Site and facility design: site selection, CPTED and layered physical security

Sub-domain 3.8 asks you to apply security principles to site and facility design. Every digital control ultimately depends on physical security: an attacker who can walk into a server room, plug a device into a network port or carry away a backup tape can defeat the best encryption and authentication. Physical security also protects the most important asset of all, people. This lesson covers how to choose a site, how crime prevention through environmental design shapes buildings and grounds, and how layered physical controls work from the property line to the server rack. Lesson 12 covers the specific controls for server rooms, media storage, utilities, environment and fire.

Principles first: life safety and layers

Two principles run through physical security. First, the safety of people always comes first. Fire exits must open, evacuation routes must stay clear and security measures must never trap people in an emergency, as lesson 1 noted. Building and fire codes set minimum requirements, and security designs must meet them. Second, physical security uses defence in depth, often described as concentric rings or zones: the site perimeter, the building exterior, the building interior and public areas, restricted areas, and the most sensitive rooms and cabinets. Each ring should deter, detect, delay and allow a response. Deterrence discourages attempts; detection notices them; delay slows an intruder long enough for guards or police to respond; and response stops the intrusion. A good design ensures that the delay at each layer is longer than the time needed to respond.

Site selection

Security begins with the choice of site, and the most effective time to influence it is before the lease is signed or the foundation is poured. Planners consider natural hazards, such as flood plains, earthquake zones, wildfire risk, hurricane or tornado exposure and severe winter weather; man-made hazards, such as nearby chemical plants, rail lines carrying dangerous goods, airports and high-crime areas; access to utilities, with diverse routes for electricity, water and telecommunications; proximity to emergency services, transport and staff; visibility, since a prominent sign announcing a data centre attracts attention; and the political, legal and regulatory environment, which affects data sovereignty. For critical facilities, a second site far enough away not to share the same disasters supports business continuity. Data centre operators often choose discreet, unmarked buildings set back from roads, with space for perimeter controls.

Crime prevention through environmental design

Crime prevention through environmental design, CPTED, uses the design of buildings and landscapes to discourage crime. It was developed in the 1970s by criminologist C. Ray Jeffery and architect Oscar Newman, among others, and rests on a few principles. Natural surveillance arranges spaces so that people can easily see and be seen, for example with windows overlooking parking areas, good lighting and low landscaping that offers no hiding places. Natural access control guides people towards controlled entrances using walkways, landscaping, fences and lighting, so that it is obvious where visitors should go and suspicious when someone goes elsewhere. Territorial reinforcement uses clear boundaries, signs and design features to show that space belongs to someone and is cared for, which encourages legitimate users to challenge strangers. Maintenance, sometimes linked to the broken windows idea, keeps the site in good repair, because neglect signals that nobody is watching.

CPTED works best combined with other measures. For example, a car park might have clear sight lines from a staffed reception, bright and even lighting, a single controlled vehicle entrance, signage and cameras. The design deters casual crime cheaply, and the technical controls handle determined attackers. Second-generation CPTED adds social factors such as community engagement, which matters for campuses and shared buildings.

Perimeter and exterior controls

At the perimeter, fences mark boundaries and delay intruders. Rough guidance often cited in study materials is that a fence about 1 metre or 3 to 4 feet high deters casual trespassers, about 2 metres or 6 to 7 feet is too high to climb easily, and about 2.4 metres or 8 feet topped with barbed or razor wire deters determined intruders. Gates, bollards and vehicle barriers stop vehicles; crash-rated barriers are tested against vehicles of specified weight and speed. Lighting should be continuous and even, avoiding dark patches and glare, and placed to support cameras. Security guards provide judgement, deterrence and response that technology cannot, but are expensive and need training and supervision; dogs are sometimes used at high-security sites. Signs warning of surveillance and prosecution add deterrence.

Building exteriors should limit the number of entrances, protect ground-floor windows, secure roof access and utility openings, and place sensitive rooms in the interior rather than against exterior walls. Loading docks are often the weakest point, because doors stay open and deliveries arrive constantly; they need separate controls, cameras and procedures so that delivery staff cannot wander into the building.

Entry control and doors

Entry control decides who may enter which areas. Badge systems using smart cards or mobile credentials grant access by zone and time and log every entry. Biometric readers, using fingerprints, hand geometry, iris or face, add strong assurance for sensitive areas and are discussed further in the identity domain. Mantraps, also called access control vestibules, have two interlocking doors so that only one person passes at a time and the second door opens only after the first closes and the person is verified. Turnstiles and security gates do a similar job for high-traffic lobbies. Tailgating or piggybacking, following an authorized person through a door, is the most common way to defeat entry control, and is addressed with mantraps, turnstiles, guards and staff training to challenge strangers.

Locks come in many forms: mechanical key locks, combination and cipher locks, and electronic locks controlled by the access system. Key control matters as much as the lock: who holds keys, whether they can be copied and what happens when they are lost. Electronic locks may be fail-safe, unlocking when power fails, which suits exit doors for life safety, or fail-secure, staying locked, which suits doors to sensitive rooms that are not on escape routes. Doors themselves should match their purpose: solid-core or metal doors, hinges on the secure side or with pins that cannot be removed, strike plates resistant to prying, and fire-rated doors where codes require them. A strong lock on a hollow door in a flimsy frame protects little.

Surveillance and intrusion detection

Video surveillance, often still called closed-circuit television, deters crime, supports investigations and allows remote monitoring. Designers decide what each camera must achieve: detection of a person, recognition of a known individual or identification of a stranger, each needing more pixels on the target. Modern systems use network cameras with analytics that can alert on movement in restricted areas, and they must be secured like any other networked device, with changed default credentials, current firmware and segmentation, given the IoT risks described in lesson 7. Recordings must be retained long enough to support investigations and protected against tampering. Privacy law limits surveillance: cameras must not intrude where people have a reasonable expectation of privacy, and signs and policies should explain their use.

Intrusion detection systems for buildings use sensors such as magnetic door contacts, glass-break detectors, passive infrared motion sensors, vibration sensors and photoelectric beams. They raise alarms at a monitoring station, which dispatches guards or police. False alarms are a persistent problem; too many lead to alarms being ignored or to police charging for responses, so sensors must be well chosen and maintained. Alarm systems need backup power and protected communication paths so an intruder cannot defeat them by cutting a wire.

In Canada, building design must meet the National Building Code and National Fire Code as adopted, sometimes with changes, by each province and territory, and security features such as locked exit doors must comply with their rules on egress, including requirements for electromagnetic locks to release on fire alarm and power failure. Privacy commissioners have published guidance on video surveillance by organizations under PIPEDA and provincial laws, emphasizing a clear purpose, minimal intrusion and signs. The Canadian Centre for Cyber Security and Public Safety Canada publish guidance on physical security for critical infrastructure.

Thinking like a manager

When a question asks for the most important consideration in physical security, the answer is almost always human safety. When it asks about the best time to address physical security, choose site selection and design rather than retrofits. When a scenario describes strangers following employees through doors, think mantraps, turnstiles and staff awareness. And remember that deterrence, detection and delay only work if someone can respond in time.