Goodbye gridlock: how to keep traffic moving with continuous data replication

Anyone travelling through a major conurbation at rush hour will know the misery caused by gridlock. It’s bad for blood pressure, for business, and for the environment. So a priority in any Smart Cities project has to be managing traffic flow more intelligently. Live feeds from connected sensors positioned across the urban sprawl, combined with real-time data from vehicles and transportation systems, and historic data about traffic flow and accidents, offer the best chance of beating bottlenecks.

The opportunity hasn’t been lost on local authorities. Cities from Manchester in the UK  and Columbus (Ohio) in the US , to Taipei City in Taiwan  and Melbourne in Australia , are already running trials to connect transportation systems. A flurry of upcoming Smart Cities events will no doubt showcase more of the latest developments.  On top of the obvious benefits of reduced congestion and delays, there is potential for cities and partners to develop and monetize new services based on the data being collected, leading to further improvements in traffic management. 

Underlying this great promise, however, is a massive assumption: that the wealth of data being collected can be processed efficiently, reliably and continuously – and by more than one party, in more than one place. 

A problem shared

Kansas City in Missouri, US, sees live data as the key to provide innovative new services . As part of a broader smart transportation initiative, it is installing cameras on lampposts to help understand traffic flows, optimize traffic signals, and allow streetcar operators to anticipate road conditions. Video feeds will track weather conditions to target snowplow deployment, and sensors embedded in the street will share information about available parking spaces via a smartphone app, so residents waste less time shunting around the streets in a holding pattern .

Four years ago, Kansas City opened up its data to third parties via an online portal, so they could experiment with it. The UK runs a similar initiative called oneTransport, an open marketplace for data  that any city can exploit to accelerate progress. Its proponents claim that more than £1bn ($1.3bn) of value could be unlocked over the next five years in the UK through the sharing of connected city data.

But there are other practical considerations for services that depend on analysis of live data feeds. A recent research paper, exploring the issues associated with Big Data in cloud computing in support of smart-city innovation , notes that “Advanced algorithms need to be optimized to handle high volumes of data, a large variety of data types, time constraints around the decision-making process, and distributed components across various geographical locations.”

With traffic systems generating hundreds of petabytes of data per hour  – data which is being updated continuously – this starts to look impossible. But is it? It certainly ought to be because of the almost unimaginable throughput of data and the need for it to be in more than one physical place simultaneously. This could be a primary data center, or cloud platform, where the data feeds arrive, plus any number of other cloud locations where the information is being combined and compared with other data sources, and detailed analysis is taking place.

Amazon lists the many ways its web services can help with smart-city data analysis at speed , but what it does not do is address the need for live data to exist in more than one place at once. The point being that high-volume movement of real-time data between locations cannot be assumed as a given. It is extremely complex to achieve.

From pain comes gain

Once cities decide to harness data to improve transportation flows, the benefits are manifold, as the city of Eindhoven in the Netherlands found in a pilot traffic-management project with IBM . The project collected braking, acceleration and location data from in-vehicle sensors, and merged this with traffic data gathered from the road for combined analytics. As a result, officials were able to respond to dangerous road conditions, accidents or growing traffic density in near real-time. Officials were also able to alert drivers to traffic incidents via smartphones and built-in navigation devices, giving them a chance to choose an alternative route. 

Toulouse, in France, is testing social-media comments about traffic issues as a means of targeting where road maintenance is needed. As a result, potholes and other problems are now fixed within a day, compared to two weeks previously, allowing more roads to remain open.

Of course, driverless vehicles are likely to redefine the paradigm again  and place even more emphasis on the need for live data to be in sync across multiple processing locations. 

The challenge may seem overwhelming, but it isn’t. As long as the data, like the traffic it’s designed to support, can merge and advance uninterrupted, cities will be able to build smarter transportation systems for everyone.