Whisker.IO Blog - "Proof of scalability: the key to success in the IoT space"

Save to PDF

Posted on Friday, January 8, 2016 by Steve Montgomery

Although the Internet of Things (IoT) is projected to grow to more than 25 billion by 2020, there are some key challenges that must be addressed.  As businesses explore the potential of IoT, many are finding compelling applications for the technology.  However, many are also finding that proving a concept is much easier than finding a solution that can scale.  In this blog, Steve outlines these challenges and talks about how Whisker.IO is the perfect solution for proving both concept and scalability of almost any IoT application... 

Market projections from researchers and industry giants predicts a huge growth in the IoT space by 2020: anywhere from 35 to 100 billion connected “things”.  A large part of that is going to consist of wireless devices that must operate from battery power, and we as an industry are going to have to solve some key problems to make those projections a reality.

Today, many companies are actively trying to create an IoT strategy for their businesses and customers, but they face some pretty serious issues.  

There is no clear definition of what the Internet of Things actually is, so companies must figure that out for themselves.  They are faced with the challenge of making a solid business case for the use of IoT technologies, but a lack of understanding of what IoT solutions are available, the limitations of existing wireless technologies that might be used, and the status quo are serious road blocks.

Usually, proving a business case involves building a proof of concept (POC) prototype that can be tested and evaluated over time.  For most companies trying to establish their IoT strategy, this is where they are at today.

With the uninhibited growth in Internet and cloud technologies, rapidly prototyping web applications is a well understood, and straightforward process.   In a short time, we could integrate a number of existing cloud technologies and some custom code to build a complete web based application for smart trashcans, for example.  They would let us know when they are full, remind us to put them out when the trash is coming, and remind us to pull them in after they are emptied using a combination of text messages, email, and even phone calls. The city would maintain a complete inventory of trashcans and might even be able to move from a schedule driven route for pick-up to a data driven route that only has the truck emptying cans that are actually full and at the curb.

The challenge comes when we actually need to connect the thing, or the trashcan in this example, to the cloud so we can collect this data.  Clearly, we cannot run wires to the trashcan, so whatever device we use must be wireless and battery powered.  

WiFi and cellular connectivity are both options that would be traditionally considered for an application like this, but as many companies are discovering, there are significant issues with either of these technologies for this application.

First, the battery life is measured in weeks or months.  Imagine the burden this would put on even a small city as they had to assign personnel to regular maintenance of the batteries in all of the tracked trashcans in the city.  It is untenable and would create too much process friction to allow mass adoption.

Second, every cellular connection comes with its own monthly fee, increasing the overall operating cost of the solution.  The cellular network is not yet ready to provide cost effective access for large numbers of low data rate devices.  

WiFi has a big problem with connectivity.  It must be in range of an access point AND have permission to communicate with that access point.  There is no way to guarantee either of those things.  WiFi range is measured in hundreds of feet, which means that the city would have to install a significant amount of infrastructure to ensure connectivity.  The monthly operating cost would be less than with cellular, but the overall up front cost and on-going maintenance cost of the infrastructure would likely be more expensive.

Third, WiFI has issues with regard to provisioning.  The process of provisioning an embedded WiFi device is highly technical, prone to errors, and must be done per device connected each and every time the access point is replaced, passwords are changed, or anything goes wrong with the connection.  As with the batteries, the provisioning could represent a significant amount of on-going maintenance that will create process friction, thereby killing the chance of mass adoption.  It is one thing to connect one thermostat in one’s home, it is another to do that process thousands of times for connected trashcans in a city.

Today, POC projects are being pursued by the tens of thousands using technologies like cellular and WiFi, but there is no clear path to scalability.  The resulting solution would be too expensive to install and too expensive and frustrating to maintain.  So most projects stall once the concept is proven and this will ultimately be a stumbling block to the projected market growth I talked about at the start of the article.

What is needed is not merely proof of concept, but proof of scalability.  Yes, we need to prove that an idea can work, but it has to scale with the application to be acceptable.

So what is needed for proof of scalability?  Good question.  Here are the key factors we have identified:

  1. It must scale downward as well as upward.
  2. It must offer very long range
  3. It must offer very long battery life
  4. It must be very easy to install and to maintain

Scaling downwards as well as upwards is an important capability.  As companies stick their big toe in the “IoT” water, they are going to do it carefully and with deliberation.  Small tests are run with small groups of devices.  If successful and fully adopted, the solution could move to larger scale, but it has to be cost effective and it must have a clear path to large scale deployment to be accepted.  Otherwise, the application will never get off the ground.

Long range and battery life have already been discussed for the most part.  A scalable wireless IoT solution must be able to reach the “things” where they live without the need for expensive infrastructure and must operate for years without maintenance.

Installation must be simple.  If an IT expert with a PhD is required to install a solution, it will probably fail.  If large amounts of infrastructure are required, it will be too expensive and will also likely fail.  And if continuous maintenance is required to keep the application running, that will also likely fail.

New technologies like the LoRa radio chips from Semtech offer solution developers the tools to address these problems; LoRa, for example, is a very long range wireless technology with the potential for long battery life if properly designed into a solution.

Combining these new technologies with proper system design will allow IoT solution providers the ability to build products and applications that can survive “proof of scalability" and that in turn will enable the IoT space to grow to the projected 35 billion devices by 2020.

About Steve Montgomery and Digital Six Laboratories

Steve Montgomery is the founder and CEO of Digital Six Laboratories,LLC. He has more than 20 years of experience with developing embedded wireless technologies for a broad range of consumer, commercial, and industrial applications.  For the past 10 years, he has been focused on M2M and now IoT applications.

Digital Six Laboratories LLC is a cloud-to-device wireless connectivity company specializing in helping enterprises connect their "things” to the Internet.  They have developed a long range, battery powered wireless solution called Whisker.IO that is capable of 4 miles of range and 5 years of battery life from 2xAA batteries.  Whisker.IO is a fire-and-forget technology that requires nearly zero infrastructure (one gateway) and zero on-going maintenance.