Many people are still getting their head around public and private clouds. Even less know about Internet of Things. However the real revolution will be bringing both together and for that you will need a proximity cloud.
What is a proximity cloud?
In five years time there will be billions of sensors everywhere. You will be wearing them on your body. They will be in your house, at your job, in your hospital, in your city, in your car/bus/plane, etc.
Now in a world where 1 billion people will have access to 10’s or even 100’s of connected devices around them, it is easy to understand that you don’t want to send all data they generate to a public or private cloud. There is just not enough mobile spectrum or fiber capacity to cater for this.
What we need is put intelligence close to the data generators to determine if your heartbeat, that video camara stream, the electricity consumption of your boiler, the brake capacity of your car, etc. are within normal limits or are outliers that deserve some more attention. Especially if video is involved you don’t want to send it over a public Internet connection if it has nothing interesting on it, e.g. the cat of the neighbours just got onto your lawn.
So houses, cars, companies, telecoms, hospitals, manufacturers, etc. will need some new type of equipment close to the data avalanches in order to filter out the 99.999999% of useless data.
An example. If you are a security company that manages tens of thousands of video cameras in thousands of companies then you want to know when a thief walks by or an aggression happens. You will train machines to make decisions on what the difference is between a human walking past or a cat. However burglars will soon find out that computer vision has a fault and when they wear a cat disguise they can get past. It is this type of events that will trigger a central cloud platform to request all videos of a local business in the last 24 hours and to train its visual model that humans can wear animal suits and then push this to all proximity clouds in all its customer premises. The alternative is storing all video streams in the cloud which would require enormous bandwidth or even worse, not knowing what happened and being in the press the next week for being the “cat-suit” flop.
As announced previously if you are into IoT, proximity clouds, robotics, next-generation networking equipment, etc. then you want to subscribe to Telruptive because I plan on giving readers early access to IoT innovations that are non-public. This message will be the first of this kind. Ubuntu will make some major IoT announcements in the coming months. The type that change industries. So if you are working on something that can be connected to the Internet and can house an ARMv7, Intel Quark or better then please reach out to me on LinkedIn and tell me what you are working on. If it meets our innovation criteria then we will let you in on some of our secrets before anybody else. We want to give your innovation wings and make 2015 a magical year for innovators…
Today I had a meeting that could be the beginning of the end of RFPs to buy software. RFPs are the tool established buyers and vendors use to keep new entrants at bay. However I haven’t met anybody that says they love writing or responding to them. The effect of RFPs on software is perverse. The main problem is that you can’t ask if your software is beautiful, easy to use, fast to integrate, efficient, effective at solving a business problem, secure, etc. Instead you ask if you provide training, because you assume it is ugly and difficult. You ask if they offer consultancy services and an SDK or connector library because you assume it is difficult. You assume you need to customise it for months because it will not be effective out of the box. But most importantly since you will be stuck with the software for years, you ask if it supports any potential feature that perhaps in 5 years might be needed for 5 minutes. It is this last set of questions that kill any innovation and ease of use in business software. A product manager in the receiving end will get funding to add those absurd features when customers ask for them. A career limiting move would be to ask for budget to reduce useless features or tell that your product looks worse than Frankenstein.
So how can you make sure that software is beautiful, does what it supposed to efficiently and effectively, is fast, nimble, easy to use, secure, scalable, fast to integrate, is future proof, etc.? You do what you do when you buy a car, you go and ask the keys of different models and take them for a serious spin and put them to their limits.
So what you propose is a three months PoC for each potential solution?
No what I propose is being able to get your hands on all different alternative software solutions and deploying, integrating and scaling them in hours or even minutes and then release a bunch of automatic performance tests and rough end-users, even some ethical hackers or competitors.
If the software does what it says on the tin, is effective, efficient, beautiful, secure, fast, scalable, easy, etc. then you negotiate pricing or use it for a minimum valuable product.
It used to be impossible to do all of this in hours but with solutions to deploy quickly private clouds and cloud orchestration solutions like Juju, we are actually planning on trying this approach with a real customer and real suppliers. To be continued…
IoT is one of the biggest potential new revenue streams but also one of the most challenging technical problems we have today.
the technical challenges
IoT is not only sensors + Big Data analytics + cloud + short-range low-energy networking and Internet. The real problem is that you have to be good at many different technologies that used to be separate and that one mistake can have disastrous effects. You have to be good at miniature sensors that need to be able to run two years on one tiny battery and use software that even the biggest geek hates to work on. At making sure IPv6 networking is adjusted to this small footprint devices with innovations like CoAP and 6LoWPAN. To learn about the world of micro-controllers open source hardware like Arduino, micro-computing platforms like Raspberry Pi and Edison, ARM Cortex, Intel Quark, etc. You also need to know about new and old low-energy networking technologies like Zigbee, Bluetooth Low-Energy, etc. Afterwards you want your sensors to be connected to a hub because otherwise you would need a SIM or Wifi in each sensor which would drain battery. So you need to make a smart hub that ideally can run apps from different developers and can support lots of new devices. However you also want devices to support peer-to-peer technologies like Thread or new standards from Intel, Qualcomm or any of the numerous standardisation bodies. You want to use 3D printing to print an attractive casing. You want to use crowd-funding to sell your Smarthub. You want mobile apps to work flawlessly with IoT. You need to know about Powerline, gesture control, in-building location tracking, voice control, etc. if you want to compete with the best smart hubs. You now need to know GPRS, 3G, 4G, White Spaces, Long-Haul Radio, WiFi or fiber broadband to communicate with the rest of the world. On the cloud side, being it public or a private OpenStack, you need to use the latest DevOps tools, Cloud Orchestration tools and containers like Docker, to deploy scale-out queues, real-time stream processing and other Big Data analytics solutions. You need to be able to train deep belief networks and push models to hubs and sensors. Recognise threatening video images. You need to be able to do rolling upgrades and continuous deployment of updates, developer apps, etc. Manage operations of millions of devices and billions of sensors. You want a store. A developer eco-system.
Now when you finally mastered all of this. Make one security mistake and a hacker on the other side of the world is able to control your house, business, city or country.
Now the business opportunities are huge as well. Save a couple of percentage on the production costs of a car and you can save hundreds of millions. Track a global epidemic or the vital signs of a billion people and you can save millions of lives. Give millions of developers a new way to channel their creativity and the Angry Bird of IoT will bring new industries. The one that changes the habits of people will be the next billionaire.
Where is the money? Industrial IoT. Where is the innovation? Home automation and wearables. You can’t pick one. You need to connect innovation with money if you want to lead the IoT revolution. If somebody else does it for you, they can make your solution irrelevant.
Charles – Chuck – Butler, a colleague at Canonical, wrote a very nice blog post explaining the basics of Big Data. It does not only explain them but it also allows anybody to set up Big Data solutions in minutes via Juju. Really recommended reading:
This is a good example of the power of cloud orchestration. Some expert creates charms and bundles them with Juju and afterwards anybody can easily deploy, integrate and scale this Big Data solution in minutes.
Samuel Cozannet, another colleague, used some of these components to create an open source Tweet sentiment analysis solution that can be deployed in 14 minutes and includes autoscaling, a dashboard, Hadoop, Storm, Kafka, etc. He presented it on the OpenStack Developer Summit in Paris and will be providing instructions for everybody to set it up shortly.