I took part in an M2M roundtable event last week in London. The gist of it was – the M2M market has promised so much, why is it taking so long to deliver? This is a question that keeps coming up, so worth exploring.
There were 15 participants in this closed event, many with deep experience in M2M, others with less – so a useful cross-section.
It was generally agreed that talk of 50 billion connected devices by 2020 and other similar numbers have set high expectations for M2M that are not particularly realistic. However, it has also focused minds on the possibilities, which is useful. In addition, M2M tends to be equated with cellular connections, whereas in fact M2M (and Internet of Things) solutions use all forms of connectivity, including fixed line and a wide variety of short range wireless options in addition to cellular, satellite and other wireless wide area network alternatives. When you also take into account the numbers of sensors that could usefully be connected, the overall numbers get very large, very quickly. It is then not particularly meaningful to focus on the precise numbers of connections there could be. Other factors become more important – like how to create value from the data generated by all those connected devices. There is no point in connecting devices just for the sake of it – doing so needs to create value that people are prepared to pay for.
In fact, M2M as a business solutions market continues to grow at a steady rate and will continue to do so for the foreseeable future. There are certainly a range of initiatives that can speed this growth up and many are in progress already, but the business market is limited by the speed with which companies can implement these solutions and the business process change this normally entails. In my last blog post I talked about the term Industrial Internet now creeping in as an evolutionary step to M2M. Championed by General Electric (GE), this is an approach that promises to make intensive use of data from very large numbers of remote sensors. We call this “M2M on steroids” because it is essentially about M2M business solutions with much more intensive use of data. By all accounts, this type of use is not far off and provides a further path for M2M business solutions growth.
But what about connected device solutions for the consumer? Are these taking off as quickly as they should be? The short answer is – not yet. Beecham Research refers to such solutions that are focused either directly or indirectly on the consumer as Internet of Things (IoT). Very rapid growth in numbers is much more a consumer market characteristic than a business market one.
The truth is – many of the IoT ideas need to be made a lot simpler than they currently are and consumers need to want them. At the same time, they need to make money for those who provide them – or at least, they need to provide value to their businesses. This is often not the case at present and the business models do not look very attractive. There is a lot more work needed for this potentially enormous part of the market to reach its potential.
We were in San Francisco two weeks ago, hosting our Digital Things executive brainstorm. This covered a lot of useful ground, including a discussion on how significant the concept of the Internet of Things actually is. Participants ranged from the highly practical – from a background in what works in the M2M market today and what tends not to work – to the aspirational – can this help humanity as a whole?
The general sentiment, as expressed in the instant voting, was that the Internet of Things truly represents a huge new opportunity and not just a new way of looking at existing trends. There were also dedicated sessions on business models for IoT, IoT platform requirements and on Big Data. You can see the agenda for the event at this link. Next stop – London, where we will be discussing a similar agenda on June 6.
It is certainly the case that much of what is being discussed right now regarding the Internet of Things is consumer related. A lot of the business models associated with consumer connected devices do not work very well at present. This is in contrast to those for business related connected devices – usually referred to as M2M – which generally do work and have increasingly proved to be highly effective. There is a lot that needs fixing in the consumer connected devices segment.
However, the term Industrial Internet is now creeping in as an evolutionary step to M2M – see our Snaps newsletter article this month on Big Data at this link as an example – and this promises to be hugely significant. Championed by General Electric (GE) and, using slightly different terms, also by Cisco Systems this is “M2M on steroids” as indicated in our newsletter article. The prospect is that – through much greater awareness of how machines are operating in their overall environment – much greater optimization of operations are feasible. This means a lot more sensors and the creation of a lot more data, with data analytics techniques being increasingly used to provide timely and precise information. This promises both further refinement to already-proven M2M business models as well as the introduction of new ones, and potentially substantial disruption in high use segments of the M2M market.
If you think there are just a few, then you would be wrong.
Beecham Research has been researching and analyzing the Wearable Technology market and its potential for the fashion industry for some time. As part of this, we thought it would be useful to create a Wearable Technology Application Chart to show just how broad this market is. You can access this chart as a free download at this link – Beecham Research Wearable Technology Application Chart.
Take a look – you may be surprised. In all, we have identified 20 application groups within 6 key sectors. Within each application group there are then many different examples already being used. At the outer edge of the chart are images showing examples of products already out in the market for each application type. And this is just the start . . . there are many more to come, whether the rumors about Apple getting involved in this are true or not.
This chart was first featured at the M2M Evolution conference in Miami on January 29, and was subsequently launched at the Wearable Technologies conference in Munich on February 4. It also featured at Wavefront’s M2M Summit in Vancouver also on February 4. The feedback we have had from these and other viewings has been very positive so far – that the chart provides a good representation of all the main applications currently in use or under development.
For those familiar with Beecham Research’s M2M Sector Map – available here for free download – there is of course a close resemblance in format. Our reason for that is, there is a similar underlying message:
This is a market with a very broad range of applications with many just waiting to be picked up and used.
Some of the key enabling technologies are now excitingly real. We will be talking about those and the real potential revenue for Wearable Technology – also known as ‘Connected Clothes’ or even ‘Digital Clothes’ – in our new report on this market, which will be available shortly.
In the meantime, we are still looking for further feedback on this chart. Have we missed anything important? Do let us know . . . !
‘Connectivity’ for consumer devices was a much bigger story at this year’s CES than in previous years. The one that particularly caught my attention was the connected fork, part of the line-up featured in this article. Why would you want to connect a fork to the Internet? In this case, the reason is to detect how fast you’re eating and maybe encourage you to slow down. That way you might discover you don’t need to eat as much as you thought – an aid to reducing weight.
This just goes to show that, if you look hard enough, you can find a service opportunity from connecting virtually any thing to the Internet. It’s all about looking at devices and objects from a different perspective: what could you do with data provided by the thing? Of course, the service dreamed up may not be compelling enough to make a fortune, but that’s a different matter.
I have been talking and writing about connecting things to the Internet in order to create new services for a long time. Up to now, though, these have been mainly business-related things, from CCTV cameras, combine harvesters, digital displays and jet engines to MRI scanners, point of sale terminals, trucks, vending machines, even clams to get an early warning of pollution in sea water and many more. What is really beginning to capture the imagination now, though, is connecting everyday things to help improve our lives. We already have the connected pillbox that glows to remind you to take your medicine at the right time. Now we have the connected fork and a whole host of other new ideas coming up.
Another developing theme at CES was wearability. The idea that we will make better use of technology and generally improve our lifestyles by wearing it, either as easily-worn devices or integrated into our clothing in some way. This picks up on a developing trend, where technology moves from functional to usable to wearable. One can draw an analogy here with mobile phones. In the beginning, we had those “bricks on sticks”. Functional yes, but not particularly usable beyond the precise purpose it was designed for – making phone calls while moving. This evolved through several generations of mobile handsets, until we had the next breakthrough which was the iPhone. This introduced a next stage of evolution – a highly usable, portable device with a virtually unlimited number of applications. So what is the next stage of this evolution? Perhaps when the parts of the device are split up and you wear them. The human interface might then change again, no longer relying on hard or soft buttons to make them work.
The point about all this is that the time for connecting everyday things to the Internet has truly started to arrive. There will still be lots of new ideas for connecting business-related things, but there is now a clearly developing momentum for connecting consumer-related things as well.
Where does IOT (Internet of Things) fit with M2M? Are they different, or are they the same thing? Is one transitioning into the other? These questions were a little esoteric a few years ago – they really did not matter. We think they do matter now though. Regarding IOT, how can you build it if you can’t agree what it is and what it’s for? At the same time, how can you forecast it accurately?
Regarding the definition part of this, Beecham Research’s view is that M2M as we know it – “M2M Now” – is transitioning right now along two different paths – what could be called “M2M Future” and IOT. This transition was outlined in the summary of our recent survey for Oracle – see this link for more details. Here’s a key extract from that report:
This body of research indicates that there are two key trends underway in the rapidly-developing M2M market.
The first is in the B2B segment. Here, M2M data from remotely-located assets and devices is increasingly being used for strategic purposes and value creation throughout the enterprise. This is a far cry from the early days when M2M data was strictly the domain of the service department and used for mainly operational purposes.
The second is in the B2B2C segment, where the opportunities associated with the Internet of Things (IOT) are now becoming apparent. IOT in the B2B2C segment will be characterised by data from large numbers of remote devices and sensors in one sector being combined with data from other sectors and with data from social media. This will impact consumer lifestyles and provide enormous potential for new services.
To avoid confusion with other definitions of IOT that are out in the market right now, we refer to our definition of IOT in the B2B2C segment as “Digital Things”. We think this draws a necessary sharp distinction between what is required to support business operations and strategies in the future (M2M Future) compared with support requirements for the consumer, or citizen in such application areas as Smart City (Digital Things). It also follows that this distinction is really at the application level rather than the network level: we expect basically the same network infrastructure will carry both application types but the application support requirements for each will be very different.
We like this definition because it builds on what is already in place, rather than seeking to replace it with something ill-defined that is not in place. We think it also makes it clearer where the new thinking needs to be: not so much in the M2M Future area that is already reasonably well understood, much more in the Digital Things area that – in our view – is not mapped out at all. We also think the standards work would be better focused on the Digital Things segment where there are currently very few platforms, rather than on the M2M Future segment where there are already plenty.
All of this has profound implications for M2M forecasting. I very much agree with the sentiment in Jeremy Cowan’s article this week – The $$ cost of M2M hype and false hopes – which questions the accuracy of recent big number M2M forecasts. It is becoming very much more important that M2M forecasts take into account likely market changes over the next few years because of the infrastructure investments required and the likely growing dependency of businesses on their connected assets. Then there are the big bets that M2M suppliers need to take to grow their businesses. To take just two examples, 5 years ago the M2M market had only a handful of M2M platforms and most of those were in the fixed line market. There were no connected consumer products like eReaders. Now there are over 100 M2M platforms in the market and eReaders are transforming the publishing business. 10 years ago there were no connected residential smart meters and no real justification for installing them. Now Italy and Sweden have already completed nationwide installations in every home. On the other hand, 10 years ago it was confidently expected by some that virtually all vending machines would be connected within 5 or 6 years. The reality now is well short of that and likely to remain so – except in Japan. Car telematics and home healthcare monitoring were other applications thought to be on the cusp of take-off 10 years ago. They still are.
No forecast is going to be entirely accurate. The longer range it is, the greater the opportunity for it to be wildly out. On that basis, it seems to me that a 10 year forecast based on projecting today’s market offerings has virtually no chance of being even remotely accurate. Segmenting a forecast according to key concepts like those outlined above should help to limit those problems.
What do Smart Cities and M2M platform standards have in common? In Beecham Research’s latest Snaps newsletter – published today – we have an article on each of these subjects, which we think should be linked a lot more closely than they currently are.
The first – The Connected City, see this link – looks at the different concepts of the Smart City and how these are taking shape. Of course, the easiest way to create a Smart City is to build it from scratch. That way you can be sure that all the services delivered are compatible and inter-operate. However, you don’t build a new city every day, except perhaps in China. Cities are generally old with a long history and already have a myriad of different services delivered in different ways, usually by very different organizations. How do you make sense of that – in the form of one over-arching technical platform through which all services are delivered – and is it worth doing in any case?
I spoke at the Smart Cities Congress in Barcelona on November 14 and had a chance to look around the show floor. I was struck by the large number of exhibitors there with large stands – from technology suppliers, service providers and system integrators to major cities – and gained the distinct impression that many think it is worth doing. I also gained the distinct impression that there is considerably more heat than light at present on how to achieve it.
The second article – The Way Ahead for an M2M Platform Standard?, see this link – looks at what is currently happening in standards development for M2M platforms, and arises out of Beecham’s attendance at the most recent ETSI workshop at the end of October. While it was clear that standards bodies around the world are trying to bring their activities together in the OneM2M initiative, it is less clear that this reflects what is actually happening in the market.
According to Beecham Research’s own count, there are already over 100, typically cloud-based, M2M platforms out in the marketplace that are being used for M2M solutions. Some deal with connectivity, some with applications and a few with both. There is no reason to believe that many of these readily inter-operate or are taking much notice of the standards work going on (TIA’s TR-50 is probably closest to current market players). So how is the market likely to achieve “one over-arching technical platform through which all services are delivered”?
There is an opportunity for the OneM2M work to be reflected in platforms for Internet of Things solutions like Smart Cities as distinct from M2M ones. It is not at all clear to us that this is the target that is being aimed at though.
We have been researching integration of M2M data with enterprise IT recently and the longer term implications of that. Today sees the publication of findings from our recent work with Oracle. Entitled “Designing an M2M Platform for the Connected World”, it assesses M2M solution requirements now and in the future. It identifies integration of M2M with IT as the second most important priority for most M2M solution requirements right now, only eclipsed by the need for end-to-end security – the subject of my last blog (see this link).
The research found that M2M data is increasingly being used more broadly for strategic purposes and value creation throughout the enterprise. Also, that it has become a means for creating new market opportunities while providing a competitive advantage for enterprise users in their own key markets. The research is based on a qualitative survey involving in-depth interviews with prominent M2M solution providers in North America, Europe and Asia Pacific regions, each providing their expert views on M2M solution requirements and trends across their customer base and live projects. That research was then backed by previously conducted quantitative surveys involving broader samples of M2M adopters.
In addition to this, I would also draw your attention to an interesting survey published late last week by Axeda (www.axeda.com), which I also wrote the introduction for. The survey is available from a link on Axeda’s home page and is also identified in the Oracle report.
Respondents to the Axeda survey were delegates to Axeda’s recent Connexion conference in Cambridge, MA, in June, where I moderated a very interactive panel session. We would categorize these as being at the leading edge among adopters of M2M technology, primarily in the Medical, Industrial and related High Tech areas. The remote assets and devices they connect in their M2M solutions tend to be large and static, are typically in constant use and any downtime is expensive – in particular the loss of critical services such as in Industrial processes or in the Medical sector.
Both of these reports provide new thinking, with the first building on the second and identifying two underlying trends in the M2M/Internet of Things space. With all the hype surrounding M2M and Internet of Things, we think these two trends will prove increasingly fundamental over the next few years.
Security in M2M has recently shot up the priority list for enterprise users, as shown in recent surveys conducted by Beecham Research. There are many reasons for this, not least the growing reliance within enterprises on their M2M systems. So the story at this link (click here for article) is not only alarming because it brings home the dangers regarding healthcare equipment, it is also timely and highlights the potential dangers for all of us.
The problem it identifies is the increasing vulnerability of computerized hospital equipment to malware infections. Such equipment is of course increasingly being connected, to improve services and reduce costs. Naturally, this makes them more vulnerable to malware, but in this environment there are additional regulatory constraints. As a result, manufacturers of such equipment often will not allow their equipment to be modified – even to add security features – because such modifications might then fall foul of regulatory restrictions. That can lead to a lot of hospital equipment not working properly for extended periods of time . . . or updated as a special exercise at higher cost.
Issues with software updates to patch vulnerabilities is by no means unique to hospitals. For all types of automation systems, installing updates can change the way that system services work or are configured, which can make the updating process itself problematic. Then there are those situations where software in automation systems is frequently in use long after the vendors have stopped supporting and updating the software. This makes older systems increasingly vulnerable to attack. Cheaper to operate, but more vulnerable.
As M2M continues to develop in new and more critical areas – such as Smart Grid deployments for example – the concerns about vulnerabilities to attack will continue to rise. The M2M industry will address these in ways that will increasingly become highly technical specialisms. A concern though is the impact of all this extra cost of security on the growth of the overall market from a business perspective. The M2M market has always been viewed as particularly cost sensitive.
Will all of this increasingly sophisticated prevention work just serve to kill the M2M patient commercially? Or are these security needs and threats leading to a new range of business opportunities, delivering real value that the market needs and wants?
I attended Oracle’s Java Embedded@JavaOne conference this week in San Francisco – running alongside Oracle OpenWorld – and presented findings from Beecham Research’s survey of M2M solution providers, which was sponsored by Oracle. One of the issues this raised was the suitability of using Java as a single software development environment across whole M2M solutions. As asked in a previous blog post, would the market gain from there being one development environment like Java for all parts of an end-to-end solution either now or in the future?
There are very different views about this. On the one hand, Cinterion (Gemalto) would certainly agree. The company launched its new M2M EHS5 cellular module this week that features Java ME Embedded 3.2, a new Java implementation aimed at small embedded devices. From Cinterion’s viewpoint this offers the opportunity to run the same version of Java across a whole M2M solution (E2E, meaning end-to-end, edge-to-enterprise or embedded-to-enterprise – whichever you prefer), thereby ensuring compatibility over the lifecycle of the solution among other benefits. On the other hand, many device designers would disagree. For them, Java still has way too much overhead for many small devices where, for example, C++ is more appropriate. These designers maintain that using Java drives up both the size of the processor and memory, which makes the design much more complex and costly than it needs to be at the edge.
In other related news this week, Qualcomm also announced it was working with Oracle to bring Java ME Embedded 3.2 to their Gobi module solution for 3G. ARM also made clear this week that it sees Java as a natural partner for both its high end and low end processor designs, only about 45% of which end up in smartphones.
This issue will run and run but Java’s momentum in M2M is clearly building. There is also the thought that – inevitably – even small devices with any on-board processing will increasingly need more power to handle more data or manage peer-to-peer communication, or both. Will Java then be a more natural choice even for these smaller devices – currently some of the more numerous devices in M2M ?