Berkeley pushed the robotics innovation boundaries by using machine learning to teach a robot what to do via trial and error and without actually telling it what to do. This is an example of how pretty soon in many industries: big data, machine learning, cloud, robotics, apps on smart devices, etc. will come together in new ways. Robots and machine learning used to be separate sciences. Now computer vision, deep learning, etc. will all become apps on robots. Robots will share sensor data with clouds where big data storage and online machine learning will continue to improve algorithms and then push updated models back to the robots. Once a model is stable and predictable, you will be able to upload it as an app to a robot store and other robot lovers will be able to use it for new purposes. Combine this with SaaS services to learn from these new uses and there will be a continues circle of new uses generating new insights, leading to new uses, leading to new insights, etc.
Translation for the science fiction lovers: Star Wars like robots will come sooner than later…
Many people don’t realize that IoT in the home means apps and app stores everywhere. Single purpose equipment will solve single problems but appliances with apps will bring synergies. GE showed the first smart fridge with apps on the Ubuntu stand in IoT World but this is just scratching the surface. Home appliances will eliminate lots of small single purpose devices. You don’t need an alarm key pad if your dish washer has an LCD screen that displays keys while the alarm is on. Neither do you need a thermostat like the Nest if your fridge already measures your room temperature and your alarm can tell if you are home or not. You also don’t need a fixed phone. Your Bluetooth speaker has a microphone. So does your TV. A Bluetooth headset is for private calls. Your modem can run the VoIP and cloud telecommunication app. You don’t need a sprinkler controller, a swimming pool controller, a garage door controller, an air-conditioning remote control, an alarm clock, a radio, etc. Apps on always on appliances will eliminate them. Expect your home to listen to voice, gesture or even heartbeat commands.
Once upon a time there was a black box. It felt bad about itself because all its brothers and sisters had a purpose in life. They were all excellent at doing something special. The black box on the contrary was just good at connecting to others. Though it was smarter than most and had a bigger housing that its siblings, there was nothing special about this black box. One day the whole family made a trip to a field. Everybody expected the field to be green. However when they arrived they found an ugly brown field full of garbage that was accumulated through the years. All the good looking siblings very quickly crashed because they expected a clean surface or got overwhelmed by the many garbage boxes that came to great them in very ancient languages that did not fit their handshake. The black box however was able to connect to all the garbage and reinvent itself over and over again. Its flexibility was limitless because it just needed to go online to get some new extension.
This was the story about the industrial gateway that choose not be sexy but instead be smarter and have more storage and above all rely on an external app store. The story will not win a price but going for a fixed function, fixed API, fix standard, etc. will result in a lot more scary story. Unfortunately the wolf will come before the ugly duckling becomes a beautiful swan…
Every time there is a new technological trend, you get market researchers asking questions to CxOs about what they think about the trend and afterwards selling the results for lots of money to whomever wants to buy the report. However if you collect reports like this and start comparing them you will see a clear trend. To the question, do you fully understand the capabilities of the new trend, lots of people will say no. To the question, what part of your business will be impacted, you will see revenue, cost, efficiency and innovation being mentioned. To the question, where are you with projects regarding the new trend, the answer is planning or evaluating, prototyping at most.
If you see this pattern, then DON’T buy the report. CxOs are not magical people that automagically understand new trends and will tell the industry what to do. They actually have no clue.
In 2010, would you rather have had the Oracle CTO set your cloud strategy or the Amazon CTO? Would you rather have the Oracle CTO define your CRM strategy or Salesforce? Oracle vs Cloudera or Spark for Big Data? Oracle vs Docker? Oracle vs Golang? Your answer should be Oracle denied the existence of cloud, SaaS, Big Data, containers, more productive programming languages, etc. until the market obliged them to embrace them. So if Oracle, which is supposed to be a leader in enterprise software, does not understand the latest IT trends, why do you think that a CTO from a telecom, retailer, airline company, etc. would better understand these IT innovations and their impact on business? The truth is that none of them does. Not even Gartner does. Gartner and others make a summary of what is happening in the industry after the first battles have been fought. They depend on collective group thinking to make a “magic quadrant” or a summary of where the herd is moving. However the real innovators already know since one to three years ahead if something major is about to happen and they have prepared solutions for this new reality. By the time the majority has realized there is a problem or opportunity, the visionary innovators already have a solution in the market. If you want to know the future you need to follow the smart elite and read what they are very excited about.
Devops have been singing the songs of Docker since early to mid 2014. Smart engineers looked at what the problems the innovators were having and focused on solving them. The end result is that Docker orchestration started fast replacing PaaS before anybody in a corporate office heard about Docker in the first place.
Current trends are Hadoop being replaced by Spark being replaced by Flink. Java is uncool, Golang and Rust are the future. Bitcoin’s blockchain will replace lots of centrally managed markets.
So if a Gartner, Accenture or Siemens tell you what IoT will look like then be careful because you don’t know what you don’t know. You should look for visions from people that don’t focus on what “the industry thinks” but instead ask the smart elite how they would solve your industry’s main problems. Chances are they create a blue ocean strategy if you listen to them. Your competitors will instead all be focusing on what their regular supplier offers which is often the same solution as five years back after it has been cloud-washed, big-data-washed, IoT-washed, etc. Do you agree mister Oracle?
With software defined radio becoming cheap, the economics of running a network will change. Fon in Spain has been for years allowing WiFi users to share their bandwidth and customers are getting a revenue share if others pay for bandwidth. Why can’t the same be done for non-WiFi communication? Why not have a crowd-sourced mobile network? If you run a mobile base station that takes up your fixed broadband then you get 50% of the revenue generated by the network. This model could be interesting for regular mobile operators as well as disruptive operators. You can easily see how using LTE license spectrum that is currently not being used by the primary owner could be repurposed, taking the white spaces concept to the next level. Cloud-based control systems can manage hundreds of thousands or even millions of micro cells. By federating all the different micro cells you get reasonable coverage. Add an MVNO agreement for black spots and some Twilio or other VoIP magic and you have a complete working network. Stitch into this network personizable value added services, like dynamic call routing [partner and kids get priority, call centers get a busy signal, etc.], customisable voice mail, etc. Even better, an App Store for VAS. Dreamscoms would offer a better experience than others because you would get paid if you participate and services would add real value, not only costs…
The first day at IoT World in San Francisco was a very big success. The press is loving apps on any smart device (drones, robots, network switches, mobile base stations, industrial gateways, home appliances and IoT home hubs). Good examples are the General Electric Smart Fridge, Erle-Copter – the first drone with apps, Acer’s aBeing One smart hub, predictive maintenance with Dataart and Microsoft, etc. However these were only the headlines. The real revolution is not a single demo of a smart device but the fact that apps can make smart devices work together and completely redefine them. Think about the possibilities of industries being changed through apps.
Let’s start with drones
Attach a camera and put an app and If you thought the “selfie stick” was the ultimate narcissism, think again. Drones can follow you around and make videos and picture from way above. But that same camera can be used to look for crop diseases or inspect an oil drilling platform for danger. Each time the app will be different but the drone can stay the same. Now take it a step further. Attach an Oculus Rift and use some gesture based joystick and you will be able to remotely control your drone. Use some of the software defined radio wonders from Lime Micro and Nuand and you can invent your own wireless protocols. Add the DigitalChaos app and your communication will no longer be traceable. Drones will be able to be flown without knowing who controls them. Scary! However both the army and the anarchists will love this idea.
The home hub
Anything with bluetooth, usb or wifi is able to be connected. At the booth we have the Wowwee Mip, lots of Spheros, the Neurosky scanning your brainwaves, the leap motion being gesture controlled, bluetooth speakers, Play-i’s Bo Xylophone robot, USB cameras connected to Spreed webRTC conference server, bluetooth oximeters, etc. The guys from The Hybrid Group and Dataart are showing how all these things can work together. Think about creating lie detectors with the Neurosky, heart monitors and oximeters. Or using them to create the ultimate learn-to-relax machine that learns what things relax you (different music can be streamed from the cloud via your bluetooth speakers, light intensity, coffee smells, etc). Add IP cameras, computer vision and a sprinkler system and you can scare neighbourhood cats by turning on the sprinklers when they enter your garden. However humans and dogs are allowed. Based on the results from the outdoor weather station, weather predictions from the web, etc. can not only control your sprinkler system but could reprogramme your alarm clock to wake up 15 minutes earlier when it rains or 30 minutes when it snows.
Putting apps on networking equipment, like the Penguin Computing switch we have at the booth, allows a switch to become a firewall [Zope], a load balancer [F5], a virtual router [Midokura], an anti-virus, an intrusion detection monitor, and be remotely provisioned [MAAS & ONIE] and monitored [Zabbix]. Since the switch has a USB slot you can still put in a bluetooth dongle and use it to control a swarm of Spheros [thanks to The Hybrid Group]. Several LTE gateways (netcom wireless, ADLink, etc.) are at display as well. allowing for the future of LTE networking to be app driven. Now imagine the capabilities for creating optimised networking apps for telecom, media, big data – genome processing, and connect them to a series of devices that are driven via ethernet or LTE. There are several ethernet-driven I/O and sensor controllers and even a sprinkler controller.
There are several robots on the stand. RTI is doing industrial control & distributed communication via a robotic arm. Sunfounder has some a very cool build-your-own robot car and even a robot spider. However think about controlling robots with unexpected controllers, e.g. a bar code scanner, mind-control, gesture control, etc. Add computer vision as an app, an autopilot app, some motors to control the car and an Uber app into your car and you can rent it to others without getting up from the sofa. Take a Trasibot – environmental research robot boat – and have scientists share the costs of sending the boat to the middle of the ocean between many research projects.
The NetcommWireless CPE and Patton VoIP gateway are examples of what can be done with software defined appliances at work. If you have to have a VoIP gateway running the whole day, why not have it talk to your bluetooth bar code scanner and credit card reader. You can make your own POS by adding a standard tablet and a SaaS service. Use a bluetooth headset or speakers to make calls, no need for fixed telephones any more. Voice control will assure you don’t have to remember any telephone number any more. You can stream music to the shop while the telephone is not in use. Have a 3D printer in the basement. Your bluetooth speakes will notify you when it finished printing. If you put the Uber app on the 3D printer then DHL will be at your door to pick up the object that was just printed by somebody else. Why invest in a lot of extra servers when your VoIP server can also control your iBeacons inside the store. As well as LED publicity panel in your shop window. Forgerock will make sure your identity is checked when it detects that your Nymi is approaching the shop’s door and will automatically open the door lock for you and can bring up the blinds.
The GE smart fridge is an amazing appliance. Why do you need apps on your fridge? The USB powered Milky Weigh measures how much milk you have in your fridge so if you forgot to check, a mobile app will visually tell you. However since the fridge has an API, our bluetooth speaker can let you know if the fridge was opened. Temperature can be shared so if it becomes more than 50C the fire truck can be called automatically. What if home appliances could be software defined? LCD displays could show you the score of your favourite team while not doing the dishes or washing your cloths. Touch control could turn your washing machine in an alarm controller and as such you don’t need to install another appliance. Put an accelerometer on the compressor and motors and GE will be able to send an expert to service your appliances before they break.
Several industrial gateways are being shown. SCADA-controlling platforms from Emutex. PLC-controlling solutions from Cloudplugs. The Energy Detective electricity monitor can show you in detail how your energy usage is and get it under control. Predictive maintenance is demonstrated with accelerometers on top of two ventilators of which one has duct tape on one of the blades. It is remarkable how much that ventilator vibrates. DeviceHive will make the lights go red when it vibrates too much so if you want a green light, just stop it from shaking. However the future of industrial IoT is all about connecting everything to everything and using Big Data analytics to reduce costs and sell more. Any industrial IoT gateway has a USB or other standard port, hence making custom cables to USB let’s you hook up to everything. Big data scale out in the Microsoft cloud is just a Juju Charm away.
Snappy Ubuntu Core
It is amazing how one technology can change so many things. Check out Snappy Ubuntu Core…
Bluetooth has enabled a revolution of any type of cheap Chinese peripheral to talk to any tablet, PC or TV. It allows many uses via its long list of standardized profiles. However its strengths are also its biggest weakness. Bluetooth is a very complex standard and more often than not it gets abstracted away by several software layers that make it somehow manageable for “very motivated” programmers. The reason for these complexities are based on a reality that soon will stop to exist. Bluetooth assumed cheap wireless communication chips need to be designed years in advance and as such any possible use case needs to be supported in silicon not in software. Making a silicon based standard that is future proof for any use case equals adding lots and lots of complexity in the form of GATT, GAP, options, etc.
Cheap software defined radio will change the foundations Bluetooth is build upon
What if you don’t need to define everything upfront but you still have cheap silicon? That is the promise of a new generation of software defined radio [SDR], the cheap version of a technology that used to cost thousands of dollars per device. Cheap SDR will allow the Github generation to define easy to use wireless protocols. Since these protocols don’t need to support any and every use case upfront, we will see an revolution in wireless technologies that was similar to how businesses moved from highly standardized electronic data interchange [EDI] to XML and schemas and more recently to free format JSON and Yaml.
How will the future of wireless technologies look like?
If every appliance at home or in a business would have a cheap SDR then communication will be a matter of downloading apps for the two devices that you want to communicate. Likely there will be a github driven series of standard-by-adoption protocols but they will see rapid evolution and succession just like we are seeing now in programming languages where Java every day gets a new competitor in the form of Ruby, Go, Haskell, Rust, etc. Programming languages no longer have widespread adoption, neither have artists and anything that the masses can define what is hot or not. Bluetooth will be like the Beatles, a memory of a cool thing that had everybody excited back than but the next generation of wireless communication will be more like Justin Bieber, very important to some, safe to ignore for everybody else…