KNect365 is part of the Knowledge and Networking Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 3099067.


Why self-driving cars can't start without edge computing

It’s that ‘future’ moment we’ve been patiently waiting for: self-driving cars. However, as AT&T estimates they will generate 3.6 terabytes of data per hour from the clusters of cameras and sensors on board, there is mounting pressure on the finer details of edge computing and how it will power autonomous cars in five years.



Car thieves of tomorrow won’t be following the model championed by Grand Theft Auto. Instead ‘Mr Robot’ style hackers will be operating in the dark web, stealing cars right from their parking spots.

Thinking about these scenarios is how telecoms providers, who will connect cars to the internet, are strategising to capitalise on this new trend. Because if you have to wait for a car to produce a security certificate from the cloud when using your car fob, you’ll be getting soaked in the rain. In those crucial first years of tabloid scepticism, people will think ‘wasn’t it easier when we just had regular ol' cars?”

What we need is real time low latency edge computing to allow for a smoother transition to self-driving cars. Critical calculations must be made close to the car, not in a data farm somewhere in Nevada. I recently used a forward-thinking van rental company when moving house, where I had to call a phone number to unlock the vehicle. While this worked fine when I arrived, taking back the van proved difficult as the automated service said ‘there was an error in locking the van’ and when I tried calling customer service, nobody picked up. The result was I had to leave an unlocked van somewhere in London!

"We need to get the architecture right so it can be exposed in a usable, safe, secure way so that it doesn't compromise the base network functions that we need," says Fuetsch. "I don't think it is a Herculean challenge; it is a big one because this is relatively new. We haven't really exposed this before."



We have adopted some cave man tactics while driving on the road. Ever flashed your lights at someone because their tail lights are out? Do you press the triangle hazard button when there is a sudden traffic jam up ahead? Ever pulled your window down to scream at a motorist: “You have a flat tire mate”?

Autonomous cars are letting us evolve from our stone age behaviour with real time V2V (vehicle to vehicle) communication. Instead of looking puzzled when someone is flashing and shouting at you, you can have a coherent conversation will each other behind your windshields.Platooning_Back_022414_Final_noTpye

This will require high capacity edge networks to free up the strain of a new wave of communication between people. Even the speeds being promised by 5G spectrum will collapse under the strain of data heavy V2V once people start swiping left and right on Tinder Car Edition. App developers will struggle to offload both the backhaul and the core of these networks. There has been a creative explosion in the sensors-drive-economy; lidar sensors, audio analysing microphones and life-saving autonomous emergency braking systems are all products of the IoT revolution.

"Edge computing fulfils the promise of the cloud to transcend the physical constraints of our mobile devices,” says Fuetsch. “The capabilities of tomorrow's 5G are the missing link that will make edge computing possible. And few companies have the sheer number of physical locations that AT&T has that are needed to solve the latency dilemma."



Safety is the number one concern when dealing with 10-ton bricks moving at 70mph. Cars are the most reliably built modern machinery when it comes to safety, and while your Peugeot 106 may break down from time to time, you will rarely see a car spontaneously combust. This was not the case when the largest cloud server farm from Amazon Web Services crashed earlier this year, sending almost 150,000 websites into oblivion.

Imagine the chaos on the roads if cars relied on these server farms when they went down, how might that affect an army of self-driving cars? Today’s wars are fought over the internet, so you can imagine how attractive a server farm is for a hacker compared to individual edge nodes scattered everywhere.

You need a reliable connection that will securely serve data packets to you from the edge of the network. Work being carried out by edge standards firms ETSI and OpenFog are paving a future for a more diversified, resilient environment. Both associations will be presenting phase 1 commitments and an introduction to the core technology at MEC Congress 2017.

“While fog computing is starting to be rolled out in smart cities, connected cars, drones and more, it needs a common, interoperable platform to turbocharge the tremendous opportunity in digital transformation,” said Helder Antunes, chairman of the OpenFog Consortium. “The new OpenFog Reference Architecture is an important giant step in that direction.”

Ryokichi Onishi from Toyota InfoTechnology Center will be presenting ‘Edge Computing Automotive Use Cases and Requirements’ at this years MEC Congress in Berlin.

Screen Shot 2017-07-31 at 11.26.46


Get articles like this by email