High-security perimeter fencing topped with razor wire at dusk
Images: iStock

Inside story: Perimeter protection

Beyond the boundary and the evolution of intelligent security
Jeremy Malies, Content Editor, In Security

In Security considers what constitutes a perimeter (our ideas are changing) and discusses techniques and products being used. Jeremy covers interoperability, resilience, sensor types, and the cyber perspective.

As usual in these features, the article looks at vertical sectors that make unusual demands on vendors and installers. These verticals include maritime, critical national infrastructure (CNI), construction, and data centres. Inevitably, there is consideration of AI which is now the not so dumb elephant in the room. What is AI achieving in perimeter protection? Are there limitations and what might be the dangers of over-hyped marketing?

Even a moment’s analysis tells observers that this sector is in flux and a shift in perspective is occurring. Facilities managers no longer merely define and seek to protect a boundary. The frameworks have become overarching; optimal perimeter protection should be a way of thinking and the in- vogue term is fusion. It has taken over from “layered protection” and “defence in depth”.

The best system designers are concerned with the potential for perimeter protection devices to combine with each other. The focus is less on spec lists and more on the intelligence (both AI and human) with which individual solutions combine. These products might be fencing, fence sensors, buried sensors, video surveillance, thermal imaging, radar, and 3D LiDAR.

Optimal solutions create successive boundaries and delaying factors that an intruder must overcome so increasing the chance of detection. Facilities managers expect the individual technologies to be greater than the sum of their parts. A candid site supervisor might concede that a determined intruder will be able to scale a fence but will often be detected in the sterile zone between this and the next fence or obstacle.

RAF Brize Norton

Nowhere in my survey is there anything more extraordinary than events in June of last year when two activists from the group Palestine Action breached the perimeter of RAF Brize Norton in Oxfordshire. They were on scooters and sprayed two aircraft with red paint to symbolize blood.

Industry observers were amazed not only at the breach but the fact that the activists were able to leave the site undetected. The group even released a video of their members riding electric scooters across the base. The political dimension should not concern us since it’s not this magazine’s remit. I focus instead only on the content of a statement to the House of Commons in the following September by Parliamentary Under-Secretary of State for Defence, Louise Sandher-Jones.

A Sandhurst graduate and former second lieutenant in the British Army, Sandher-Jones has an impressive résumé. But the parliamentary record, Hansard, tells us that she was sent out to bat with a description of the airbase as merely “needing further investment”. There was also a reference to “years of under-investment and hollowing-out under the previous administration”.

Brize Norton is now reported to be operating an improved system for detecting and tracking attempted breaches with links to a control room that has been overhauled. There has also been investment in drones, fencing, and increased manned patrols. Palestine Action a well-resourced group with determined activists but it’s salutary to reflect on what jihadists might do at a military base.

This incident is unprecedented. It warranted the headlines but does not truly help technical observers identify trends. If we want to learn about the sector through repeated behaviours, it’s best to assess what end-users are saying about their expectations and recent experiences. And we should drill down into specific verticals within perimeter protection.

perimeter4 6oct26

Hostile vehicle mitigation

As a technical publication, In Security should focus on perimeter protection incidents when they have been part of an attack by a named organization on a defined site with a physical boundary. Saying this, I don’t want to undermine the gravity of attacks on pedestrians in public places during markets, festivals, feast days, and historical commemoration days. But the cases particularly relevant to my survey involve vehicle attacks against permanent perimeters protected by gates, turnstiles, barriers, and bollards. The facilities are likely to be a military installation, critical national infrastructure (CNI) or government building.

A relevant case study for this survey – and one that might hold potential for hardening the sector – is an incident from March 2025. At Bannu in Pakistan (it’s close to the border with Afghanistan) a pair of suicide bombers drove explosive-laden vehicles into the outer wall of the Bannu Cantt military compound. The pair were affiliated to the Pakistani Taliban.

The explosions breached the wall after which six attackers entered the compound. They were killed by commandos. It’s important to note here that the successful perimeter breach (the collapse of part of the wall) was a planned component of a larger assault. Six patrol staff and six attackers died and there were numerous injuries including to civilians. The lesson here is the need for successive lines of fencing and sterile zones rather than a wall that is an integral part of the facility.

perimeter3 6oct26

Industry soundbites

Bottlenecks and vulnerabilities

Nemer Mdradas, a regional director with Gallagher Security, is clear on his expectations of perimeter protection. “Zero intrusions is not a goal – it’s a must!” Mdradas’s press statements identify high rates of false/nuisance alarms as a pain point for perimeter security. He believes that unwanted alarms dilute situational awareness. Mdradas consistently uses evocative turns of phrase and points to “AI video analytics that is added as an afterthought” as having much to answer for in terms of system failures. He is optimistic about intelligently monitored pulse fencing systems meeting operational needs. The subtle challenges come when legitimate activity is being carried out close to the areas that are likely to be breached by intruders.

Mdradas is cyber-aware, as all physical security executives must be, and on record as insisting that system design should ensure that physical access points never become backdoors for cyber intrusion. These are demanding but inspiring industry goals.

Avoiding passive boundaries

There is now a consensus that perimeter protection should not be a passive boundary. It’s healthy that end-users insist that perimeters ought to be the scene of confrontation (digital, usually through voice address) rather than retrospective analysis after a breach has occurred.

Responses should come while a situation is gathering momentum. And perimeter protection can be aimed at legitimate site users as well as intruders. Thus, it requires only basic video analytics and voice messaging for a site operative to be told that he has entered an area that requires a hardhat, hazmat suit or similar but is not using them.

Northern Irish accent!

Mobile-deployable perimeter protection video surveillance units that are more or less independent and can be placed around the edges of estates often end up looking like Daleks. This is a matter of context-sensitive design and can be seen across several manufacturers. These units are invaluable for construction sites, railway lines, and temporary event locations that may have limited power and Internet connectivity. Much is being achieved with solar battery power and 4/5G cellular connections.

A proactive aspect of these units is their automated talkdown messages to intruders with warnings that their presence has been detected, reported, and acted on. I’ve been amused at trade fairs to note that the most intimidating accent for a talkdown message is the upward gliding vowel sounds used in Belfast! And these talkdowns are increasingly customized to precise site location and context. Intervening sufficiently early to stem escalation is the obvious goal.

An area where perimeter protection has seen improvements is the ability of automated systems to assess the threat or otherwise from an intruder. Certain intruders such as a wayward (perhaps bored) but not seriously ill-intentioned youngster or a disorientated country walker should be identified, alarmed upon, and interacted with if only for their own safety. But the perception of threat level must be proportionate, and discrimination must be intelligent. Optimum perimeter protection is a combination of mechanized products and focused reaction from manned patrols.

Digital twins

A consolation of spending 30 years in this industry is the opportunity to observe true evolution. Writing rail case studies over decades has seen me assess integrated software management platforms for transport hubs that represent the facility and its technology assets as a 3D model. The best vendors optimize site awareness for operators and reduce training time for new recruits. Managers have given feedback to the effect that they have been able to make the best possible decisions with the best possible information.

But there is now a wholly different methodology for this – digital twins. This is a dynamic, digital representation of facilities and their security systems. The approach allows combination of the site’s layout with live updates from fence sensors, video surveillance, radar, thermal imaging, access control, and SCADA.

It should be noted that this is not 3D modelling of a building and its perimeter. There is true connection between the actual site and its digital twin so helping operators understand the events and movements they are monitoring remotely. The result is improved analysis of developing threats allowing the most appropriate response.

perimeter5 6oct26

The verticals

Maritime perimeter protection

Having grown up around the port of Newhaven and spent time at the nearby port of Shoreham, I follow maritime case studies closely. Consultant friends tell me that at major ports they must design for a variety of activities going on across what is usually a diverse estate. As a result, fence-mounted cable sensors are often calibrated zone by zone to factor out nuisance alarms.

In this way, an area that is seeing crane usage would be adjusted in a different way to an area that is receiving only foot passengers. And within the perspective of illegal migration, often from the Schengen Area to the UK, a passenger terminal might have sensors adjusted to maximum sensitivity with managers willing to tolerate a higher level of false alerts than is the norm.

Ports handling everything from bulk cargo to cruise liners not to mention ferries and leisure craft put a myriad of demands on security managers. A complication is that it’s common for a port’s terrain to vary between open coastline, industrial areas, and passenger areas.

Zonal adjustments at ports

On fences with detection systems, the manufacturers’ algorithms focus on the kind of sharp or abrupt frequency that might be consistent with attempts at breaching a perimeter on foot or in a vehicle. Lower frequencies caused by expected ambient movement around the port and extreme weather are rejected as alarm sources.

Fibre optic sensors transmit laser light across optical fibre with the alarms being generated by vibrations that are deemed meaningful according to setting. A different approach is vibration sensors that react to physical changes caused by alarm-worthy activity such as climbing, lifting, and cutting the port’s perimeter barriers. A taut-wire system will observe the tension in a wire, detecting and reporting on changes that could be consistent with an act of intrusion.

An airport and its “corridor” mode

Look at how end-users prioritize the benefits they have gained from successful installations when talking to trade media. Number one is often the convenience and the economies of being able to oversee their facility and drill down into different disciplines from a single platform.

This was a key takeaway when facilities managers from British Airways Engineering at Heathrow Airport worked with Johnson Controls. The trendy phrase is “ecosystem” and it is descriptive in that it suggests an interconnectedness that was relevant in this case study which encompassed an area equivalent to 109 football pitches.

The site involved 10 hangars and 35 aircraft stands supporting an airline fleet. Perimeter protection was relying significantly on foot patrols. The case study shows that video surveillance exploiting intelligent algorithms is now a credible perimeter protection mechanism and should be regarded as such alongside fencing and radar.

Two hundred Johnson Controls Illustra cameras were used at Heathrow, the model being one that can operate in corridor mode. Naturally, this can be a mode in which the camera’s sensor resources are used economically by showing a tall narrow image such as a fence and the strip of ground alongside it.

But that can be a simplistic interpretation. Video surveillance of perimeters often has the camera in a regular landscape mode, particularly as there is a need to distinguish between human intruders and large wildlife and perhaps alert on miscreants who have scaled one fence and are proceeding through a sterile zone to the next fence. Radar has also been used on this project and it should be noted that radar detection can prompt a PTZ camera to align with the threat. The Heathrow project has impressed as part of a drive towards unified operational intelligence.

perimeter2 6oct26

Perimeter security at a federal data centre in the US

You would be delusional to expect vendors to reveal locations and names of data centres they serve with perimeter protection solutions. It’s enough that the case studies are informative and give insight into the challenges and equipment used. The project I now describe is useful for my overall intentions here because it involves both a fibre optic solution and LiDAR.

RBtec Perimeter Security Systems have recently supplied their RaySense Distributed Acoustic Sensing (DAS) solution across a 2.5 km perimeter at a data centre run by a US federal agency. And there is a familiar aspect to this story that runs through more modest installations. It was an absolute requirement that the perimeter boundary should function throughout the installation process and not even be weakened during fit-out. A coverage gap was unacceptable. How many times have we heard that? And quite rightly at any mission-critical facility.

RaySense is a fibre optic intrusion detection system that is mounted onto existing fencing infrastructure. The cable itself is the sensor; there are no power or electronics requirements along the fence. The documented performance of threat discrimination and the rejection of nuisance alarms is impressive.

The federal data centre has a continuous fenced boundary with vehicle and personnel gates. This case study saw replacement of a legacy copper-based sensor system that was unfit for purpose. The site is prone to electrical storms and copper was behaving like an antenna. Five kilometres of sensor cable were installed and the perimeter was configured into 11 zones of operation. (These reflected how the site is patrolled by staff and how alerts are responded to.)

The company also supplied their PulseVI LiDAR units which are giving three-dimensional tracking of movement in the areas that are open for vehicles and people.

RaySense contains no copper and this was a clincher in terms of product specification. The RBtec literature for the project includes a pull-quote from its team who note (with wry humour) that copper on a fence in an area that is prone to storms is a no-no. “The weather will find it every time.”

perimeter6 6oct26

Aviation perimeter protection case study

Sydney Airport began as a paddock and dates from 1919. It now has three runways and deals with 44 million passengers a year. The facility’s 24km perimeter has been secured by Gallagher Security (an executive of this company has been quoted above) with its Pulse Fence solution.

The airport’s 24-km perimeter is unusual in that it includes a sea wall and the Pacific shore. It had been relying on a standard chain-wire fence topped with razor wire (a height of 2.44 metres which has been an aviation standard) until this was deemed unfit for purpose. Alerts on perimeter breaches were coming only through manned guarding patrols and video surveillance. An active fencing solution was needed, one that could work from within a physical barrier and detect proactively. A Gallagher system was specified by integrator HRC Security.

The Gallagher monitored Pulse Fence system is based on a grid of energized, high tensile wires with multi-zone management capability. The grid can be constructed inside a new or existing perimeter fence. It integrates with video surveillance and other disciplines through simple IP networking. The fencing delivers a short and safe shock to the intruder. This is an energized pulse sent around the perimeter fence line. It meets stringent safety criteria and zoned fences allow adjustments according to need.

Conclusions

A variety of stakeholders (going right up the chain to architects) are involved in securing perimeters. There is now an evolutionary process from “detect everything” to “be discriminating as to what matters”.

This article has assessed actual case studies and quoted engineering executives describing real-life challenges. In other words, I’ve sought to avoid hyperbole. A distinction that needs to be drawn is that headline-grabbing innovative technology should not distract us from basic issues such as poor site design from the get-go, inadequate detection coverage, nuisance alarms through imprecise settings, and slow responses by patrol staff who may lack motivation.

There are however, as Ian Dury taught us in a song, “reasons to be cheerful”. Hostile vehicle mitigation continues to develop from crude individual barriers into sophisticated, integrated designs. Increasingly, ONVIF compliance means that intelligent fencing can work alongside perimeter-protection-specific surveillance cameras.

Perspectives are broader now; the industry has moved on from merely protecting a boundary understanding threats across a whole facility. The best attention of the brightest people and best-of-breed products should be focused on generating reliable alarm responses from both sensors and humans.

Walk round an industry trade fair and a conscientious sales executive is less likely now to be fixated on selling you a stronger bollard. The focus from premium manufacturers is creation of a generally protective environment. Vehicle approach points are managed, vulnerable routes receive special attention, access is supervised, and there are ways to counteract the consequences of a penetration.

Most importantly perhaps, my survey here tells me that the perimeter is no longer seen as a simple boundary. It’s the area where incidents are assessed by human intelligence and increasingly discriminating AI models. The situation is then deemed to non-threatening or escalated to involve the full resources of the site and if necessary outside agencies.