A man seated at a table on which multiple electronic devices are placed.

Standardization in IoT: The Case for a Common Language of Intelligent Things

I first published this Standardization in IoT article on LinkedIn on 6 July 2017. I am republishing it here after editing for clarity and conciseness.

Are you disciplined enough to buy products from the same tech company so that your mobile device can reliably communicate with your TV, watch, laptop, tablet, refrigerator, home entertainment consoles and perhaps even your car?

Or do you wish that you didn’t have to check whether your phone can communicate with the smart TV you received as a gift. Or whether your smart watch can control your speakers?

An increasing number of consumer goods are connecting to the internet. Are walled gardens the future of tech? Or can we expect our intelligent devices to communicate with each other seamlessly, regardless of manufacturer?

This article examines, from a customer’s perspective, the story of the Internet of Things and the strategic case for standardization. Sources not hyperlinked are listed in the references section.

What is IoT?

Gartner listed the combination of Artificial Intelligence (AI) and the Internet of Things (IoT) as one of the most promising Strategic Technology Trends for 2017. The study found that independently functioning smart devices will soon collaborate. This means that users will find themselves surrounded by intelligent things that work in tandem.

Gartner (n.d.-a) defines Artificial Intelligence as “technology that appears to emulate human performance” either in aiding work on complex tasks or replacing the need for human input in “nonroutine tasks”. AI becomes more useful as the training data increases. Indeed, in 2013, it was estimated that 90% of all data generated in history had been done so in the previous two years. AI is being used to parse that data, much of which is unstructured (O’Leary, 2013).

The Internet of Things (IoT) refers to a network of objects that communicate information, about their environments, to each other. These devices do not need physical proximity. For example, vehicles can communicate with home appliances via the internet. A person driving home might send a signal to their home management software to activate the heating system. Thus, IoT is a technological implementation that aims to improve “life quality” via the sharing and analysis of data between different objects (Bujari et al., 2017).

Opportunities and Challenges of IoT

IoT is not limited to serving a single household at a time. It can scale to larger environments such as hospitals, traffic management, industrial operations, and public services (Bujari et al., 2017). Connecting any and all devices is a step up from the past few decades that focused on connecting computers in the ‘dot com’ era and people in the ‘social media’ era (Taivalsaari & Mikkonen, 2017). Taivalsaari and Mikkonen (2017) go on to say that the ability to network nearly every device is due to increasingly affordable computing hardware. As a result, many organizations, are capitalizing on the ease of entering the market. Bujari et al. (2017) estimated 20 billion objects to connect to the internet by 2020. This is orders of magnitude higher than the scale of millions at the time the paper was written. They also reported that Gartner estimated the revenues from IoT technology to reach $300 billion in 2020. In response to the burgeoning growth of this technology, European governments have earmarked millions of dollars to research and projects associated with such ventures. Competition has resulted in diverse offerings of IoT platforms. There were 115 such platforms in 2017. The technology was fairly young then, but in all likelihood, the major players will become apparent by 2025 (Taivalsaari & Mikkonen, 2017).

The presence of so many platforms presents a challenge for many stakeholders. First, each manufacturer or solution provider is tailoring their IoT platform to match their suite of products. As more kinds of products connect, a user may come to possess products from various manufacturers with different ways of collecting data and connecting to the internet. In such a case, the user may need different ‘controllers’ to control the functions of different devices (Cerf, 2015). If a common platform or standard were available, the user could use one device to manage the others around it.

Having each device manage its own ‘stack’ could lead to greater demands on computational power (Sigtia, Stark, Krstulovic, & Plumbley, 2016). Data collection at hardware level, data analysis in embedded software or in the cloud, and communication with the cloud all contribute to demands on computational power. This could increase costs of building the device. Letting nearby controllers, such as smartphones, analyze data from these devices could reduce computational power requirements and free up resources for other tasks.

A common IoT platform could also better manage the security and privacy of a user’s data. Each device having a different data security protocols could lead to security breaches or privacy violations. A common standard would enable manufacturers to create safeguards for data at appropriate points in the ‘stack’.

Standardization in IoT is the need of the hour to make these products easy to use and to expand choice. Many manufacturers have teamed up to work on a common platform.

The Status Quo

Implementation of IoT

The market potential of IoT is extensive precisely because of the variety of applications made possible through this technology. Home appliances that communicate with each other can make the home smarter. Similarly, medical objects, such as prescription bottles and vital sign sensors, working in sync can autonomously monitor patients (Bujari et al., 2017). To add value to their product, device manufacturers are now implementing IoT solutions to their product features (Banafa, 2016). Banafa says that these manufacturers are designing and implementing the entire ‘stack’, i.e., the hardware that enables connectivity and the cloud based network that collects data and provides the ‘online’ features of these products. The ‘stack’ includes the device itself, which contains the hardware that collects data from the environment through sensors or other aids (Taivalsaari & Mikkonen, 2017). This hardware then relays data to the cloud via gateways that act as channels. The cloud is where the data is gathered and analyzed in depth. The intensive computing required to gather, compare and analyze data from different sources is made efficient by the use of cloud computing (Zhou, Cao, Dong & Vasilakos, 2017).

Cloud computing is not always reliable for applications that are time-sensitive such as “telemedicine and patient care” (Dickson, 2016). Dickson found fog computing to be an emergent solution. Fog computing delegates computations at the ‘edges’, or in other words, at IoT gateways like routers. IoT devices could send data to such gateways in their vicinity to analyze their data and receive results back sooner than if the data had been sent to a cloud server. Cloud and fog computing have been suggested as solutions because it is not feasible to put large amounts of computing resources in every IoT device. Thus, having devices from various manufacturers be able to talk to a common gateway would be made easier by the implementation of a common standard.

Banafa (2016) found that because each manufacturer tends to develop their own ‘stack’, an inconsistency can be observed across the different cloud services used by them. He further states that a “standard model” is needed to perform common tasks inherent to the process such as processing and storing data. Such a model will enable different devices from different manufacturers to work together seamlessly. The user-facing side of the ‘solution’, usually the ‘app’ that controls the device or shows information related to it is generally tied to that particular device. With the rising ubiquity of IoT devices, a user may be inundated with a plethora of apps to control all the different devices in their possession. This is not sustainable from the customer’s point of view (Taivalsaari & Mikkonen, 2017). Instead, a unified model will let a user control all other connected devices through one device. This will fully realize the quality of life goal of the technology. Ultimately, for such a network of devices to function satisfactorily, they must be stable (Naraynan, 2017).

The Case for Standardization

Ease of use is not the only thing that necessitates a common platform or language. Connecting to the internet opens up a device to malicious attacks thereby negatively impacting the user’s security and privacy (Bujari et al., 2017). The authors further said that solutions that guard against such attacks cannot be localized to the application or device but must encompass all stages from the device to the cloud and vice-versa. This is further complicated in a network of different devices with different functions and that therefore require different security levels. For example, from a security point of view, a wearable device that measures heart-rate may be considered to be less important than one that measures blood sugar level. A breach of this data may also lead to privacy concerns if the data is shared with insurance companies. This could have a material impact on the user’s financial position. According to Bujari et al. (2017), Lin & Bergmann (2016), and Tiburski et al. (2016), when handling user data an IoT solution must consider:

  • Authentication: this refers to the identifiability of an object,
  • Confidentiality: this secures data and makes it available only to “authorized” sources,
  • Integrity: this refers to data being unaltered in transit or in storage,
  • Fault tolerance: this refers to the ongoing provision of security services if any problem is identified

It should be assumed that any IoT device can be targeted for data theft (Lin & Bergmann, 2016).

Competitive Collaboration?

A standardized platform for IoT must, therefore, address these issues of efficiency, ease of use, security and privacy among others. The constraints at the device level are low computation capabilities (Tiburski et al., 2016). The current situation is similar to that of the 1970s and the multiple networks involved in ICT which were later unified by the TCP/IP protocols (Bujari et al., 2017). Work is being conducted into finding similar common protocols for IoT. Some of these, from a security standpoint, are being developed by the Internet Engineering Task Force and the Open Mobile Alliance (Tiburski et al., 2016). However, if history is any indicator, the major players involved in the technology will standardize the platform. In the past few years, many different industry groups have developed and proposed their own standardized platforms for adoption by the wider industry. One of the most influential is the AllSeen Alliance, consisting of Microsoft, Qualcomm, and LG among others, which has created AllJoyn as a common platform. The Industrial Internet Consortium (includes Intel, GE and AT&T, among others), the IPSO Alliance, Open Internconnect Consortium (includes Dell, Samsung, Broadcom) and the Institute of Electrical and Electronics Engineers (IEEE) are all working on their own solutions. However, the willingness of industry competitors to work together to address this problem indicates progress.  

Achieving consensus has been difficult, but some members of the AllSeen Alliance have joined the Open Interconnect Consortium (renamed to Open Connectivity Foundation) in the past year. These efforts may be further impacted by the lack of cooperation from giants like Apple, Google, and Amazon. According to the executive director of the OCF, today’s IoT situation is similar to the “walled-gardens” prevalent in the early Internet age. Considering the three giants’ current focus on ‘connected homes’, AI personal assistants and so on, they do seem to believe they can power through on their own.

References

Bujari, A., Furini, M., Mandreoli, F., Martoglia, R., Montangero, M. & Ronzani, D. (2017). Standards, security and business models: Key challenges for the IoT scenario. Mobile Networks and Applications, 1-8. DOI:10.1007/s11036-017-0835-8.

Cerf, V. G. (2015). Access control and the internet of things. IEEE Internet Computing, 19(5), 96-c3. doi:10.1109/MIC.2015.108.

Lin, H., & Bergmann, N. (2016). IoT privacy and security challenges for smart home environments. Information, 7(3), 44. DOI:10.3390/info7030044.

Naraynan, K. (2017). Addressing the challenges facing IoT adoption. Microwave Journal, 60(1), 110.

O’Leary, D., E. (2013). Artificial intelligence and big data. IEEE Intelligent Systems, 28(2), 96-99. DOI: 10.1109/MIS.2013.39.

Sigtia, S., Stark, A. M., Krstulovic, S., & Plumbley, M. D. (2016). Automatic environmental sound recognition: Performance versus computational cost. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 24(11), 2096-2107. doi:10.1109/TASLP.2016.2592698.

Taivalsaari, A., & Mikkonen, T. (2017). A roadmap to the programmable world: Software challenges in the IoT era. IEEE Software, 34(1), 72-80. DOI:10.1109/MS.2017.26.

Tiburski, R. T., Amaral, L. A., de Matos, E., de Azevedo, D. F. G. & Hessel, F. (2016). The role of lightweight approaches towards the standardization of a security architecture for IoT middleware systems. IEEE Communications Magazine, 54(12), 56-62. DOI: 10.1109/MCOM.2016.1600462CM.

Zhou, J., Cao, Z., Dong, X & Vasilakos, A. V. (2017). Security and privacy for cloud-based IoT: Challenges, countermeasures, and future directions. IEEE Communications Magazine, 55(1), 26-33, DOI: 10.1109/MCOM.2017.1600363CM.