Tag Archives: Internet of Things

The IoT — a bottom-up approach

In the early 1980s, the Apple II computer combined with the first spreadsheet program (called Visicalc) revolutionized how financial analysis was done. Suddenly through the use of these new tools, complex models could be created by anyone that could type numbers and formulas into a grid. Almost overnight, financial decision making was transformed through the ability to use more data to accurately predict business outcomes versus simply guessing.

But this transformation did not happen from a corporate edict. Back in those days, Apple II computers were brought in the back door of businesses by employees that understood the power of the tool and the value it could provide.   It was not part of the IT infrastructure and was not part of any corporate computing initiative. Businesses did not offer to buy Apple II computers and did not offer training or support and, frequently, if an employee was “caught” using unauthorized equipment, they would be in hot water.

Yet this grass roots movement changed the face of technology and evolved into computing as it is known today. And the companies that lagged in the new technology deployment were left in the dust by their competition.

What does this have to do with the Internet of Things (IoT)? Plenty. We believe that history is about to repeat itself.

As designers and developers of IoT systems, from our vantage point virtually all industries can benefit from the IoT. If you are involved in providing a product or service, you can always gain from having a deeper understanding of how satisfied your customer is, how they use your product or service, and how you can make your product or service even more compelling to that customer. The IoT allows for the collection of near real time data about your product or service — and this data obtained can revolutionize your business and can leapfrog you ahead of your competition.

So why aren’t more companies deploying IoT solutions today? Yes the technology is new.   But we believe that many people are waiting for someone to tell them how to “do” the IoT.

As a case in point — some Internet of Things providers have taken the approach of doing things from “the top down” and recommend that their clients study the IoT from a corporate perspective. They talk their client companies into developing a corporate vision of how IoT technology can be used across the entire organization, followed by the development of a 3-5 year plan to successfully implement the strategy successfully. They engage with their clients initially with an expensive study that produces a vision and plan before any real IoT benefits are deployed and realized.

We believe there are several problems with this approach. First, IoT technology is changing so rapidly that any type of long term plan defined today will most certainly need to be redone as things evolve. Technology changes, costs are plummeting, and new solutions will be developed that are not even on the horizon yet. Second, it has been our experience that it is very difficult to envision all of the ways that a single IoT deployment can benefit an entire organization. Many times a project is developed with a specific goal in mind, but once the deployment is completed, there are typically many other benefits that are realized once the data obtained is fully digested. And third (and probably most importantly) this approach delays the benefits to the organization by deferring any IoT deployments until the entire vision is defined and agreed to.

Our preference is to promote a “bottom-up” approach to IoT.   In a similar vein to Agile software development, we believe that a strategy in which projects are formed based on today’s real, identifiable needs within the organization and developed/deployed iteratively while realizing new value from the project as new information is learned, is a more successful way to realize the benefits of the IoT . Each project stands on its own merits and has specific and tangible goals that can be achieved in short deployment cycles that build on the previous cycles.

The benefits to this approach are many — smaller efforts with specific ROI can allow the organization to gain the benefit of IoT technology quickly to improve their products and services and reduce their operating costs. Using an iterative deployment approach, organizations can learn about the value of data acquired and rapidly build on the success of their smaller projects as they apply this learning on new deployments.

So the choice is yours. You can wait for someone to study your situation and deliver a high level analysis of your corporate IoT situation that will be obsolete in 6 -12 months or you can start deploying IoT technology today on a smaller scale, delivering near term results that can be expanded as you learn and as the technology evolves.

Truly innovative organizations find the right tools and put these tools to work to gain the advantage on the competition, learning as they go.   Does this describe your business? Or does it describe your competitor?

The Internet of Things is not a smartphone app

apple-iphone-technology-blur

Recently, I’ve noticed an interesting trend in product introductions in the Internet of Things (IoT) space.   There seems to be a lot of manufacturers that want you to believe that wirelessly controlling a device by a smartphone app is “the Internet of Things”. Well of course it depends on your definition of the IoT, but simple control of a device via Bluetooth from a smartphone app doesn’t really qualify as an IoT solution to me.

In my opinion, to qualify as an IoT “product”, the product needs to:

  • Be a device that is collecting data and optionally controlling a remote device
  • Have the ability to communicate the data collected to the cloud
  • Have the ability to perform data analytics on the aggregation of the data collected to derive new value

To illustrate the difference between simple local control and a true IoT application, consider the example of a device designed to turn a light bulb on or off from your smartphone. Aside from it being pretty cumbersome to do (start the app on your phone, select the right screen, scroll to find your particular light bulb, then tap to turn it on — I think I’ll just go flip the switch on the wall), simple controllers do not make use of other devices that might work in concert with the light bulb controller — say a motion detector. A more sophisticated light bulb controller that worked in conjunction with a motion detector by passing messages over the local area network could automatically trigger the light bulb controller to turn the light on when someone entered the room. No clunky smartphone interaction is required. By having the ability to have separate devices work together, the process is simplified and much more useful.

Next — consider the integration of a cloud-based server that can communicate with the light bulb controller. By sending the device’s status to a server, I can track when the device is on or off from the cloud. Now imagine that I do this for every light bulb (or any other appliance) in my home.   I now have data available that shows the state of all of my devices in real time and data that shows how each device is utilized over any time interval that I choose. I also gain the ability to control my device from the cloud — I no longer need to be within Bluetooth proximity to control my lights — I can do it from virtually anywhere that I have an Internet connection. This gives me additional capability to create cloud-based software to set up a schedule when I want specific lights to go on or off in my home automatically. Again – no smartphone app required.

Lastly, after accumulating light bulb data over time, I can now look at it as an aggregation and derive important and valuable information from it. For example, I can track my overall energy usage over a day, month, or year. I can tell if a bulb is burned out (i.e. if it is not consuming electrical current) and needs replacement.   I can combine my lighting control data with data from other systems to provide a comprehensive picture of my home or office energy usage / savings.

If I am the light bulb controller’s manufacturer, I can use the data accumulated in the cloud to determine how my customer is using my product (how often, how many times, at what times, etc.). I can also derive quality information about all of my devices in the field — time between failures, most common failure, etc. The data allows me to understand my market and use this knowledge to create better products and more compelling solutions to stay ahead of the competition.

In our small example of the light bulb controller, you can see the value of not only controlling your devices locally, but the additional value of combining the function of the controller with other devices, the value of interacting with your device via the cloud and using data analytics on that data to derive new value. Thinking beyond simple light bulb controllers, imagine applying this same mind set to IoT solutions for factory floors, agricultural applications, transportation, oil/gas rigs, medical, home automation, and any other industry that involves process control.

Now that is the Internet of Things.

Real world wireless communication for the Internet of Things

radio-mast-561650_1920

When implementing an “Internet of Things” (IoT) solution, an important consideration is the wireless technology(ies) that will be employed to get data from your devices (sensors, gateways, etc.) to the cloud. As with most technologies, there is a confusing array of options — each with its strengths and weaknesses depending on the specific application.

Frequently the discussion around wireless technologies centers around specs, throughput, standards, etc., but what is really needed is a practical guide to implementation for specific applications.   Based on our experience in the field deployments we have made, we would like to share some of our thoughts regarding the current state of wireless technology for IoT.

Before the Tech Talk

Before getting into the bits and bytes of any particular communication standard, consider the following factors and decide on the characteristics of your device and its deployment:

  • Will my device be stationary or mobile?
  • Is there a power source that can be utilized for the device or will my device rely on its own power?
  • What is the range necessary to transmit / receive data to and from the device?
  • How often will my device need to communicate to the cloud?
  • What is the volume of data that will be sent?

These factors will be important as you decide on a communication technology for your system. As we will discuss, each form of wireless communication has specific attributes that may either make it a preferred choice or that may disqualify it from your selection process.

Some common wireless communication technologies to consider

We’ve captured a few of the most popular wireless communication technologies used in IoT deployments below along with some useful information to help during the selection process:

  • BLE — which stands for Bluetooth Low Energy. Also known as Bluetooth 4.0, BLE is useful for short haul communication (usually less than 100m) between devices. It has a moderately high data throughput rate and is deployed frequently in battery operated applications due to its low power requirements. It is really a “point to point” communication scheme between devices without networking capability.

Some other key points regarding BLE:

  • Every modern smartphone has BLE compatible radios and software stacks, making communication between your device and a smartphone relatively easy to implement.
  • Bluetooth technology is evolving quickly and specifications change accordingly, making backwards compatibility a potential issue for your device.
  • BLE can be implemented in an embedded design for < $5 in quantity
  • Not really well suited for direct communication to the cloud as it does not support HTTP for REST interfaces. It typically requires a gateway that can transfer data on the BLE side to a HTTP communication path (typically Ethernet or WiFi).       We have used smartphones as gateways for this purpose.
  • Zigbee —  is a radio technology used to create personal area networks and is very useful in certain IoT deployments.   Zigbee is also a short haul communication scheme, but has the additional benefit of being able to operate in a “mesh” network in which all Zigbee nodes can pass messages between other nodes, effectively increasing the distance that can be covered. This is useful in multi-point deployments where you may have multiple sensors distributed in a concentrated area (i.e. manufacturing floors, home automation, etc.).

Some other considerations for Zigbee:

  • Typically Zigbee is very power friendly, with low stand-by current draw and relatively low power consumption during transmission.       This makes it suitable for battery powered devices.
  • Zigbee has found a following over the past few years, particularly with home automation OEMs.
  • Data throughout is modest (up to 250k bits/sec)
  • Secured via 128-bit encryption keys.
  • WiFi — started as a wireless extension of Ethernet for computer networking. It was widely adapted in the consumer and industrial spaces and has become the standard for networking PCs, tablets, notebooks, and associated peripherals. It has extremely high data throughput — however this comes at the expense of its power requirements. Range typically is up to 100m, however this can be extended with specific antenna types.

Some other key points regarding WiFi:

  • Ubiquitous
  • Universal smartphone compatibility
  • Affordable — can be implemented in an embedded design for < $5 and frequently the connection to the Internet can be obtained for free
  • Well suited to communicate through the Internet to a cloud server — universally supports TCP/IP and HTTP to provide REST interfaces to cloud services
  • Well protected and controlled from a security standpoint
  • Relatively high power usage
  • Sometimes difficult to connect IoT devices to corporate IT infrastructure because of corporate policies on security.
  • Cellular — utilizing a network established initially for voice communication, cellular offers a wide bandwidth data capability that does not depend on local access points or routers. Its range therefore is virtually limitless (as long as you are in an area covered by cellular coverage).   However, you will pay for each byte transmitted / received with this technology and of all of the technologies discussed, it has the highest requirement for power consumption during transmission.

Some additional points regarding cellular:

  • Cellular benefits from security models already established for regular voice communication
  • This is the only technology discussed that charges per byte of data transmitted or received. While these rates have come down recently (and will continue to do so), it may be a factor in your deployment
  • Network coverage can be an issue is specific deployments — especially indoors on factory floors, etc.
  • Power consumption is high, especially during data transmission.       However, if your device has a power source readily available, this may not be an issue
  • Device cost is the highest in terms of the devices discussed here — typically $20 in volume. This price will come down as higher volumes of these devices are manufactured for IoT applications.
  • Other
    • A ton of other “standards” in the Low Power Wide Area Network (LPWAN) arena are beginning to surface — including SigFox, Dash 7 Alliance Protocol, LoRaWAN, nWave, Weightless-P and Weightless-N, IEEE 802.11ah, and LTE Cat-M. These standards are emerging as a direct result of needs in machine-to-machine communication.
    • Some of these new standards have started to gain traction, particularly in specific industry segments.
    • Estimates show that today the cost of a typical module ranges from about $5 to $20, depending on the specific technology. However, the cost per module will eventually fall below one to two dollars in volume quantities.

So in summary, selecting a wireless communication technology for your IoT deployment requires careful consideration of a number of factors.   It is critical to understand where your device deployment will happen, what your data requirements are, and how the devices will operate after deployment.   Power is always a consideration, as is the overall cost of the device itself and the costs associated with on-going support.

We’ve seen IoT wireless communication projects suffer due to the mismatch of actual needs versus the technology chosen — spending the appropriate amount of time and energy on this decision will avoid expensive redesign / rework after the fact.

The Fragmented Internet of Things

gears

The allure of the “Internet of Things” (IoT) to today’s business leaders is compelling. Having the information to run an organization more efficiently with the realtime data that the IoT provides can result in lower operating expenses, increased profitability, and higher levels of quality. OEMs that adapt the IoT in their product strategies can offer better functionality for their end users while also building an ability to acquire deep knowledge about how their products are used in the marketplace. So it’s no wonder that companies are rushing forward to find help in implementing their IoT initiatives.

However, the help these organizations employ to achieve their IoT goals often is long on promise and short on delivery. Many companies claim to be experts on the IoT, but they frequently only provide portions of the overall solution path. For example, sensor manufacturers may have the hardware element necessary to acquire important data, but frequently do not have the ability to move this data to the cloud and make it useful. Similarly, cloud providers are happy to provide interfaces for devices, but often do not provide end-to-end support for the variety of sensors, gateways, and other hardware devices sending that data.   Wireless carriers are good at moving data around, but they often do not have relationships with the hardware manufacturers producing the data or the cloud providers hosting the data. And mobile app developers can create applications to allow data to be displayed and/or manipulated, but frequently have no idea how the data is acquired and what to do if the flow of data stops.

Most of the time, the customer is led to believe that they are getting a complete solution with a single vendor, only to find out that their vendor only supplies a piece of the IoT puzzle, with the other pieces being left open for definition and implementation. The result is a confusing journey with other vendor’s products and services, trying to create a complete, end-to-end solution, often resulting in a partially working system and frequently ending in the failure of the project.

Clearly, the customer needs a complete, holistic solution. And here is how it should be done:

  • Step 1 : Define the business problem / opportunity

Implementing an IoT solution should be no different than any other new business initiative. First start with a definition of the problem to be solved or the opportunity that can be created by using the IoT. Then determine the costs, risks, ROI, and other factors that should guide the project.   Technology should not be the focus of this study — the organization should see a clear benefit at a business level at this stage.

  • Step 2: Define the total IoT solution that solves the problem / creates opportunity

Next, a discussion should take place between the organization and an IoT firm that has a proven track record in providing end-to-end IoT solutions. This firm should strive to understand the business need and then help create a solution that will meet the project requirements.   When considering the technology required for the project, the firm should have complete and proven solutions for:

  • Data acquisition — sensors, monitors, gateways, wireless communication
  • Device control — configuration and real-time control of remote devices
  • The Cloud — data protocols, storage, customer Internet portals, data analytics
  • Mobile apps — data display, device setup/configuration, alerts/alarms
  • Step 3: Implement the solution and provide on-going support

Lastly, the IoT firm employed should provide an easy to understand proposal for implementation of the entire IoT solution being considered. The proposal should cover “build” versus “buy” of the various hardware devices, services, and software needed for the system solution. It should also include not only the initial installation and operation of the system, but also a plan for on-going support for maintenance and scalability as the organization grows.

By considering the IoT project as an entity and not as a piece of hardware or software that is sold by a particular vendor, a customer can dramatically increase the likelihood that their IoT project will not only be able to succeed in the project’s stated goals, but that it will be done without the confusion and uncertainty that occurs with many current IoT projects that are trapped in the Fragmented Internet of Things.