Bridgeworks features in this article from TaaS Magazine (pg 56) to discuss the growing influence of data management for connected, autonomous and electric vehicles.
May 11, 2021
There are many predictions about connected and autonomous vehicles, some of them suggesting that fully autonomous, levels 4 and 5 vehicles will begin to become commonplace on public roads from 2025.
A study by Vynz Research says the global connected and autonomous vehicle market size was 17.7 million units in 2019; and it predicts that this will reach 51.2 million units by 2025 – a compound growth rate of 17.1% during the period of 2020 to 2025.At present, most vehicles aren’t fully autonomous, yet still increasingly rely upon data to operate.
Growing data volumes
With their emergence will be a growth in data. Rich Miller writes in his article for Data Center Frontier, ‘Rolling Zettabytes: Quantifying the Data Impact of Connected Cars’: “The Automotive Edge Computing Consortium (AECC) is working to help stakeholders understand the infrastructure requirements for connected cars. At Edge Computing World, AECC board member, Vish Nandlall, outlined the group’s findings on the volume of data created by autonomous cars and the challenges they will create.”
Vish Nandlall, VP at Dell Technologies and board member of the Automotive Edge Computing Consortium, said in his talk, ‘Driving the Zettabyte Edge’ at Edge Computing World in December 2019:
“Automotive data volume will drive the edge and we’re going to hit zettascale volumes.” before adding: “We’re really starting to challenge the limits of the cloud technologies we’ve been using. It’s a challenge to the infrastructure and cloud communities, and a challenge to the automotive community.”
The AECC also said that “data traffic from autonomous vehicles could surpass 10 exabytes per month by 2025 – about 1,000 times the present volume.” Such are the amounts of data involved with the future of connected and autonomous vehicles that Nandlall believes vehicle manufacturers will have to become hyperscale computing companies with a single vendor operating at a zettascale. In other words, they will become landlords of amounts of data real estate.
CAVs and data
Data will need to be analysed and, to some extent, stored within connected and autonomous vehicles – particular when 5G coverage is poor. Data will have to be transmitted to and from the vehicles, too.This data will need to be offloaded and transmitted to data centres for backing up and storage, as well as to enable big data analysis.
While much of the data relating to driving will be carried out at the edge, close to the sensor on each vehicle, there will be a need to store it somewhere other than only in the vehicle. This is to allow historical data to be stored efficiently, which will be as crucial to autonomous vehicles as any other kind of data – including for insurance and vehicle maintenance purposes.
Autonomous vehicles will also need to connect with everything in the world around, including road infrastructure and other vehicles. These vehicles will need to use artificial intelligence and machine learning to obey road rules, respond to traffic signals and learn from common and uncommon (unexpected) situations to avoid accidents.
On top of the driving and vehicle data, there will be more data to and from the vehicles for media, entertainment and for big data analysis. This data analysis will be needed to enable machine learning, allow the vehicle to adapt to new situations; Permit software developers and engineers to provide patches and upgrades and for manufacturers and their ecosystem partners to improve their vehicle’s overall performance.
With the steering wheel being handed over the vehicle itself, with no human driver, there will be time for its passengers to enjoy music and streaming services, play games or shop while on the go. Big data will therefore be used to analyse their behavioural patterns and to make suggestions based on their preferences, location, browsing and shopping habits.
Addressing data ownership
With the data there will come legal challenges, which means there will need to be an eye on regulatory compliance, too. Dr. Stephan Appt, LL.M. Partner, Head of German TMT at Pinsent Masons, writes in his article ‘Legal challenges of data-driven mobility’:
“Data is central to the way new connected, autonomous and electric vehicles operate. This requires automotive businesses to address issues such as data ownership, data access and data sharing, as well as the protection of personal data, at the outset of vehicle development.”
“Sound data governance, beginning at the earliest stage of development, can help original equipment manufacturers (OEM) and their suppliers address the legal challenges that arise around data privacy and security, as well as broader data-related issues of availability, accessibility, consistency, integrity and efficient data and risk management.”
Appt adds that data trusts have the potential to improve data sharing within the automotive sector, despite the fact that stiffer data protection regulations are anticipated. He thinks that automotive manufacturers should explore such options as data trusts and believes they should understand how the regulation of data within the sector is likely to evolve.
SD-WANs and WAN Acceleration
With data growth, and with it the need to improve data management – including to prevent cyber-attacks from causing significant operational and financial damage to organisations – SD-WANs have become a popular choice. David Trossell, CEO and CTO of Bridgeworks says they are a great technology, but with them latency and packet loss can still dog them. Increasing bandwidth won’t resolve latency and packet loss issues, either.
To boost the data transfer, egress and ingress capabilities of SD-WANs, he argues that they can benefit from having a WAN Acceleration overlay, and while the chosen connectivity for connected and autonomous vehicles is likely to be 5G, the data will at some point go over a wide area network.
Trossell therefore provides his top 5 tips for best practice data management, relating to connected and autonomous vehicles – particularly relating to the transport, storage, regulatory compliance, and data analysis over WANs:
- With the projected data volumes, data as it is collected at the edge will have to undergo immediate prioritisation to separate the data that requires immediate response from that ofnon-urgent auxiliary data.
- Holding data at the edge for longer than required could overload the edge storage capacity. The quicker this can be dispatched to the central Data Ocean the better.
- Moving this data over large distance will affect data throughput unless tools are used to mitigate Latency and Packet Loss.
- With Global car manufacturing and the European GDPRphilosophy of data protectionbeginning to roll out across other countries, therewith be a need to encrypt this data whilst in flight and storage – tradition data deduplication products will struggle move this type of data over distances.
- With all the differing types ofdata being created some will have a shore life cycle,whilst other data type such as GDPR datawill require storage for possibly a number of years and has to be backed up for regulatory requirements.
The data management challenges are never-ending, and they will continue with the end of the Brexit transition period at the end of 2020. No matter what your views are on Britain leaving the European Union, connected and autonomous vehicle manufacturers (as well as their ecosystem partners) will be keeping a close eye on any legislative and regulatory changes that occur in the U.K.
Evolving data models
The same applies to elsewhere in the world, including in the EU, because data protection has to evolve with new situations and with new technologies. This includes the development of new insurance business models. At present, when someone drives a car and is involved in an accident that they caused, they are liable. Without a human driver, this will change and may involve OEMs being held accountable for the failure of a connected and autonomous vehicle to avoid an accident. So, data will be at the heart of how they operate, and at the core of how new insurance polices are developed.
With the prospect of electric cars being pushed forward a decade by the UK government, with a ban being introduced on new petrol and diesel engine vehicles in 2030, there are also calls for new ways of charging vehicle owners tax. At present, tax is charged on petrol and diesel sales, but once petrol and diesel cars are taken off the road, there will be a need for a new model. This could involve implementing systems into the cars to permit charging owners on a per mile travelled basis. Data will again play a key role to determine how much people will need to pay for their journeys.
However, the need to ensure that the Treasury’s coffers remain healthy, has to be balanced with the ambition to encourage more people to buy electric vehicles. As autonomous vehicles are also likely to be electrified, despite the demands that their automated systems create. As many people are still wary of the idea of being driven by a computer, the Government, as well as industry, will need to be careful how they implement new pricing models. They will need to be enticing and not punitive. However connected and autonomous vehicles develop, good data management, data security and data protection will also be imperative. Without which, consumers might not jump aboard.