From Cloud To Edge: The Smart City And Personal Data

November 09, 2018

The Smart City is as much a vision as it is a range of products and services where sensors and data play centre stage. In the current paradigm, sensors collect data throughout the city – at the edge – and transmit these to a server anywhere on Earth – the cloud – for processing, sharing and delivery of smart services. Could the edge not take a more prominent role than observing citizens and the city?

Whether a service is cloud-based or not can make a big difference. Just imagine Amazon's Alexa working offline and not storing on Amazon's cloud what you tell it – or alternatively, the iPhone X face recognition needing access to iCloud. Cloud or edge: for the engineer, it is a question of speed, latency and cost; for the citizen, it is also a matter of data ownership and privacy.

City authorities collect and store a wealth of data and increasingly make it available to everybody, possibly in real time, as anonymous open data. These platforms are believed to foster new, innovative services and to stimulate local entrepreneurship. Cities like Lyon and Amsterdam have gone a long way in this direction, and they showcase a wealth of interesting applications. However: very few business models exist for scaling these new services citywide, and they often rely on subsidies. Subsidies are also needed for sensor deployment and data processing. Take air quality monitoring, for example, where low-cost sensors have entered the market. Thousands of new low-cost sensors still cost a lot of money, and most city budgets are too stretched to fund a citywide deployment.

Why this lack of viable business models, even if no new hardware is needed? Anonymous, open data is useful throughout a city's administration, and it has a lot of indirect benefits for citizens. In my view, however, it cannot generate a direct Return on Investment unless you use it together with personal data. This is why Google is mining cities' open data, to merge it with personal data from the search engine and geo-localisation …

City authorities store a lot of personal data too, in the frame of public services. Generally – and fortunately! – they do not re-use it, not least because citizens assume that personal data collected by city authorities will not be used outside of their initial scope, even if the same citizens easily accept that search engines and social media do exactly this.

Let me add that city public services are normally behind a pay wall of either fees or subsidies, which is different from the Internet model of search engines and social media. Could this model of free services against giving away personal data replace the established model of public services behind a pay wall?

I would like to illustrate why this is not the case, and the consequences, using the example of the self-driving car, or CAV for "connected autonomous vehicle". The CAV is actually not autonomous at all but relies on ultrafast communication with other vehicles and an outside infrastructure of intelligent sensors, an infrastructure that does not exist today. Nobody can tell today who will be in charge in cities and how revenues will be generated and shared, even if the CAV tests underway in Phoenix and Pittsburgh are 100 percent privately funded. This free market model must not be scaled across large cities because mobility as a service is too important for any city to abandon it to private initiative alone. City authorities must be part of it, and not just passengers.

Being able to regulate CAVs is actually a political battle cities are currently fighting with legislators against the free market and data city lobbies. Winning it matters not least to avoid the superposition of multiple smart city IoT infrastructures: one for CAVs, another one for waste collection, yet another one for air pollution and so forth. In the next ten years, smart driving will probably become the backbone of the smart city.

In Grenoble, for example, we will take back all car parks into city administration. The private operators in charge today maximise income instead of using them to reduce traffic congestion and encourage citizens to shop downtown, not on the Internet. Next year, our car parks will be part of Grenoble's offer of mobility as a service. And at the right moment, CAVs will be included into this offer, not as luxury goods for the happy few, but as a shared ride-hailing service fully integrated into a global offer where you pay for a ride whether you use a shared bike, the tram, a CAV – or all three of them.

If an infrastructure or a service is publicly owned or under public supervision, the same naturally applies to the data it generates. Where private investment exists, city authorities should not hesitate to regulate data ownership. Good regulation actually attracts investment as it creates trust. Finally, city authorities should prefer a pay wall – either by use charges or subsidies or a mix of both – to the Internet free service model. A service must be worth its money, for the citizens or the city's budget, including for data-driven services.

Today, cities store personal data mostly in silos, typically on an edge server. A smart city will open these silos of personal data if a citizen retains the right to decide each time whether he or she agrees, or not, to use of his or her data outside the silo. This approach is called "self-data", and cities like Lyon are experimenting with it since several years. Its implementation ideally involves Apps and smartphones because it's practical and safe – very much like payments on the go. And it's ideally suited to the crowd sourcing of data from sensors at home or in citizens' mobile devices.

A big obstacle is that all commercial vendors of smart city solutions try to lure their city customers into closed vertical systems with a cloud on the top, and sensors and actuators at the bottom. The communication protocols, the application interfaces and the middleware might be standardised or even open-source, the different systems remain isolated from each other because the battle is about access to the data in the cloud. Despite a lot of talk, truly integrated citywide systems do not take off as quickly as, for example, similar platforms for industry 4.0. Market growth for IoT is real but mostly concerns renewal or optimisation of existing systems, for example traffic or energy management.

The few existing integrated platforms also often are pilots or limited to a few districts, because citywide systems would be more expensive to set up and operate than any expected cost reductions. In a nutshell: there is a lot of smart city technology push, but the market pull still has to happen.

How to move forward from this apparent deadlock? Let me recall that Europe has lost the battle to connect people. The Internet, the smart phone and the cloud are all in the hands of Asian players and US. The next battle is to connect not people but 50 billion devices. Europe has two assets to win this battle: its industries in sensors and edge data processing, which control 50 percent of the world market and its cities, which are world leaders in governance, public investment and citizen empowerment.

With this Eurocentric view in mind, I see five guidelines how to move forward:

  1. A smart city should be strong when it comes to policies, objectives and budgets. No wonder innovation leaders like Copenhagen, Barcelona or Amsterdam regulate the impact of AirBnB and Uber on their inhabitants and city centres. Transparent regulations and clear policies attract sustainable long-term investment. Global, or European standards are equally important and cannot be replaced by corporate technologies. Long-term infrastructure investment by cities is impossible without such standards upon which everybody, including private investors, can rely in the long term.
  2. The Smart City is less about technology than about citizens who want a better quality of life in a heavily constrained environment, citizens who do not want to be left behind or excluded, citizens who wish their privacy to be respected. This is why smart city projects in Grenoble focus on real-life problems like traffic congestion and air pollution. They should have a scalable business model and a positive impact on many, if not all citizens. Don't waste your time to foster solutions without a problem.
  3. Merging open and personal data is often useful, but it is the citizen who decides. This means that personal data – electricity meter readings, geo-localisation, traffic recordings, library loans and so forth – and the control over it must remain in the city, close to the sensor. Implementation of self-data in Grenoble will involve two elements: a universal data repository for every citizen, which for security reasons, is server-based and supervised by the city administration. Second, a smartphone App allowing citizens to permit access to this data in an easy and safe way.
  4. Experiments and pilot projects are needed but the proof of the pudding is the transition to mass use, which will often involve a public service, a pay wall and/or subsidies. If scale-up and a business model are not built into the pilot, it's pointless to start an experiment. In Grenoble, we will use mobility-as-a-service to roll out our self-data implementation. Our global mobility platform will store personal data in a new universal data repository. From the first day, close to 40 percent of all citizens will be connected and able to liberate all of their personal data for services they are interested in.
  5. Sensors throughout the city matter but even more important is that they are intelligent and able to communicate directly with citizens. Artificial intelligence in the edge is a hot research topic – do not leave machine learning to the cloud! Communication also matters because today's smart city networks and protocols are incompatible with mobile phones and home WiFi. 5G networks are a huge opportunity to change this, with China and Huawei clearly moving into this direction. Imagine smart traffic lights or parking sensors directly interacting with cars and citizens within, say, 200 metres! Citizens could likewise agree to exchange their personal data only within their district. This is a far-flung vision but so was the Internet when it was imagined in Europe in the early 1990s.

Claus Habfast is a municipal councillor for the City of Grenoble and Vice President of the Greater Grenoble City Area.

Overlay Init

Curated By Logo