Natalie Sauber, market intelligence lead at Arcadis, explores how autonomous and connected vehicles are driving a new generation of data centre requirements.
When it comes to the future of transport, Connected and Autonomous Vehicles (CAVs) are among some of the most disruptive technologies we’re seeing today. This gradual shift towards driverless vehicles brings many potential benefits, not least of which in helping to increase road safety, reduce traffic congestion and pollution, minimise fuel consumption and greatly improve the passenger experience.
The huge amount of investment from major automotive manufacturers coupled with new entrants from the technology sector brings the potential for truly transformative change in the way that people and goods are transported.
By 2025, the industry predicts that CAVs are estimated to generate over $200 billion in annual global revenue for auto manufacturers and others. However, a less obvious consideration is that these vehicles will also transmit over 100 petabytes of data to the cloud each month.
To put this in perspective, a petabyte is just over one million gigabytes (the equivalent of digitising approximately half of the printed material ever published). Furthermore, by 2050, the data volume between vehicles and the cloud will require 10 exabytes per month, approximately 10,000 times larger than the present volume. One exabyte is around a thousand petabytes, and equivalent to 250 million DVDs of data.
Clearly, the transfer and management of this volume of data will create unprecedented challenges. But, for every challenge there is an equal opportunity, and data centre providers must prepare themselves for this growth, both in the cloud and at the edge.
Connecting at the edge
Connected and autonomous vehicles will soon communicate in real-time between each other (V2V) and vehicle-to-everything (V2X) for enhanced safety and convenience. For CAVs to operate safely and efficiently, reliable ultra-low latency vehicle-to-infrastructure (V2I) communications is critical.
As it stands today, the 4G cellular network can facilitate management of certain tasks back to the CAV, such as weather and road conditions and real time information on accidents. However, providing high definition maps and identifying obstacles and objects moving near the vehicle requires constant reliable processing with low latency.
Far better is the 5G network currently being rolled out, which processes data up to 1,000 times faster than the existing 4G networks. The 5G cellular base station network will support connectivity and control for the CAVs, as the V2I wireless system will be capable of managing all the different tasks required by the AV, while at the same time operating under constant network connectivity.
Impact on data centre infrastructure and management
The transfer into hyperscale cloud facilities for big data analytics and management precedes a fully integrated (V2X) eco-system. However, delivering that data to the data centre is challenging, with estimates suggesting that approximately $1 million needs to be spent on data centre infrastructure for each AV.
If we believe the predictions that 10 million self-driving cars will hit the road by 2020, data centres are going to require immense investment in computer tech, networking, storage and other infrastructure. We seem to be at a crossroad, as there is a dire need for new approaches on data centres to meet the demands of the new mobility value chain, from data ingest to deep learning and simulation capabilities.
We know that data centres continue to see huge levels of investment across the world. The Hyperscalers and cloud providers are bringing more capacity online at an unprecedented rate, pushing the boundaries of lean construction methods.
Colocation providers are developing ever more sophisticated systems to allow their customers to manage data generated traditionally in servers, as well as that created at the edge. There is also a need to carry out analytics in real time, as well as the big data crunching that is carried out in the cloud.
Consequently, the use of Artificial Intelligence (AI) will shape a new generation of data centres equipped with new hardware and software architectures enabling large-scale, real-time analytics and machine intelligence. Soon, computer and storage will be optimised by introduction of edge and quantum computing to work seamlessly as an extension of the overall IT system. The is a definite need to increase network capacity between vehicles and the cloud by implementing edge computing and more efficient network designs.
Dedicated data centres or research institutes for CAVs are scarce. Only recently did the UK see a significant financial investment made towards elevating this to the next level. The West Midlands Combined Authority (WMCA) will receive up to £20 million to allow Warwick Manufacturing Group (WMG) to create the UK Mobility Data Institute; a focused research centre to collect, process and analyse transport data generated by autonomous vehicles, along with smart charging of electrified vehicles.
OEMs have taken control and ownership of their data centres
A number of car manufacturers have also pledged to build data centres over the coming years. In 2016, Toyota announced plans to build its Data Centre for Connected-Car Data. Meanwhile BMW is planning to use IBM's cloud infrastructure to deploy its new platform for collecting data from connected cars. In 2017, Ford announced plans to build a $200m data facility centre to store connected car data in Detroit. The American manufacturer predicts that their data storage requirements will rise from today’s 13PB to an astounding 200PB by 2021.
The recent increase in data centre construction activity is clearly accelerated by the industry readying itself for these data consumption increases, and the draw on construction resources is going to increase in line with demand.
Future proofing global supply chains is going to be an essential step in meeting these requirements, as is careful site selection due diligence, which enables access to the best site locations and the critical infrastructure necessary for their future data centre needs. By effectively bridging the gap between data centre providers and city policy makers, can we improve network implementation times and significantly make a difference to the industry.