Laboratory for Traffic and Transport Engineering: Solving urban congestion with a scalable game

About Laboratory for Traffic and Transport Engineering

The Laboratory for Traffic and Transport Engineering (LICIT) is a French laboratory researching the dynamic modeling and regulation of traffic flows.

Industries: Other
Location: France

About NRX

NRX is part of the IT Link Group that has specialized in Google Cloud technologies since 2005, supporting its clients in their digital transformation.

LICIT powers its traffic simulation platform with Google Cloud Platform, automatically scaling to peaks in simultaneous users with Google Kubernetes Engine and keeping costs down with flexible usage.

Google Cloud Results

  • Scales automatically with Google Kubernetes Engine to accommodate peaks in simultaneous user sessions, enabling large-scale data collection
  • Scales down to zero when not in use, reducing running costs
  • Removes the need for investment in physical hardware, keeping the infrastructure flexible for future projects

Potential to scale up to 10,000 simultaneous users

Urban air pollution is a major threat to human health: according to a 2018 WHO report, "more than 80% of people living in urban areas that monitor air pollution are exposed to air quality levels that exceed the World Health Organization (WHO) limits." Finding ways to improve traffic flow and reduce vehicle pollution is central to improving the quality of life for city dwellers. That's one of the goals of LICIT, the Laboratory for Traffic and Transport Engineering. Based in Lyon, France, LICIT is a joint research unit of the French Institute of Science and Technology for Transport, Spatial Planning, Development and Networks (IFSTTAR) and the National School for State Civil Engineering (ENTPE).

"Gathering enough mobility data is a key issue for LICIT. Being able to access sufficient data is what enables us to evaluate and test traffic control strategies, and to do that we need to be able to accommodate many users simultaneously."

Cécile Becarie, Research Engineer, LICIT

"At LICIT, we specialize in the dynamic modeling of traffic flows and their control," explains Cécile Becarie, a Research Engineer at LICIT. "The way we do that is by building simulation platforms to model how traffic flows, tracking how it changes over a number of hours in a town center, for example. Gathering data across an entire city is very complex, so using simulation tools enables us to build a large data set in order to improve our models."

In order to gather data about user decision-making regarding mobility across a large city, LICIT built a simulation game where players have to make transport choices to complete given tasks. In order to be able to connect many users simultaneously, it needed a powerful infrastructure that also had to be cost-efficient to run. Working with implementation partner NRX, LICIT chose Google Cloud Platform (GCP) for the job.

"Gathering enough mobility data is a key issue for LICIT," says Cécile. "Being able to access sufficient data is what enables us to evaluate and test traffic control strategies, and to do that we need to be able to accommodate many users simultaneously."

Scaling up, scaling down

When building a project that needs to be able to power many simultaneous sessions, but will only be used intermittently, traditional hosted infrastructures can be costly to maintain. "The simulation game we built involves interaction between users, whose trajectories impact on one another. In order for it to be interesting and to collect sufficient data, it has to be able to scale up to accommodate many players," says Cécile. "But between organized sessions where players log on simultaneously, the number of users is zero."

"The high degree of flexibility we can achieve is thanks to Google Kubernetes Engine. When the platform is not being used, we can simply scale down to zero. Google Cloud Platform offered an advantage as Kubernetes was originally designed by Google, even though it is now also open source."

Nicolas Roux, CEO, NRX

To build a scalable platform that was powerful but cost-efficient, LICIT worked with IT Link Group affiliate and cloud experts NRX to build the game architecture on GCP using Compute Engine and Google Kubernetes Engine, with Stackdriver and Cloud Storage. "The initial architecture had been built with on-premises hosted servers in mind," explains Nicolas Roux, the CEO and Founder of NRX. "We conceived a highly flexible infrastructure that will be able to scale seamlessly from 200 simultaneous users up to 1000, or more."

"The high degree of flexibility we can achieve is thanks to Google Kubernetes Engine. When the platform is not being used, we can simply scale down to zero. Google Cloud Platform offered an advantage as Kubernetes was originally designed by Google, even though it is now also open source," says Nicolas.

A long-term partnership

The success of the project has been the result of the long-term partnership between IFSTTAR and the IT Link Group. "IT Link has been working with IFSTTAR for many years, and has an excellent understanding of how their custom-designed simulation software works," says Thomas Greneche, Commercial Director (South) at IT Link. "When Nicolas suggested putting in a bid for the tender, we felt it would be a great fit. Working together, we were able to bring NRX's understanding of cloud-based technologies together with our understanding of what LICIT needed."

"IT Link has also been a Google partner since 2005, so it has been a great collaborative project based on knowledge exchange and long-term partnerships," says Thomas.

Solving new problems

Thanks to the infrastructure built on GCP, LICIT has been able to avoid investing in expensive physical hardware or hosted service subscriptions, and scale its costs according to usage. "I estimate that if we had carried out this project using an on-premises solution, the costs would have been two or three times higher," says Nicolas. "As it is, LICIT only pays for infrastructure usage during its data-collecting sessions."

"We have tested the game with 200 users so far, and can easily scale to 1,000 or more. We could scale to 10,000, although beyond that, we would reach limitations due to our software, but not because of GCP!" says Cécile. "That means we are able to gather sufficient data to produce models to improve traffic issues in large cities."

"This is the first time the laboratory has ever built a simulation game like this. Now we know that we are able to gather the large data sets we need, so we can go even further with more complex models to gain greater insights into traffic behavior and offer better solutions."

Cécile Becarie, Research Engineer, LICIT

Now, LICIT plans to develop more versions of its simulation game to answer other problems and even serve as pedagogical tools. "We are currently speaking to a French city about an online version to teach its inhabitants about ways to decrease traffic congestion. We are also developing a version that enables engineering students at ENTPE to evaluate traffic control measures," says Cécile.

"On the technical side, the next phase could be looking at Machine Learning to use something like TensorFlow to predict user behavior," says Nicolas. "We could also think about adapting the platform to use BigQuery to analyze data collected and visualize it with Google Data Studio."

"This is the first time the laboratory has ever built a simulation game like this," says Cécile. "Now we know that we are able to gather the large data sets we need, so we can go even further with more complex models to gain greater insights into traffic behavior and offer better solutions."

About Laboratory for Traffic and Transport Engineering

The Laboratory for Traffic and Transport Engineering (LICIT) is a French laboratory researching the dynamic modeling and regulation of traffic flows.

Industries: Other
Location: France

About NRX

NRX is part of the IT Link Group that has specialized in Google Cloud technologies since 2005, supporting its clients in their digital transformation.

Google Cloud Platform logo

12 Months FREE TRIAL

Try Kubernetes Engine, BigQuery, and other Cloud Platform products with $300 in free credit and 12 months.

TRY IT FREE
Google Cloud Platform logo

12 Months FREE TRIAL

Try Kubernetes Engine, BigQuery, and other Cloud Platform products with $300 in free credit and 12 months.

TRY IT FREE