Memoori: The major trends shaping the physical security as we move into 2023

Memoori Research Senior IoT and Security Analyst Owen Kell shares his insights on the trends that will shape the physical security industry in the year ahead.
By: Owen Kell, Senior IoT & Security Analyst, Memoori Research
Ongoing geopolitical & trade tensions threaten to put the brakes on Chinese expansion
China’s leading video surveillance giants have become increasingly dependent on foreign markets. In terms of global revenues, giants Hikvision and Dahua Technology have retained their dominance through 2022, reporting strong overseas sales growth in H1 2022, and generating revenues of over $16.7 billion in 2021.
Their future growth is, however, threatened by a combination of weak domestic demand (largely due to the knock-on effects of the Chinese government’s increasingly controversial zero-Covid approach and resultant regular lockdowns of major cities), ongoing trade tensions between China and the West, and a weakening global economic outlook.
Tensions over alleged human rights violations and the ongoing trade war between China and the U.S. has seen continued legislative moves, new sanctions, and tit-for-tat trade barriers erected that have hugely disrupted the flow of both physical security products and key product components critical to ongoing innovation between the two nations, as well as the ability of their respective manufacturers to trade in their respective markets.
Furthermore, the recently announced US government export restrictions imposed on leading chip designers NVIDIA and AMD will stymie China’s ability to cost-effectively carry out the kind of advanced computing required to remain competitive in the fields of computer vision and natural language processing AI.
Restrictions on Chinese physical products sales to particular market verticals (or indeed a total bar on sales of all of a particular company’s technology) continue to increase not only in the U.S., but also in other Western nations, and it seems that the geopolitical and trade tensions are set to get worse before they get better.
Chinese chipmakers are not yet capable of replicating the performance of these advanced NVIDIA and AMD chips, so as a result, Chinese AI researchers may be forced to revert to using multiple lower-end chips to replicate the processing power.
AI adoption goes mainstream
After several years of false dawns and over-hype, through a series of incremental steps, AI-enabled solutions have become increasingly commercially viable in the physical security space. The pace of change over the past 12 months has been particularly remarkable, with new research papers that push the boundaries of what is possible being released on virtually a weekly basis.
We have observed significant improvements in terms of the speed, accuracy and cost of machine-learning solutions for practical applications in security-related fields including complex facial recognition, cutting-edge video surveillance scene processing, audio analytics and robotics/drones, to the extent that leading AI algorithms in several areas now far exceed human capabilities for several use cases.
While the most advanced AI relies on significant computational power and processing capability, a combination of improved edge processing capabilities at the device level and flexible access to increasingly affordable cloud-based computing resources now make accessing these innovations a viable option for many firms.
The race is now on between vendors to integrate next generation AI-enabled security functionality into end-user focused applications in a way that facilitates accessibility, practicality and ease of use for everyday security scenarios. AI is increasingly replacing the manual effort required for some roles, with data interpretation that formerly required human input now being handled algorithmically, automating processes or steps to make managing security that much easier.
For security staff, this will lead to a reduction in the amount of time spent monitoring screens or watching out for alarm notifications, and more time spent conducting higher-value work. AI will be leveraged to help analyze, evaluate and prioritize data feeds, and then provide real-time prioritization and recommendations on security issues requiring the attention of the security staff.
Making this kind of AI functionality accessible to security professionals will also require integration of AI/ML tools into existing software and platforms in a way that minimizes the need for coding or technical expertise to operate, as well as providing accessibility through multiple media including handheld devices.
While concerns over the ethical usage of AI and algorithm biases will persist, the increasing levels of integration of AI tech into all manner of everyday services and solutions (particularly into digital media generation) will lead to increasingly widespread acceptance of the technology in society as a whole.
An increasingly cyber-conscious customer base
Smart buildings are experiencing an explosion in the volume of IoT devices being deployed, as well as ever-increasing levels of convergence between IT and OT networks. These factors, combined with the growing sophistication of malicious actors and increased reliance on cloud services mean that smart buildings are becoming increasingly vulnerable to cyberattacks.
The last 18 months have seen a huge rise in ransomware attacks, as well as rising costs per incident of cyberattacks on businesses around the world[1]. All too often, IoT-enabled building automation or physical security systems have acted as the “soft underbelly” of organizational cyber defense, with multiple high-profile cases of security breaches serving to highlight the risk and vulnerabilities posed.
At the customer level, awareness of cyber risk, cost implications and even the adverse effect of cyber risks on building owners/operators’ ability to effectively insure their buildings, is steadily increasing. We have observed increasingly strict cyber policies adoption among more sophisticated clients over the past year to mitigate these growing threats, with many companies adopting a “Zero Trust” approach to network architectures, as well as continuous verification as well as investments into systems hardening.
Here again, AI will also play an increasing role – both in cyberattacks and cyber defense.
Memoori’s recent report into AI & Machine Learning in Smart Buildings found that AI tools and building blocks for launching an offensive AI-driven cyberattack have already been developed by bad actors, with several incidents identified by researchers indicating that AI had been used to execute attacks faster or to gain deeper access into a system. In terms of protecting against the threat, AI is increasingly being deployed to provide cyber-risk analytics for improving organizational resilience and understanding cyber risk by improving threat intelligence, prediction, and protection as well as enabling faster attack detection and reducing the need for human cybersecurity experts.
As demonstrable cybersecurity capability moves from being a “nice to have” to a “must have” in the eyes of an increasingly cyber-conscious customer base, security solution vendors seeking to differentiate themselves will need to invest in areas including security by design, gaining cyber certification, demonstrating cyber standards compliance, and independent testing and validation of their products to help differentiate their offerings.
Other trends to watch out for in 2023
Other trends we’re actively monitoring that we believe will continue to significantly impact the market for Physical Security include:
- Ongoing global supply chain woes, which look set to continue well into 2023, impacting stock levels, price inflation and component availability.
- Continued blurring of the lines between cloud and on-prem physical deployments as increasing numbers of end users embrace hybrid deployment models for at least part of their security solution.
- Rising demand for security systems integration and interoperability for better reporting and control of other building/business functions for applications including occupancy analytics, energy efficiency and improving the employees or tenant experience.
The ideas presented in this article draw on the findings of several recent Memoori’s research reports into cybersecurity for smart buildings, IoT in smart buildings, and AI applications for smart buildings, and will form part of Memoori’s forthcoming annual research into the Physical Security market, due for publication in December 2022.
[1] https://www.ibm.com/security/data-breach
Novaira Insights: Video surveillance market growth continues; price has increased, too.
Video surveillance is a key element in security. So how has the video surveillance market fared recently, and how is it expected to perform in the years to come? asmag.com spoke with Josh Woodhouse, Lead analyst and Founder, and Jon Cropley, Principal analyst, of Novaira Insights, which recently released a video surveillance market report.
By: William Pao, Senior Reporter
According to the report, “The world market for video surveillance hardware and software,” the global video surveillance market grew 16.4 percent in 2021. An easing of restrictions on movement and efforts to meet pent-up demand post-COVID were cited as some of the key growth drivers. Indeed, amid project resumptions and construction booms in various parts of the world, a continuation of growth in the video surveillance market is all but expected.
Price hikes resulting from supply shortages
However, the report cited the global average price of a network camera increased by over 7 percent last year, making 2021 the first year in which the global average price of a network camera increased rather than decreased. A main reason cited by the report was a shortage of components used for production of video surveillance equipment, resulting in higher prices for those components; this then forced video surveillance equipment vendors who were unable to absorb such cost increases to raise the prices of their own equipment.
Especially, vendors were faced with a shortage in semiconductors, which are the basis for image signal processors and SoCs that are key components in IP cameras.
CHIP Act may help somewhat, but not in short term
“A shortage of semiconductors was particularly problematic. However, there was also shortages of resistors and materials such as some plastics and metals,” Woodhouse and Cropley said, adding that the CHIP Act recently signed into law in the United States may improve the chip shortage situation somewhat, but not anytime soon.
“It is likely that the CHIP Act will lead to greater production of semiconductors in the U.S. in the longer term. However, it will take years for new production facilities to be built and for volumes to ramp up. The impact of the CHIP Act is therefore likely to only start being felt towards the end of the forecasts in our report (our forecasts run to 2026),” Woodhouse and Cropley said.
According to both, general inflationary pressures will force vendors to increase prices yet further in 2022 and 2023. This, then, is expected to produce an impact in the video surveillance market. “SIs and end users will purchase fewer surveillance cameras than they would do if prices weren’t rising. This will mean that camera unit shipment growth will be lower in 2022 than it was in 2021. Growth will then be even lower in 2023 than it was in 2022,” Woodhouse and Cropley said. “The global average price of a network camera and an analog camera is forecast to fall again in 2024. However, a more prolonged period of high inflation presents a serious risk to this forecast.”
As for next year, the report said the global video surveillance market for hardware and software is forecast to grow at 11.7 percent in 2022 and will be worth an estimated US$28.2 billion. “We are forecasting much lower growth in 2023 followed by gradual recovery with growth increasing from 2024 onwards,” Woodhouse and Cropley said.
Cloud adoption increases, especially in U.S.
Technology-wise, a gradual trend to using the cloud for video surveillance also continued in 2021, particularly in the Americas region where the market for cloud video management software exceeded $150 million, the report said. It forecasts the number of cloud-connected surveillance cameras in the Americas will grow on average over twice as quickly as new network camera shipments between 2021 and 2026.
Especially, the report found the United States has been quicker to adopt cloud for video surveillance than most other countries in the world. Woodhouse and Cropley explains why this might be.
“It is down to a mixture of factors. A major factor is that it has many organizations with distributed sites, each with a small number of cameras. Furthermore, these organizations operate in a large country using a common language and a common set of rules on data residency, privacy etc. Bandwidth availability and cost have been more favorable than in some other countries too,” they said.
OSSA: Orchestrating the Digital Data Flow
Open Security & Safety Alliance (OSSA), a collaborative initiative focused on creating a framework for standards and specifications in the security, safety and building automation space, shares their thoughts on the unfolding trend that will change the security industry in the year to come.
By: Gijs van den Heuvel, Chair of Marketing at the Open Security & Safety Alliance (OSSA) and Manager Strategy and Partner Collaboration at Bosch Security Systems
As an industry alliance, we have a good idea of what’s in the works by our members when it comes to innovating across important areas within security, safety and beyond. OSSA representatives hail from some of the most influential companies in this space – such as Bosch, Hanwha Techwin, Milestone Systems and VIVOTEK – and together are determined to continue bringing forth an open, data-driven ecosystem.
The IoT is all about connecting things to make life easier, more intelligent, more intuitive and more productive. OSSA workgroups have projects underway that are primed to unify and elevate the market as a whole, so that there will be headspace to grow for all participants.
Already, OSSA organizations specified hardware and software conditions to make it possible to run third-party (AI) analytics applications securely on existing, brand-agnostic computer vision devices (starting with cameras). Working together spurred a handful of specifications for building common components (e.g., for core system requirements, cyber security directions and application interface APIs) that can be utilized jointly.
Now, a notable future trend our members are funneling expertise into is bringing about new levels of data and information sharing to all we do when it comes to smart cameras and related IoT devices. How can we achieve next-level dissemination of extremely valuable data flows sensed by products that collect data in their brand-specific siloes? How do we establish open but secure pathways to easily share and uniformly interpret data to connect the dots when it comes to surveillance and activities being captured and contained on edge-computing devices?
Millions of “things” generate, accumulate and house heaps of factual insights that – unfortunately all too often – remain untapped and stagnant once recorded. If corralled, connected and optionally given an artificial intelligence/ML scrub, this information brings a bigger picture of what transpired across, for example, surveillance systems. Taking frames or moments from various devices and drawing relationships between them to form a cohesive “data flow” opens up a massive new corridor of IoT-based possibilities. According to our group, there resides tremendous value in making data from one source consistently interpretable for another.
Imagine harnessing content from a camera that captures a car and applies its license plate app to read the details. This car crosses in front of the building, and minutes later another camera across the facility records an individual scaling a fence in the vicinity where the car was last detected. Finally, a third camera on the opposite side of the building records an individual onsite in an unauthorized area, and an open platform app detects a gun in hand. The system immediately then alerts security personnel about a potential threat. Tying this crucial information together seamlessly across a natural path of data flowing from security and safety devices takes us to a new frontier delivering not only what’s “seen” but more importantly what’s “sensed.” This is the future if device manufacturers and other stakeholders participate in an industry-driven ecosystem.
OSSA members are working on a set of generic, vendor-neutral data Application Programming Interfaces (APIs) to enable this type of uniform consumption and data interpretation across cameras or computer vision gateways in adherence with OSSA standards. It is also designed to apply across other device types within the IoT security and safety domain to provide a loose coupling between any pair of send/receive applications that collect, digest and interpret data. Enabling content from various brands and device types to be interpreted in an open forum brings newfound levels of storytelling and safety to our security situations.
Coupled with progress that continues around allowing for easy integration of third-party AI analytics applications on “Driven by OSSA” video cameras and gateways in an agnostic fashion, OSSA facilitates cooperation on many levels. There’s a front-row seat for anyone interesting in expanding this collaboration framework to together lift our industry to new levels of openness, innovation, interoperability and success.