock your door. It will do a lot that I can’t even think of today.
However, no single firm (even not Google or Amazon) will be able to deliver a complete solution for all possible use cases. I do not expect that you will ever install Google light switches in your house. But your Google personal assistant should be able to control your light. This means that this is a classic platform business.

As with all platform businesses there isn’t rooms for a large number of platforms in the same area. Usually you see one to four main players with a number of niche players around. But who will be the platform leader in personal assistants?

Amazon

Amazon is in a good position. The Amazon echo is the only device that is already on the market. They also have the skills and infrastructure to build a service like this. They also know how to provide useful APIs to developers. With its cloud computing platform Amazon already powers a large number of consumer services – even if the consumer doesn’t know about it.

Apple

Apple have been highly successful building platforms. With the iPhone Apple created the first large platform since Microsoft Windows. However, Apple don’t like infrastructure business. Their platforms are build with one goal: to sell more Apple hardware. This has changed a bit with Apple Music, but I still don’t see Apple competing with Amazon, Google and Microsoft for the platform with the lowest price and the most features. While Apples HomeKit might be an interesting approach it is limited to Apple devices. Do you want to live in a smart home that can only be controlled with Apple devices?

Facebook

While Facebook is working on AI, it is mostly targeted to optimise the user experience within Facebook. Also experiences with companies bought by Facebook shows that Facebook isn’t interested to build an open platform that programmers connect to if it is not highly integrated into the Facebook application.

Google

Google is in a similar position as Amazon. They know how to operate large platforms and provide APIs to developers. They have access to the larges volumes of data available to train their system. However, one thing they consistently failed was to market hardware. The mobile phone business taken over from Motorola was sold again. The Chromebook is an interesting device, but did not get traction on the market. The best-known hardware from Google is the Nest thermostat – an acquisition. However, it seems that Google doesn’t really know what to do with Nest.
Google had to biggest success in hardware with the Android platform with a strategy that already worked for Microsoft: sell on Operating system and sell licenses to hardware vendors.
Maybe that is also a good strategy also for personal assistant platform.

Microsoft

Interestingly I did not see a lot in this area from Microsoft. I’m pretty sure that Microsoft is working on this, but I can’t see their strategy in this area yet. Their APIs and platforms are becoming more and more open, but Microsoft is still seen as the “Windows company” in the consumer market.

Others

There are a lot of startups that work in this area. However, I think most of them are just working on a cool technology and want to be bought by one of the large players. I don’t expect that one of these
Samsung is a classic follower. I’m pretty sure we will see some kind of voice-controlled device from Samsung also in the next 12-24 months. However, Samsung has never been successful building a platform business and I don’t see that they are doing anything to change this.

Looking at the strengths and weaknesses of different players in this area, I would expect Google to be successful when they focus on building a platform and APIs instead of selling hardware. However, Amazon is an important competitor and it will be interesting to see how they and the other competitors will position themselves in this area.

Security in IOT

If you build your own home automation, security is something you should think of from the beginning. You don’t want somebody turning of your heating while you are away for skiing for a week. It is also not too much fun if somebody turns on the lights in your sleeping room during the night.
Joshua Corman explains the security risks in IOT applications (home automation is kind of an IOT application) and what to do to build secure systems.

Click on the image below to view his keynote from the 2016 IoT conference
Screenshot 2016-05-17 09.31.54

Z-Wave goodbye? Not yet

If you are interested in security, you might already know Steve Gibson’s podcast “Security now”. You might not agree with all of his opinions, but he collects quite a lot of information what’s happening in computer security. Make sure you understand what part of the podcast is advertisement and what is real information – as it is not always obvious.

At the latest episode, Steve has a look at an attack to Z-Wave that had been shown recently. As the podcast is always long, you should skip to 1:41 (minute 101).

goodbye-zwave

Steve has some valid points against Z-Wave. Especially the fact that most of the standard isn’t publicly available means that there can be flaws in the design or the system that can’t be easily corrected. However, if you look at the presentation the podcast if referring to, you will see that the guys did not break the Z-Wave encryption. It just wasn’t available in the network they hacked into as it isn’t widely used today.

Does this mean Z-Wave is secure? Not really, but it also doesn’t mean that Z-Wave is insecure in general. We just don’t know yet how secure it is. Would I use it for a door lock? Most likely not. But it might be still a relatively inexpensive choice for some non-critical functionalities.

References

Voice control for your home?

I was never a big fan of voice dictation or voice control. When I started to experiment with this technology in the 1990s it was far from perfect. Even today, I use Siri on my iPhone very rarely. However I see some interesting use cases in home automation. What about going to bed and say “turn the lights off”? This should be quite simple. You can already buy products that do things like this. Check out this cool KNX room controller: imgres

It not only looks good, it is also expensive. But is this really the future? It is a start. But there are much better use cases. Both Amazon and Google are working in real assistant services. The Amazon Echo is already on the market in the US, Googles Home will be available soon. I would expect that Apple is working on something similar.

One use case that I can easily think of is the following:

me: “Turn the lights off”
control: “Have a good night, I will wake you tomorrow at 7”

That’s still quite simple, but what about:

me: “Turn the lights off”
control: “I see that you have an early meeting tomorrow. Should I wake you already at 6:30?”

Or this:

control: “It is expected to snow tonight. This might cause delays in public transportation tomorrow. Should I wake you a bit earlier than usual?”

There are many other interesting use cases. Will it be reality soon? I don’t know but I really want to experiment with this.

Innovation begins at home

Andy Stanford-Clark, IBM Distinguished Engineer for the Internet of Things, will explain how his hobby of home automation and his passion for energy saving converges with the Internet of Things.
This has led to projects that have helped alleviate energy poverty in social housing, and are helping commuters from the Isle of Wight check if the ferries are running on time.
Andy will explain that “it’s all about the data” and how these solutions are built from instrumented, interconnected, intelligent devices. He will illustrate his talk with examples from his home, including his twittering mouse traps!

Click on the image to view the video on Vimeo.

Screenshot 2016-05-17 09.22.36

Is wireless the future?

Looking at new companies in the home automation industry it looks like many of them prefer some kind of radio frequency data transmission. This makes sense as customers don’t want to have additional cables in their home. But is this really the future of home automation? I’m not sure. Here are some arguments against wireless control:

  1. Reliability: with more and more devices communicating wireless, interoperability problems can become more and more problematic.
  2. Power supply: with some exceptions (e.g. EnOcean), even wireless devices need a power supply. This mean either cabling or batteries that need to be changed regularly
  3. Security: Wireless devices are easier to attack than wired devices (at least if these are not connected to a public network).

Especially the security aspect is important in my point of view. Many home automation devices will be used for 10-50 years. While it might be reasonable to buy a new music player every 5 years, you don’t want to change your wall switches every 10 years. The development of encryption algorithms has shown that no algorithm is secure forever. This means the software has to be updated from time to time. Newer encryption algorithms might even need more powerful hardware. Do you believe your supplier will provide software updates for the devices during the next 20 years?