December 17, 2016
I never miss an opportunity to make the point that Security and Privacy are the two big challenges for IoT, and that privacy is the bigger one. Security problems can be fixed since we accept apologies and forget rapidly. Was it billions of accounts Yahoo? Privacy issues are different since it is about trust and without a proper architecture no service can cope with new requirements on privacy.
Most people respond along the lines of “I have nothing to hide” or “that’s the way people are these days”. But people living in countries where they don’t trust the government have a completely different point of view. In most western countries we have quite tough policies about what you can ask people in interviews for employment. But given the data available to employers today (provided from users by signature!) combined with data analytics, they already know much more than you can imagine, so they don’t need to ask.
This is the naive era and it will come to an end, soon! Your personal data is yours and you should only give it away if you think it is a good idea! And many organisations are equally naive today! Critical data has to be kept safe! Web services for IPR management might not be a good idea for example. At least not unless you know where the servers (and their backups) are.
The only architecture I know of today which can support future requirements on privacy, is that users own their data and opt in to share it. I’m working with Springworks in the automotive industry today. In our company, enabling mobile operators to connect cars, the owners of the cars own the data generated in the car and they opt in to insurance companies, road side assistance companies etc to get some of their data. Car manufacturers typically argue that they own the data.
Here is a good example of what will drive new requirements on privacy – a report from Democratic Media on how wearables are used to collect and sell health data. Is this something we want? I don’t think so. Consumers and enterprises will raise new requirements, and governments will follow with legislation. Proper architectures for privacy and trusted partners will be kings.
July 17, 2016
The telecom industry got a head start in the early days of IoT aka M2M. Ericsson took the lead with the 50B connected devices mantra which translated well into an obvious role for huge operators with millions of connected users and massive investments in wireless networks. And it went without saying that the obvious providers of technology and solutions would be the telecom vendors already in the family. This was a clever and well executed plan by primarily Ericsson since the operators were looking for the next massive growth opportunity after having connected most people and given them data buckets. Operators went for it with one primary caveat – we will not only provide connectivity.
Now, only few years later IoT has gone from a connectivity focused baby to a data centric young adult. IoT is only a new phase of Internet – we’re adding things to people and businesses already connected – which explains the speed of development. From a usage point of view, Internet including IoT and the enabled services and solutions will remain customer focused and very fragmented and the operators will continue working hard to figure out their role beyond connectivity.
Here’s what I would do. The two major challenges for IoT are security and privacy. The smaller one is security since we are used to repair and plug holes as we go, and since we tend to have a very short memory. But privacy is really challenging since nobody knows how the requirements will develop over time. When your data is out you can never get it back why any service will need a solid privacy architecture to cope with future requirements without having to rebuild from scratch. The user must own his or her data and decide who should get hold of it.
But privacy is not only a technical issue why we need trusted partners to help us manage our data. Today most of us trust serious banks to manage our financial data as an example. Our trusted partner needs big muscles to force and enforce proper agreements and obviously have to walk the talk themselves. Operators already manage a lot of our private data and provided they do that well, this is a great position to build the trusted partner role on. The trusted partner role can also be required to build and orchestrate a fair and robust sub-eco system in a specific market like smart homes or connected cars.
So why is privacy of importance anyway? Today most users of internet applications sign up without even reading the agreements. One often hear “I’ve nothing to hide” as the key argument to avoid the hassles of reading and thinking. But with today’s capabilities in data analytics, all digital traces and information we leave behind can quite easily be used to picture an individual, understand how a product is designed, see patterns, understand security procedures and arrangements and so on. In the wrong hands this can be really bad and this type of information is already very useful for burglars to know when people are away and industrial espionage for example. A small example of what could, technically, be in use today is recruiters knowing without asking that a female candidate is pregnant, don’t exercise or have a heart problem.
I am absolutely convinced neither consumers nor policymakers will let this development continue as today. And this is where I believe operators with a relevant brand can find their role in IoT beyond connectivity. A role which will be increasingly important and valuable as far out as I can see. Two good examples of operators I see moving in this direction today are Telefonica and Telia Company (see FAQ 6 and 7).
October 23, 2015
IoT, IoE, M2M or whatever we call it has three primary deliverables: sustainability, security and efficiency. Each of these are massively important in all organisations. And like if that wasn’t enough, they are almost always interlinked – a more secure solution is often more efficient and saves a little of our planet for example. So what’s stopping all of us from running after these benefits? It actually takes quite a lot to get there but done correctly it’s almost always worth the effort. A simplistic description of what it takes could look like this:
• understand the current business well enough to see what issues or opportunities to work on
• time and mandate to work on them
• a couple of ideas on what to do
• simplistic understanding of what’s in the toolbox today
• prototyping and user testing over and over and over again
• make a business case, plan or something like that
• guts to bet on it
• development and proof of concept testing
• pilot testing
• focus on processes and people
Prototyping with user testing is probably the single most important part since without a clear view of exactly what to do, what users want and so on, the projects most often become very expensive failures. And up until recently it was very complicated and expensive to make working hardware prototypes. But today we have a range of affordable fantastic devices for prototyping like iBeacons and TI sensortags. And together with developer tools like Evothings Studio a working prototype is only hours or maybe a day or two away from real life tests on mobiles and pads. Further more, it is too abstract for normal human beings to visualise what adding IoT to something will allow us to do. We need to try it out to understand.It is my experience that most organisations are stumbling on other things than technology when trying to figure out how to utilise IoT in their business. At the same time adding IoT for the right things and at the right time will be game changers across industries. The countries who can enable citizens to do a fair share of what the healthcare system is doing for them today (like the banks did) will make massive gains in efficiency. The manufacturer of fridges who figure out how to provide them as a service to their customers will change the game, increase predictability and help save the planet. Avanti!
None of these things are even close to rocket science, most of them have been done many times before and very few have anything to do with technology in general or IoT in specific. The thing that we are less used to is to actually connect the things physically. It’s about hardware and hardware is challenging by definition – ask the next VC you come across. Design for manufacturing, mechanics, design, labour cost, DOA, electronics, material, tools, power, freight, etc are typical things we need to be on top of. And that require quite many people to be involved.
PS. Don’t forget that the key difference between when we connected people and organisations to the Internet and now when we connect things, is that ignorance will not be an acceptable excuse.
January 7, 2015
Smart Homes is a much talked about opportunity for IoT. It has what it takes to attract a lot of companies and people including those owning, managing, visiting and living or working in them. And beyond that also companies selling products and services for them. And the concept of smart homes is fluffy enough to include the three big deliverables of IoT: sustainability, safety and efficiency, as well as things like economy, comfort, fashion and entertainment.
We all know that a specific solution for every single task or device isn’t good enough. So the approach to make a remote controller for the toaster, one for the fridge and one for the kitchen fan (I actually saw a dedicated remote controller for the fan in Italy and I’m still thinking about the use-case) will not make the job. And we also know that “this is THE network for the SmartHome” approach isn’t taking us there since we already have a lot of different infrastructure and networks in houses and we have a number of different more or less technical requirements on them.
The combination of these two insights makes it hard to come up with Smart Home solutions that will capture large parts of the market, especially if we leave aside new buildings where one can start from scratch. I suggest something like a cluster approach to the challenge where we try to combine infrastructure, applications and tools to provide attractive solutions for larger parts of the Smart Home challenges. Let me give you a couple of examples what that could look like:
- The ultimate Media solution which uses IP networks to stream content easily and flexibly to and from devices and services (bring the best from Sonos, AirPlay, Spotify, Netflix, etc). Once installed you could add, change and remove hardware and software components easily.
- A really secure managed infrastructure with a tough SLA for applications and services that require an infrastructure to really trust. Applications could be alarms, door locking systems, smoke detectors and other things you are ready to pay extra for if the service is guaranteed.
- A kitchen app that interact with all your favourite Internet services for cooking and shopping, your kitchen appliances regardless of brand (or not?) and maybe energy monitoring and advice relating to the kitchen.
I have concluded that we will have at least three networks in our homes: an unmanaged Wi-Fi network which is already there, a managed very secure network with top-notch quality of service and a more generic but still managed network for things like home appliances. WAN solutions will generally speaking be too expensive and will just complement the LANs the way they always had. But some devices connected directly to a mobile network and/or narrow band WAN infrastructure like Sigfox will most certainly be part of the solutions.
One of the most interesting projects I’m currently involved in is a joint effort between a number of Swedish real estate owners and members of our alliance for Swedish IoT entrepreneurs (SMSE) with real estate focus. The request from the real estate owners was “a secure, robust and open service platform for multi-dwelling buildings” which they can install now, keep for years and have app developers to start bringing innovation to tenants, owners and maintenance staff.
We’re working with several technologies and one of the most interesting one is the well established Internet chat protocol XMPP since it provides a promising open architecture to deal with data integrity and privacy issues. We have already pulled together the bits and pieces required to run our first hackathon creating mobile apps on building automation systems talking XMPP. I’m looking forward to the next few months of this project which hopefully include a major hackathon demonstrating the power of this approach.
October 16, 2014
IoT is hotter than ever! Gartner just placed IoT at Peak of Inflated Expectations in their Hype Curve and forecast 5-10 years until the Plateau of Productivity. But our baby is growing much faster than anything else we have seen before and I stick to my previous view that our baby will be grown up, but still young, 2016 after only three years as teenager. Good enough technology and infrastructure are in place since a couple of years and when organizations started to go from Powerpoint and thinking to trials and pilots we reached the teens. Four things have been missing to leave the teens behind: solid participation from the IT players, a hot M&A market, active and seriously engaged enterprises and efficient easy-to-use prototyping tools for users of IoT. All these things are starting to happen now which is one of the reasons why I dare to challenge Gartner on their projection. But there is another aspect of IoT which is underestimated: how the value is created.
In most cases we build something and when it’s done we start harvest. And if customers like what we built it takes off. It might take a couple of years at least to plan, develop and start produce, then we start market and after another year or so it might take off. That explains Gartner’s 5-10 years to Plateau of Productivity, if one ever gets there. But IoT applications deliver value when the information created is distributed to an IT system, shows up in an app, makes an alarm go off somewhere or change a road sign. Initially all IoT applications had to be created end to end – from sensor to terminal – which made them expensive to make and maintain. But now we can leverage existing networks, platforms, tools, terminals and applications making it much cheaper and quicker. So far we have seen this primarily in the consumer market where a connected sensor providing data to an app has been good enough. In enterprises data management and delivery is more complicated and changes in processes and business models takes time, but they are getting there. When they do, the operational value (cheaper, faster, etc) will be obvious and the strategic value (brand, innovation, employer attractiveness, etc) will be visible in the horizon. One only has to look at what GE is doing with Industrial Internet
to understand that the impact will be massive.
Everyone promoting the story about billions of connected devices delivering data to impressive Big Data systems creating trillions of $ benefits clearly put IoT at the peak of inflated expectations. But all hard-working organizations and entrepreneurs working on industry or company specific IoT applications, well-integrated and cleverly implemented, are changing the world very fast. These efforts will start pay off soon putting competitors who haven’t started yet in a very difficult situation similar to when the frequent flyer program happened or Richard Fosbury jumped 2.24 at the Olympic Games in Maxico City 1968 using a “redicolous” new technique.
The only major difference between when we connected people and businesses to the Internet and when we connect things is that ignorance will not be an acceptable excuse this time. Beyond some clever start-ups the winners will be organizations who best understand when and how to improve their business using IoT solutions.