Connected versus Intelligent Devices in the IoT – and in Saunas

Aapo Markkanen
March 10, 2015

 

One of the most profound questions affecting the Internet of Things at the moment is where the smarts in smart systems will reside. The first phase of the IoT – an extension from its terminological precursor, M2M – has been based on the premise that the device itself is rudimentary and any intelligence in it comes from the cloud level. Moreover, in many cases “intelligence” has not been a priority to begin with, and the application has been developed to deliver only remote control or servicing initiated by a human operator, without any real need to capture and process data.

We at ABI Research refer to this approach as the connected device paradigm. For players involved in M2M/IoT it has been more of a necessity than a choice. The connectivity problem simply has been solved before the computing problem, so it made sense to design the early systems in this way.

Advances to computing and power consumption, however, are beginning to pull intelligence from the cloud to the network edge. The increase in so-called edge intelligence, in turn, is making the available architecture choices more nuanced and allowing organizations deploying the IoT to enhance their physical assets and processes in novel ways. As a result, the industry is now on the verge of the intelligent device paradigm. It could be summarized to come with three advantages:

Latency: Processing data at the edge reduces latency from the actions. Use cases that are highly time-sensitive and require immediate analysis of, or response to, the collected sensor data are, in general, unfeasible under cloud-centric IoT architectures, especially if the data are sent over long distances.

Security: By and large, sensitive and business-critical operational data are safer when encrypted adequately at the edge. Unintelligent devices transmitting frequent and badly secured payloads to the cloud are more vulnerable to hacking and interception. Additionally, enterprises may need to secure and control their machine data on the pre-cloud level for compliance reasons.

Total Cost of Ownership: The paradigm shift can reduce the IoT systems’ total cost of ownership, or TCO. Intelligent devices are more expensive upfront than less sophisticated alternatives, but their TCO over a long service life can prove substantially lower, owing to reduced data service costs and extended battery life.

The edge intelligence can take place on the endpoint device as well as on routers, switches, and other gateway/hub devices. The latter method has the benefits of having a less constrained form factor and often access to the power grid, so the chances are that for the time being most IoT practitioners will find it to be the more compelling of the two.

The gateway model, in which the gateway device communicates with endpoints over a short-range wireless technology, appears to be gaining traction in both industrial and consumer-facing settings. The ongoing innovation in mesh networking is likely to give a further boost to gateway-driven intelligence over the next couple of years, by opening up more extensive short-range designs.

At the same time, it is important to understand that the cloud level, as such, will by no means be going away. Rather, it is likely that the IoT will reshape the cloud as a concept, and make it something more distributed than it has been during the “digital-first” internet era, with its vast and centralized data centers. For the physical-first Things, the cloud may well turn out to be a network of local (city-level) or even hyper-local (neighborhood-level) layers that deals with the data as close to the source as viable.

Ultimately, no single IoT architecture can fully address all possible use cases. The key learning here should be that organizations betting on the IoT are finally starting to have real technology choices at their disposal, and those choices should depend on the characteristics of the use case – where and when the data need to be processed, what the security requirements are, and what the estimated TCO looks like. For some use cases, the best set-up is still more about being merely connected, whereas for some others there’s a case for investing in smartening Things up and making them truly intelligent.

Sauna stoves

To demonstrate how the difference between “connected” and “intelligent” device paradigms plays out in real life, let me cite a carefully selected, if potentially unorthodox, example from the consumer market and tell you a little about sauna stoves. On a global level, the sauna stove is an admittedly niche product category, yet in my native Finland it represents a market with a remarkably high installed base. According to the most reliable estimates, there are slightly over two million saunas in the country, against a total of 2.5 million households. Each sauna has one stove, which is heated by wood or electricity.

In recent years, the Finnish sauna industry – starting from its high end – has been experiencing an increasing uptake of connected sauna stoves that can be turned on and off remotely, usually from a mobile application. In this paradigm, the sauna itself remains fairly dumb and unaware of its physical context, so while the connectivity element provides a nice dose of everyday convenience for the end user it poses also a serious risk for life and property if used carelessly.

See, many sauna users tend to use the stove not only for bathing purposes but also for drying sports gear and other personal accessories on the rocks that cover the stove’s fireplace or electric heating elements. Naturally, the latter, secondary use case can be applied safely only when the stove is turned off. If the rocks are hosting anything inflammable when the stove is on, the drying process turns into an immediate fire hazard. This makes the primary use case problematic when cloud-enabled remote control is added to the equation, since as a user interface it is inherently riskier than the traditional one that requires the user to initiate the heating by applying a match or a switch in the same physical space as the device.

According to the national safety authority, fire incidents caused by saunas are on the rise in Finland (from 52 in 2010 to 156 in 2013), after declining for years, and based on its study the trend can be largely attributed to the growing popularity of remote control. This adds an exotic flavor to the matter of “IoT security”, and goes to show rather tangibly how equipping a physical-first device with connectivity, but with no intelligence, can have very negative consequences.

In this example, the solution to the problem is easy to see. Ideally, a sauna stove would have a set of sensors that prevent remote control if they detect anything on the rocks that does not belong there. The data readings would be processed either by the stove itself (i.e., the endpoint) or by a smart-home hub (i.e., a gateway), and besides the improved product security the paradigm shift could add further value. Adjusting the heating cycle (to factor in a prolonged evening run, interworking with a wearable activity tracker), or the temperature (to factor in different preferences amongst the bathers, interworking with a smart showerhead), are examples of such next-generation features that spring first to my mind.

Sadly, though, even the intelligent and learning sauna does have one fundamental downside. It is not compatible with wood-burning stoves, which in terms of heat quality are superior to electrically heated ones. The ideal user interface does not always equal the ideal user experience.

 


 

Oleg LogvinovPrincipal analyst Aapo Markkanen leads ABI Research's Internet of Everything Research Service, contributing to various research activities related to Internet of Things, M2M, and big data. In his research, he explores areas such as predictive analytics, product lifecycle, quantified self, contextual awareness, cloud platforms, and IoT developers. Before joining ABI Research, Aapo worked as an analyst at IHS, where he was responsible for providing market intelligence and strategic analysis on the European telecoms sector and its leading players. He holds BSc and MSc degrees in management studies from the University of Tampere, Finland.

 

 


 

Subscribe to the Newsletter

Join our free IoT Technical Community and receive our Newsletter.

Subscribe Now


Calendar of Events

2017 IEEE International Conference on Software Quality, Reliability & Security (QRS 2017)
25-29 July 2017
Prague, Czech Republic

2017 IEEE International Symposium on Electromagnetic Compatibility, Signal and Power Integrity
7-11 August 2017
Washington DC. USA

The 3rd International Forum on Research and Technologies for Society and Industry (RTSI 2017)
11-13 September 2017
Modena, Italy

The International Workshop on IoT applications in Intelligent Transportation Systems and Logistics
Workshop co-located with the 2017 IEEE International Conference on Service Operations and Logistics, and Informatics (SOLI 2017)
18-20 September 2017
Bari, Italy

IEEE IoT Vertical and Topical Summit
18-20 September 2017
Anchorage, Alaska, USA

IEEE International Conference on RFID Technology and Applications
20-22 September 2017
Warsaw, Poland

The IEEE World Forum on Internet of Things (WF-IoT)
5-8 February 2018
Singapore

See More Events


Call For Papers

IEEE Internet of Things Journal
Software Defined Networking for Internet of Things
Deadline: 1 August 2017

Theories and Applications of NB-IoT
Deadline: 15 August 2017

Towards Positioning, Navigation, and Location Based Services for IoT
Deadline: 1 September 2017

Integrated Computing: Computational Intelligence Paradigms and Internet of Things for Industrial Applications 
Deadline: 15 September 2017


Past Issues
May 2017
March 2017
January 2017
November 2016
September 2016
July 2016
May 2016
March 2016
January 2016
November 2015
September 2015
July 2015
May 2015
March 2015
January 2015
November 2014
September 2014