The level of video surveillance across 41 venues and their environs at the 2024 Paris Olympic Games is predicted to be one of the records broken at the 2024 Games, far outstripping past Olympics.
The French government sees it as a necessity to prevent terrorist attacks, but the extent of the planned video surveillance coverage has already been causing controversy. Privacy campaigners have expressed concerns about how the data collected will be used and AI-driven analysis managed.
Many are worried the plans will set a precedent for other events, with Germany potentially to follow suit at Euro 2024.
Taking the stance that analysis of video footage will help to protect millions of attendees, the government has tried to quell fears by setting out what information will be collected and how it will be processed. It has been at pains to point out that facial recognition is not being applied to the footage.
The AI-powered surveillance algorithms use computer software to analyze images captured by the cameras in real time and are trained only to recognize and send alerts when noting pre-defined unusual behavior or objects.
This includes abnormal crowd movement and density, weapons and abandoned objects, fire breakouts, and any contravention of traffic direction rules. When the system detects suspicious activity, it is scrutinized by humans before decisions are made about taking further action.
Raising the Bar for Performance and Resilience
To gather this intelligence, hundreds of cameras are going to generate vast amounts of unstructured video data to store and analyze in real time. Critically, the surveillance system will need to be highly resilient to ensure that if any application or hardware fails, there’s immediate failover so continuity is maintained.
Losing any footage could have significant security implications. Crucial moments of suspicious activity might be missed, impeding efforts to recognize potential threats and prevent attacks, or stop dangerous situations from escalating. Timely investigation of incidents or crimes could be jeopardized or prolonged without video evidence to reconstruct events, seek out perpetrators and corroborate eyewitness accounts.
If uninterrupted video recording isn’t available to verify what’s happening on the ground, there’s a bigger risk of false accusations against innocent bystanders. And losing vital sections of footage wouldn’t go down well with the public, heightening concerns about reliability and transparency.
The extent of the impact would depend on the criticality and length of what’s missing, but any loss could create security gaps and complicate real-time analysis and response. Not a position the French government would relish with millions of spectators and global media watching.
Unstructured Data is No Longer a Hurdle
Handling this massive volume of unstructured information efficiently and securely requires the latest video surveillance recording and storage technology, capable of scaling to support thousands of cameras, with the highest levels of uptime, and with the ability to support and run multiple surveillance applications, like advanced crowd intelligence analytics.
Legacy network video recorders (NVRs) were not built to support video surveillance operations at scale, as each application on the system, like analytics, access control and video management, would need a dedicated server to operate.
"Losing any footage could have significant security implications"
Furthermore, standard NVRs, application servers, and storage systems are not designed for always-on video surveillance environments. While NVRs can survive one or two hard disk failures, they cannot protect against multi-faceted failure scenarios.
This includes problems such as multi-disk failures, disk controller issues, motherboard failures, system power supply outages, chassis damage or network switch failures. Any combination of these could result in significant downtime and breaks in video recording.
Once a tough hurdle to overcome, this is now feasible with a new breed of surveillance platforms. New solutions are replacing traditional racks of NVR servers with a single compute and storage software-based platform for video recording and advanced analytics. Customers can easily host more than 10,000 cameras on a single system.
Plus, they offer robust data protection and automatic failover across an entire cluster to maximize uptime, virtually eliminating data loss or downtime caused by a failed component or server. In a mission-critical scenario like the Paris Olympics, organizers cannot afford to stop recording or lose access to previously recorded video.
While there may be application failover available with some older video management system applications to continue recording, they will still lose access to the previously recorded video if the NVR fails. There won't be enough time during an event like this to retrospectively bring up the failed system and try to access the old video. This continuous access to recorded video is only available in a new age software-defined distributed storage architecture.
With the added advantage of running multiple physical security applications on a single, unified infrastructure, a cluster can include compute and GPUs for running AI workloads, as well as flash or disk storage, depending on specific requirements.
Effectively, organizations and authorities requiring mass surveillance for large-scale events like the Olympics require a system built for mission-critical video surveillance, whereby they benefit from an easy-to-manage, cost-effective system that can accommodate the massive growth in video data and AI analytics.
Security and Privacy on the Same Team
Critics are concerned that such comprehensive surveillance in Paris will be the thin end of the wedge, arguing that facial recognition and further AI-powered analysis could easily be added into the system at a later point – and used for more than just sporting events. They warn that any individuals or groups at rallies could be identified and monitored, comprising their right to privacy and free speech.
However, used responsibly, these new solutions will provide vital safety and security information – with the ability to pinpoint and review only the footage where an incident may or has occurred – to protect all types of large gatherings.
Additionally, they can assimilate intelligence insights to enhance the customer experience at major events, for example, directing attendees to avoid queues for entrances, exits, tickets, concession stands and other facilities.
In the meantime, it remains to be seen whether there are repercussions from surveillance at the Paris Olympics or whether most people are happy to accept some privacy trade-offs in the interests of keeping everyone safe. Recognizing that security and privacy must go hand in hand will need public approbation to help authorities strike a balance that works for the majority of society.
Image credit: Svet foto / Shutterstock.com