The Pendulum of Innovation: Sensor Networks

Soofa’s VP of Engineering, Bob Breznak, recently went to Sensor Expo 2015. While there are a ton of sensors out there, they are mostly the same. This lead us to discuss the never-ceasing pendulum of innovation. One one side of the pendulum, we have the device. On the other, side the network. As new technology emerges, we are continually swinging from emphasis on device to network, to the creation of a new device built upon the former network and the development of a new network. And on, and on, endlessly. This is a really exciting and cool cycle that has resulted in the development of the modern world as we know it. The more devices and networks created, the more that can be created! They build upon each other, similar to the concept of a flywheel popularized by Jim Collins (and loved by Jeff Bezos).

For a while, the sensor market had an emphasis on producing the best devices and driving costs down. Now, there are affordable sensors for anything you could dream up, and numerous developer kits and companies dedicated to helping you use sensors! The fight for dominant status in the sensor market has mostly concluded, with a few major companies emerging as victors.

We know this is inevitable of any new technology - many enter the market, tackling the low-hanging fruit, before a few companies more or less arbitrarily become dominant. When the products all do the same thing, company success really depend on adoption and marketing. Now, people are trying to solve a new problem: "How can we make the creation of sensor networks, and the retrieval of data, simple?"

That's how it always goes: there's a race to create the new technology, then a push to simplify that technology and make it accessible to everyone. Sensor network technologies are the new low hanging fruit, and the pendulum has swung back towards the network side. There are countless companies dedicated to making installing sensors and interpreting their output easier. Take Samsara, which just raised $25 million:

“Breakthroughs in semiconductors are making sensors smaller and radically cheaper, and wireless Internet connectivity is becoming ubiquitous. Yet traditional enterprise-scale sensor systems still require complex infrastructure, from hard-wired networks to servers, middleware, and custom code... Samsara disrupts the traditional sensor model with an integrated, software-centric solution. Our products combine plug-and-play sensors, wireless connectivity, and rich cloud-hosted software, all tightly-integrated for simple deployment.” From Samsara website 

Sound familiar? They join Helium, Wovyn, Dust Networks (Linear Technology), and so many other companies in their quest to simplify sensor networks.

A diagram from the Helium website provides a good visualization for how these companies can simplify the process of "sensorizing" something.

From  Helium

From Helium

The funny part about this cycle is that the new creation is based on the work of companies that it will soon push out of the market they created. Take server spaces - servers were built from scratch by different tech companies for years. Companies building massive data farms were able to be the top performers of the time, and naturally the resources to do so were mostly limited to large companies. Then, Amazon Web Services (AWS) came around. Today AWS is the dominant server provider, but their service could not have been built without the work of companies before. So, as software progresses & technology gets cheaper, barriers to entry are lowered. Companies can then come along and be genuinely competitive with incumbents who built the very infrastructure they stand upon.

Sensors have been around a while, but it wasn’t until smartphones required them and enabled further sensor applications that the sensor space exploded. A few years ago, creating a new low-cost sensor and commoditizing it was a lucrative and simple path. Now, the fight is about commoditizing the network infrastructure to make these devices work cohesively. That was the main thing Bob saw at Sensor Expo - there are a ton of companies trying to be the company in sensor networking!

We are left, then, with some open-ended questions. There are so many different transmission methods of information from sensors - Bluetooth Low Energy (BLE), RFID, LoRa, Wifi, and traditional connectivity via internet connections (cellular, Wifi, internal networks). It’s hard to establish a singular sensor networking platform when there are so many uses for sensors and ways they share their data! So, will a singular networking paradigm eventually evolve? Right now, this seems unlikely - after all, different transmission methods exist to suit different needs. In the future, though, there may be general dominance of a few categories of transmission, along with the creation of new methods to suit needs that haven't arisen yet. 

Also, we are in the early stages of widespread sensor usage. How could a company possibly create a winning network paradigm when we don’t really know what our connected networks will require? The answer, I think, is that they can’t. That’s not to say that no companies in the sensor network space will succeed - they definitely will; there is a clear need for simplified networking. It’s just most likely that the dominant sensor networking paradigm - if there is one - will be created by a company building off of these companies' (Samsara, Helium, Wovyn, etc.) early work. After all, isn’t that how the pendulum swings?