Page 6

EETE JULAUG 2014

Clearing away the fog of computing for the Internet of Things By Nick Flaherty Intel is bringing its expertise from silicon manufacturing to the Internet of Things with two new ‘Ignition’ Labs in Europe to drive the development of IoT systems with a specific focus on the Internet of Things for smart cities. The labs – in Swindon, UK and Munich, Germany – follow the first two in Istanbul and Stockholm last year. The labs bring together the whole eco-system of developers, from hardware to software and services, in one room to create a complete, interoperable solution over the space of several weeks. This was a lesson learned from Intel’s silicon manufacturing expertise, and particularly from Doug Davies who was previously the general manager of the Intel’s Arizona fab before taking over as general manager of the IoT group at Intel. “What we learnt from our fabs was that when something goes wrong, they don’t have a meeting a week for a month, they pull people off the line and put them in a room until its solved,” said Karen Lomas, director of IoT Smart Cities for EMEA at Intel. “So when Doug Davies came out from manufacturing to run IoT he saw there’s a lot we can learn from running a fab and this is what they do, bring the experts together, fix the problem and then deploy it to the factory floor.” The eco-system is the vital element, she says. “Intel doesn’t want to be a system integrator, we provide hardware and software expertise as part of the process, then we can figure out where the value is and where everyone can make money. Then the lab space is to physically make the products work together.” The proximity to London as a large, congested city was an important factor, says Lomas, as it is predicted to add 100,000 people a year. This equates to adding a city the size of Birmingham by 2030, and that has huge challenges for the infrastructure that can be tackled by Karen Lomas, director of IoT for smart cities for EMEA at Intel in the newly launched UK Ignition Lab with Gerry Hodgeson, managing director of IoT services company Cascade 3D. the Internet of Things. “We felt that the prioritisation is where the ecosystem partners are, and that’s in the big cities,” she said. “Some of the greatest innovations come out of need so there are some fantastic innovations come out of Istanbul and we’ve gone for some of the other leading cities such as Stockholm where we have very strong business partnership and very strong engineering expertise.” One areas of focus is the gateway, using processors such as Atom and the lower power Quark, to control the interface to sensor networks. This is particularly important for connecting legacy systems securely to the Internet, she says, and is part of the move to ‘fog computing’ with more intelligence at the edge of the network. “Intel believes the numbers are growing exponentially and when you are in the billions of devices range what you really want to do is have the data where it is needed to make the decision – the edge provides the ability to make the decision where they are needed.” She points to a recent project for smart toilets in Heathrow Airport’s newly re-developed Terminal 2 in London, which uses boards and middleware from partner Eurotech, as an example of fog computing. “With smart toilets you have a people counter, and notify cleaners every 50 people,” she said. This gives cleaner toilets, reducing queues, improving customer experience and allows customers to be directed to other toilets if one is particularly busy. “That’s all done at the edge, that’s not in the cloud,” she said. “The cloud link is storing the information for predictive analytics and trend analytics – if the toilets are hardly ever used should they have a different use such as a store cupboard. Different compute capability needs to be done at different times with different data sets. If you look at the IoT gateways we have Quark and Atom products and depending on the specification of the gateway you can do more or less of the information partitioning, depending on what the need is.” Latency is also important. Having local processing provides quicker feedback and reliability if there is a problem with the communications link, as well as reducing the load on those links. “If you think about billions of bits of data if you are truly looking at systems of systems there is so much data coming out and there’s the delay as well – that determines whether you need to do that in the cloud or just send exception 6 Electronic Engineering Times Europe July/August 2014 www.electronics-eetimes.com


EETE JULAUG 2014
To see the actual publication please follow the link above