printer friendly version

Open systems support hybrid surveillance

Back in the day, surveillance was an on-premises solution with powerful servers recording video streams, and video management systems (VMS) were used to replay and search footage when an incident occurred. Today, the options have broadened, and AI has significantly expanded technology’s analytical capabilities, both on servers and on the camera.

End users, therefore, have the option to select the best technology for their requirements, whether on-site, at the edge (with AI-enhanced cameras), or in the cloud. In reality, there is no best solution, only a hybrid combination of what works best for users in different scenarios.

SMART Security Solutions asked two leading surveillance technology providers to explore hybrid surveillance systems to identify the options available in the local market for different scenarios. Our answers come from:

• Rudie Opperman, manager for engineering and training MEA at Axis Communications, and

• Walter Rautenbach, MD of neaMetrics, local distributor of TRASSIR.

Walter Rautenbach.

SMART Security: Is the local market adopting hybrid surveillance solutions or holding onto traditional on-site servers and video processing?

Opperman: Hybrid surveillance architectures are becoming increasingly common, particularly in mid- to large enterprises and critical infrastructure environments. Traditionally, many organisations relied fully on on-premises systems with centralised servers and storage. However, the rapid advancement of edge processing capabilities, combined with the scalability and flexibility of cloud services, has shifted the conversation from ‘either/or’ to ‘what combination makes the most operational sense.’

Today, many deployments incorporate:

• Edge-based analytics and event filtering at the camera level.

• On-site servers and VMS for operational control and compliance.

• Cloud-based services for centralised management, long-term storage, AI analytics, or remote access.

Hybrid solutions are especially well-suited to:

• Retail and multi-site commercial environments.

• Critical infrastructure (utilities, energy, water).

• Transport and logistics.

• Education and healthcare.

• Smart city deployments.

In these sectors, resilience, scalability, and cost optimisation are key drivers behind hybrid adoption.

Rautenbach: Hybrid surveillance is gaining significant traction, though adoption varies considerably by customer size and sophistication. Traditional all-on-premises deployments – a server, an NVR, and a VMS – remain the dominant model for smaller or single-site installations. For multi-site organisations, enterprises managing legacy infrastructure, and anyone looking to add intelligence to an existing camera estate, hybrid architectures have become the natural answer.

To understand where hybrid fits, it helps to be clear about what ‘hybrid’ means in practice. In our experience, it rarely means ripping out existing cameras and replacing them with smart edge devices. More often, it means layering intelligence onto what is already there, which is precisely why TRASSIR has always positioned itself primarily as a software and AI analytics company rather than a hardware vendor. Our VMS and analytics modules are designed to run on virtually any IP camera via ONVIF, RTSP, or native integration, which means customers can introduce AI capabilities without being locked into any single camera manufacturer or forced into a full hardware refresh.

In terms of industries where hybrid is most relevant, retail chains are leading the adoption, driven by the need to manage dozens of distributed sites from a central dashboard. Logistics and warehousing follow closely, where a mix of on-site intelligence (for real-time stock and safety monitoring) and centralised reporting makes operational sense. Transport hubs, industrial facilities, and mixed-use commercial properties are also strong candidates – essentially any environment where multiple sites need centralised oversight, but cannot compromise on local reliability.

SMART Security: When looking at a hybrid system, what points should be considered when deciding on edge, cloud, and/or on-premises?

Opperman: Edge analytics are most valuable when:

• Immediate response is critical (e.g., perimeter detection or intrusion alerts).

• Bandwidth is constrained.

• Privacy filtering (e.g., metadata extraction instead of raw video transmission) is required.

• Only relevant data should be transmitted or stored.

Modern edge devices can significantly reduce server load by filtering and processing events before data leaves the device.

On-premises infrastructure remains important where:

• Regulatory or compliance requirements mandate local storage.

• High-resolution video retention is required.

• Network reliability is inconsistent.

• Latency-sensitive operations demand guaranteed performance.

On-site VMS platforms also provide strong operational control and resilience in critical environments.

Cloud services are increasingly used for:

• Centralised management of distributed sites.

• AI-based analytics and advanced search.

• Long-term or redundant storage.

• Software updates and system health monitoring.

The cloud offers scalability and operational simplicity, especially for organisations managing multiple geographically dispersed sites.

Integration and unified management

A key success factor in hybrid deployments is interoperability. Modern management platforms are evolving toward a unified operational view, often referred to as a ‘single pane of glass,’ in which edge devices, on-premises servers, and cloud services can be monitored and managed through a single interface.

This is why open application platforms at the edge that allow a wide range of analytics to run on the device itself are so critical. They ensure that data is processed in the most efficient location before being passed to the VMS or cloud service. Open standards and API-driven architectures are essential to ensure hybrid systems remain flexible and future-proof, rather than locked into rigid ecosystems.

Rautenbach: When it comes to edge surveillance, true on-camera AI edge processing, the kind that runs deep learning inference fully independent of any server, is still not a main focus in the TRASSIR ecosystem. Most TRASSIR cameras in our Trend, Pro, and Ultra lines include hardware analytics for basic on-board functions such as motion detection, line crossing, and intrusion zone alerts. These are valuable for reducing bandwidth usage and triggering recordings without server involvement, but they are distinct from the full AI analytics for which TRASSIR is known.

However, this is changing. Our newly launched AutoPass camera is a significant milestone. It is the first TRASSIR camera to run our ANPR analytics module natively on-board, achieving 98-99% licence plate recognition accuracy without requiring a server in the loop. It supports multi-country recognition, integrates directly with access control systems and barriers, and represents the direction we are moving toward as on-camera processing power continues to grow.

In the near term, edge intelligence in a TRASSIR hybrid system is best understood as the on-site server layer rather than the camera itself, enabling us to provide analytic intelligence to current deployments independent of camera capabilities.

On-premises core

On-premises processing is where TRASSIR’s core intelligence sits. Our NeuroStation servers run TRASSIR OS and use CPU and GPU resources to power the full suite of AI analytics modules – Neuro Detector, Neuro Counter, Face Recognition 2.0, AutoTRASSIR (LPR), ActiveStock, Fire and Smoke Detector, Pose Detector, and more – processing streams from any compatible camera in real time.The key driver of this decision is the depth of analytics. If you need accurate, server-grade AI processing with tight response times and no dependence on internet connectivity, the on-site server is the right layer. Data sovereignty and compliance requirements are also strong drivers. For many customers, particularly in banking, government, and industrial sectors, footage must remain on-premises, with an option to synchronise to a central location.

A practical advantage of TRASSIR’s software-first approach is that these analytics can run with third-party NVR hardware and on customer-owned servers running Debian OS with a dedicated GPU capability, not just on TRASSIR’s own NeuroStation hardware. This matters enormously in hybrid scenarios where customers want to retain existing infrastructure investments.

Image credit: Gemini.

Centralised monitoring and system integration

Centralisation often means centralised monitoring, access, and administration across distributed on-site systems. TRASSIR CMS (central monitoring station) is designed to unify multiple on-site VMS servers and remote locations into a single operational environment. It provides centralised user and rights management (including LDAP and Active Directory), consolidated monitoring, and archive synchronisation between sites for redundancy and resilience.

Remote access and multi-site viewing can be organised through the CMS without requiring full cloud-based video storage or analytics processing. This approach is particularly relevant in regions with bandwidth constraints or regulatory limitations, where keeping video locally, while enabling central visibility is essential.

A single CMS Station can connect up to 5000 servers, making it suitable for large enterprise-scale deployments. Load distribution and connection optimisation within CMS ensure that viewing and management traffic is handled efficiently, even across geographically dispersed sites.

SMART Security: In a hybrid scenario, what equipment and solutions does your company supply to make hybrid surveillance work for customers?

Opperman: Axis has long focused on distributed intelligence at the edge, open system architecture, and scalable integration. In hybrid deployments, Axis contributes through:

• A portfolio ranging from advanced cameras with deep learning processing units (DLPU) to network audio and intercoms, serving as the intelligent edge. These devices run powerful analytics directly on board using the AXIS Camera Application Platform (ACAP).

• Support for open VMS platforms and third-party integrations.

• Secure device management and lifecycle services.

We also provide tools that bridge these domains, such as AXIS Device Manager for large-scale device management and cloud-connected services like AXIS Camera Station Secure Entry for access control, allowing partners to build comprehensive solutions.

Our technology is designed to operate within open, interoperable ecosystems rather than closed proprietary stacks. This flexibility enables system integrators and end users to design hybrid architectures aligned with their operational and regulatory requirements.

Training and partner enablement

Successful hybrid systems depend heavily on system design expertise. Axis invests significantly in partner enablement through the Axis Communications Academy, which offers structured technical certification programmes, design and architecture guidance, and specialised training in areas such as cybersecurity and network best practices. We also provide ongoing technical support and integration resources. Hybrid environments require a strong understanding of networking, data management, cybersecurity, and analytics integration. Our goal is to equip partners with the skills needed to design resilient, future-ready surveillance ecosystems.

Rautenbach: Because TRASSIR is fundamentally a software and AI analytics company, our contribution to a hybrid solution is primarily the intelligence layer, and that intelligence can sit across all three tiers of the hybrid architecture described above.

At the camera level, TRASSIR manufactures a wide range of quality cameras to suit various analytical needs, enabling fully integrated turnkey deployments. Our Trend, Pro, and Ultra cameras provide basic hardware analytics for bandwidth-efficient triggering, with the AutoPass marking our first step into full on-camera AI. These cameras work best when paired with TRASSIR’s server analytics, but any third-party camera that meets the quality and resolution requirements can feed into the same system.

At the on-site server level, TRASSIR NVRs and VMS platforms, deployed on NeuroStation hardware (or compatible third-party servers), provide GPU-accelerated AI analytics.

At the management level, TRASSIR CMS provides the unified oversight, remote access, and redundancy that ties a distributed hybrid deployment together.

On partner training: TRASSIR operates a formal training and certification programme (trassir.com/certificates/) to ensure integration partners can confidently design, deploy, and support hybrid deployments. This includes those that mix TRASSIR software with third-party cameras and hardware. We also provide pre-sales support, a storage and analytics calculator, and dedicated marketing materials.

For more information contact Axis Communications, +27 11 548 6780, terri.miller@axis.com, www.axis.com

Share this article:

Further reading: