The internet of what?

At the launch of Orange Silicon Valley’s new IoT lab last week, Peter Marx, the CTO of Los Angeles, spoke about some of the challenges facing him as a he makes Los Angeles a smart city (one of his first fun observations was that no-one wants to live in a dumb city).
The city as a digital platform is going to drive the future. Even as we become more connected, more and more people are choosing to live in cities. The World Health Organization estimates that worldwide 54% of people live in cities (this up from 34% only a generation ago).
What improves?
One constant in the discussion of smart cities is a focus on end user experiences rather than technology itself. Working that through Marx raised an important question which a large number of IoT implementations will have to deal with. As IoT is inherently passive, and requires additional effort to implement, the questions of what are you going to actually do (switch traffic lights more efficiently, direct airport passengers to the shortest TSA lines, change power sources, manage toll pricing dynamically etc.), and how you measure improvement, become key. It’s much easier to feel excited about technology or new services if you can do something you couldn’t do before (share a photo with a single click, stalk your ex-girlfriend without having to embarrass yourself by talking to anyone else, monitor your health with a wearable and so on). These were novel experiences. For many IoT will incrementally improve things.
Marx commented that if he delivered a 14-17% improvement in traffic congestion, would people notice, and how would they think about it? Traffic is a moderately self regulating system. In gridlocked traffic, people who don’t have to go out in a car, don’t. As soon as a system improves people will go out. It’s entirely possible that a 15% improvement in performance will be reflected in 15% more cars per unit time, but not necessarily an overall reduction in journey time (i.e. more efficient packing of goods/people/things might reduce variability, and improve total system throughput). To be seen as a success the system likely needs to deliver clear individual value.
Autonomous and smart(er) cars being guided through the traffic congestion create another layer of connectivity challenges. You won’t want to get into a new car when you cross the border between the cities of Los Angeles and Santa Monica. This is not so much about connecting things as connecting systems. The water system needs to talk to the power system, traffic has multiple systems (and more in geographies with toll roads and congestion charges). How can you ensure that an autonomous agent works in an optimized way across a route?
The internet of … trees?
It’s not the internet of things, or even the internet of everything, apparently it’s the internet of trees. Los Angeles has a surprisingly large number of trees, and strict rules about replacing them if repair work requires removing them. The city expects that in the next year there will be 200,000 new trees planted, all of which will be tracked with sensors. In comparison Los Angeles has 45,000 connected traffic signals, and another 28,000 pedestrian lights.
How do you connect the dots?
Connecting trees, traffic lights water systems and cars is driving new set of capabilities focused on lower bandwidth communications. Samsung and others have invested in Sigfox, with LoRa alliance presenting itself as an alternative. Of these two, Sigfix appears more focused on smaller messages, with LoRa, based on more traditional cellular technology focused on bigger and bi-directional messages (nice comparison of the two). It seems likely that these systems will be driving smart capabilities at the edge, so for example, traffic lights will be driven by algorithms at the light, fed by just enough overall system information, rather than relying on a central connection for every decision. That said the more interconnected the systems, the higher the chance that centralized capabilities and communications will be important- this will be an interesting area to watch. Data collection might work the same way, in a previous blog in this sponsored series my colleague Jon Collins discussions some of the issues relating to IoT and local data.
It was great to hear from someone on the edge (pun intended) deploying IoT capabilities in the real world. The challenges faced connected multiple interconnected systems in a smart way are becoming clear, and we are starting to see solutions. How big will those trees be by the time every city is smart?

White space broadband gets green light in UK

The British telecoms regulator Ofcom has formally approved the deployment of white-space broadband technology in the U.K., following trials.

White space broadband uses the empty buffer zones that are placed between TV channels to stop them bleeding into one another. The broadband technology glues together these patches of spectrum and, as pilots around the world have shown, it can do so without interfering with the TV transmissions. This is achieved through the use of databases that tell the client device which spectrum it can use in which location and at which time.

Ofcom said on Thursday that it hopes the technology can be deployed in the U.K. by the end of this year.

“This decision helps ensure the U.K. takes a leading role in the development of innovative new wireless technology,” acting Ofcom CEO Steve Unger said in a statement. “It is also an important step in helping the U.K.’s wireless infrastructure evolve effectively and efficiently.”

White space technology may work, but few countries have thus far authorized its use due to concerns over interference. The only commercial deployment I have so far seen was that of a student-oriented network in Ghana, with [company]Microsoft[/company]’s involvement, which went live last month after an on-campus pilot.

The lack of widespread regulatory movement on white space broadband has already forced some in the industry to look to different spectrum for supporting their new internet-of-things networks. That’s a pity, as it works very well for sensor networks, and indeed it’s being tested out in the U.K. for flood defenses and smart city webcams and sensors.

Apart from that, the technology is also good at delivering web access over long distances and into buildings — just like TV broadcasts, funnily enough — and therefore has a lot of potential for both urban and rural broadband provision. As I saw for myself at a [company]Google[/company] trial in my native Cape Town, it’s no fiber competitor but it can make a real difference in areas where fixed-line providers are loath to roll out decent infrastructure.

Smarter cities could start with a simple light bulb swap

Simply swapping out the old-school incandescent, sodium or metal halide lamps with newer LED bulbs that can also contain an array of sensors can push a city on the path to becoming smarter, said Wim Elfrink the Executive Vice President for Industry Solutions and Chief Globalisation Officer with Cisco. According to Elfrink, while cities are installing the LED lamps they often elect to put in video surveillance and even Wi-Fi access points too.

“We see this as an enormous inflection point,” Elfrink told me during an interview on Thursday. “This could blanket a city in Wi-Fi and enable the city to offer citizen services and we are always looking for what will be that inflection point. The simple light bulb could be it.”

He pointed out that having the video surveillance means that computers could count the number of people in the area and reduce or increase the light based on the amount of foot traffic. More people require less light, cutting down on energy. Of course, not every city will be comfortable with video surveillance in all public areas, and Elfrink says most will start with small pilot projects such as one that is being deployed in Chicago.

The Chicago street lamps don’t use cameras to track people, instead they count the number of people by tracking the number of cell phones as they ping asking for Wi-Fi hot spots. They also track temperature and various weather and pollution data via sensors. In Copenhagen, Elfrink estimates that by placing Wi-Fi access points in about 92 percent of the street lamps you could blanket the entire city with Wi-Fi.

The video wall at the DOLL Visitor Center in Copenhagen showing  the live feeds of the streets and data visualizations.

The video wall at the DOLL Visitor Center in Copenhagen showing the live feeds of the streets and data visualizations.

And while the Wi-Fi wouldn’t be part of the actual LED bulb, many of the sensors and maybe the camera could be. In fact, many of the smart bulbs coming out in the consumer and commercial markets double as speakers or sensors of some kind or another. Beacons or ambient light sensors seem to be the most popular in the commercial space.

Such plans also require the city to own or have access to dark fiber so it can offer its own services, but the data it can gather and the potential savings it can realize are substantial. For example, there is the obvious savings from more efficient lighting being turned to the most appropriate level, but adding more sensors means the city can better predict weather patterns and position snowplows in areas where the snow is likely to hit hardest, before it happens.

Elfrink says that every city should have an information and telecommunications technology plan, much like they have an urban and economic development plan, so they can understand where the data from these endeavors should go and who owns it. These plans should also detail how the data should be shared across these city instead of being locked up in a single department. For example, if the lighting and power department claimed this data and didn’t share, it would be much less valuable.

Even as cities start with one-off deployments in smart parking or perhaps a connected street lamp pilot, it’s worth thinking about how to bring these disparate forays in the digital realm into a holistic platform that citizens, governments and developers can access and make use of in ways that protect privacy and ensure security of city assets. That can be an overwhelming project, but as Elfrink says, just start with a single light bulb.

Updated: This post was updated January 26, 2015 to correct the spelling of Wim Elfrink’s name.

IBM to start crunching connected car data for Peugeot

IBM is putting its data analytics to work on information collected from Peugeot’s in-car sensors, ostensibly combining it with data from traffic infrastructure and smartphones to create better car apps and more network-aware vehicles.