13.2 Physical Barriers, Fencing, Gates, Bollards (ASTM Crash Ratings), and Vehicle Mitigation

13.2 Physical Barriers, Fencing, Gates, Bollards (ASTM Crash Ratings), and Vehicle Mitigation

The outermost physical defensive boundary of a mission-critical data center—Level 1 of the concentric security model—serves as the first line of defense against deliberate kinetic attacks, criminal intrusion, and catastrophic vehicular threats. A major design-basis threat confronting modern hyperscale and enterprise facilities is the Vehicle-Borne Improvised Explosive Device (VBIED) and direct vehicle-ramming attacks aimed at breaching the facility facade or disabling outdoor electrical substations, fuel storage yards, and emergency diesel generators.

To safeguard critical infrastructure, the Data Center Design Consultant (DCDC) must integrate three synchronized engineering elements: physical standoff distance, crash-rated perimeter barriers and bollards, and anti-climb perimeter fencing integrated with Perimeter Intrusion Detection Systems (PIDS). Mastery of vehicular kinetic energy calculations, modern crash ratings (ASTM F2656), and environmental site geometry is essential for designing resilient perimeter protection.

---

Vehicular Threat Mitigation and Standoff Distance Principles

Standoff distance is the physical buffer distance separating a potential explosive detonation point (such as an uncontrolled public street, boundary fence, or parking lot) from the nearest structural building envelope or critical infrastructure component (e.g., utility transformers, diesel fuel tanks, cooling towers).

   [ Public Roadway ]
         |
         v (Hostile Vehicle)
   [ Perimeter Fence & Crash Barrier Line ]
         | <---------------- Outer Clear Zone (about 20 ft) ----------------->
         | 
         | <============ STANDOFF DISTANCE (target about 100 ft) =============>|
         |                                                                      |
         | <---------------- Inner Clear Zone (about 30 ft) ----------------->  |
         |                                                                      v
   [ Internal Patrol Road ]                                              [ Hardened Facade ]
                                                                         (Data Center White Space)

Blast Physics and Hopkinson-Cranz Scaling

When an explosive device detonates, it generates an instantaneous shock wave characterized by a high peak positive incident overpressure ($P_{so}$) followed by a negative phase suction wave. As the shock front strikes a rigid structural surface (such as a data center concrete facade), the pressure is amplified by reflection, creating reflected blast pressure ($P_r$) and reflected impulse ($i_r$) that can cause catastrophic wall shearing, column buckling, and progressive building collapse.

Blast wave behavior is mathematically characterized by the Hopkinson-Cranz scaled distance formula:

$Z = \frac{R}{W^{1/3}}$

Where:

• $Z$ = Scaled distance parameter (expressed in $\text{ft/lb}^{1/3}$ or $\text{m/kg}^{1/3}$)
• $R$ = Physical standoff distance from the detonation center to the structure (in $\text{ft}$ or $\text{m}$)
• $W$ = Equivalent mass of TNT explosive charge (in $\text{lbs}$ or $\text{kg}$)

Peak blast overpressure falls steeply and non-linearly with standoff distance (roughly $P \propto 1/R^3$ in the near field, approaching $P \propto 1/R$ in the far field). Consequently, increasing standoff distance is the single most cost-effective engineering countermeasure against explosive blast damage.

Standoff Criteria in Practice

ANSI/BICSI 002 does not assign a fixed standoff distance to each Availability Class. It advises maximizing the distance between vehicle-borne threats and the building and consulting security specialists and law enforcement about the likely device size. Department of Defense UFC 4-010-01 (DoD Minimum Antiterrorism Standards for Buildings) gives standoff tables for military buildings, and federal and commercial security planners commonly aim for roughly $100\text{ ft}$ ($30\text{ m}$) or more where the site allows.

Where a dense urban parcel makes adequate standoff impossible, the building envelope is analyzed with blast modeling and hardened: thickened reinforced concrete walls, composite wraps, blast-resistant glazing, and catch systems sized to the design-basis threat.

---

Crash-Rated Barriers: ASTM F2656 vs. Legacy DoD Ratings

When a hostile vehicle attempts to ram through a perimeter to eliminate standoff or strike outdoor equipment, the perimeter barrier must absorb and dissipate the vehicle's entire kinetic energy, bringing it to a complete halt within a strictly controlled penetration distance.

The Kinetic Energy Equation

The dynamic force exerted by an impacting vehicle is governed by the kinetic energy formula:

$KE = \frac{1}{2} m v^2$

Where:

• $KE$ = Kinetic energy in Joules ($\text{J}$) or Foot-Pounds ($\text{ft-lbs}$)
• $m$ = Vehicle mass in kilograms ($\text{kg}$) or slugs ($w/g$)
• $v$ = Vehicle velocity at the moment of impact in meters per second ($\text{m/s}$) or feet per second ($\text{ft/s}$)

Because velocity is squared ($v^2$), doubling the vehicle's speed quadruples its destructive kinetic energy. A medium-duty truck traveling at 50 mph ($80.5\text{ km/h}$) possesses 2.78 times the kinetic energy of the exact same truck traveling at 30 mph ($48.3\text{ km/h}$). Perimeter road geometry must be designed to eliminate straight-line acceleration approaches.

The ASTM F2656 Standard

For decades, vehicle barriers were specified using the U.S. Department of State SD-STD-02.01 "K-ratings" (K4, K8, K12). In 2007, that standard was replaced by ASTM F2656 / F2656M (Standard Test Method for Crash Testing of Vehicle Security Barriers), now in its 2023 edition, which provides a more granular test protocol. Lighter vehicles (SC, FS, PU) are tested at 30 to 60 mph and heavier vehicles (M, C7, H) at 30 to 50 mph.

+---------------------------------------------------------------------------------------------------------+
|                               ASTM F2656 VEHICLE CRASH TESTING MATRIX                                   |
+-----------------------+-------------------+--------------------+------------------+---------------------+
| TEST VEHICLE TYPE     | VEHICLE CODE      | VEHICLE WEIGHT     | IMPACT SPEEDS    | CORRESPONDING       |
|                       |                   |                    | TESTED (MPH)     | ASTM RATING         |
+-----------------------+-------------------+--------------------+------------------+---------------------+
| Small Passenger Car   | SC                | 2,430 lb (1,100 kg)| 30, 40, 50, 60   | SC30 ... SC60       |
+-----------------------+-------------------+--------------------+------------------+---------------------+
| Full-Size Sedan       | FS                | 4,630 lb (2,100 kg)| 30, 40, 50, 60   | FS30 ... FS60       |
+-----------------------+-------------------+--------------------+------------------+---------------------+
| Pickup Truck          | PU                | 5,070 lb (2,300 kg)| 30, 40, 50, 60   | PU30 ... PU60       |
+-----------------------+-------------------+--------------------+------------------+---------------------+
| Medium-Duty Truck     | M                 | 15,000 lb          | 30, 40, 50       | M30, M40, M50       |
| (Common DC Threat)    |                   | (6,800 kg)         |                  | (Replace K4/K8/K12) |
+-----------------------+-------------------+--------------------+------------------+---------------------+
| Class 7 Cab-Over      | C7                | 15,873 lb          | 30, 40, 50       | C7-30 ... C7-50     |
|                       |                   | (7,200 kg)         |                  |                     |
+-----------------------+-------------------+--------------------+------------------+---------------------+
| Heavy Goods Vehicle   | H                 | 65,000 lb          | 30, 40, 50       | H30, H40, H50       |
| (Tandem Dump Truck)   |                   | (29,500 kg)        |                  |                     |
+-----------------------+-------------------+--------------------+------------------+---------------------+

ASTM Penetration Ratings (P-Ratings)

ASTM F2656 measures the dynamic penetration distance past the barrier's pre-impact reference line (for trucks, the leading edge of the cargo bed). Penetration beyond 30 m is rated P4:

+---------------------------------------------------------------------------------------------------------+
|                                 ASTM F2656 PENETRATION CLASSIFICATIONS                                  |
+-------------------+-----------------------------------------+-------------------------------------------+
| PENETRATION RATING| PENETRATION DISTANCE (CARGO BED)        | OPERATIONAL APPLICATION                   |
+-------------------+-----------------------------------------+-------------------------------------------+
| P1                | <= 1.0 meter (<= 3.3 feet)              | Specify where protected assets (fuel,     |
|                   |                                         | generators) sit within a few meters       |
+-------------------+-----------------------------------------+-------------------------------------------+
| P2                | 1.01 to 7.0 meters (3.31 to 23.0 feet)  | Acceptable only if assets are clearly     |
|                   |                                         | more than 7 m behind the barrier          |
+-------------------+-----------------------------------------+-------------------------------------------+
| P3                | 7.01 to 30.0 meters (23.1 to 98.4 feet) | Acceptable only if assets are clearly     |
|                   |                                         | more than 30 m behind the barrier         |
+-------------------+-----------------------------------------+-------------------------------------------+

Critical Comparison: A barrier rated M50-P1 stops a $15,000\text{ lb}$ truck moving at 50 mph with less than $3.3\text{ ft}$ ($1.0\text{ m}$) of bed penetration. Conversely, an M50-P3 barrier stops the same truck, but permits the cargo bed to travel up to $98.4\text{ ft}$ ($30\text{ m}$) past the barrier line before halting. If an outdoor diesel fuel storage tank is located $25\text{ ft}$ behind the fence line, an M50-P3 barrier could let the truck reach and rupture the fuel tank before stopping, whereas an M50-P1 barrier stops it well short.

+---------------------------------------------------------------------------------------------------------+
|                          LEGACY DOD K-RATINGS VS. ASTM F2656 EQUIVALENTS                                |
+-------------------+-------------------+-------------------+--------------------+------------------------+
| LEGACY DOD RATING | TEST WEIGHT (LBS) | SPEED (MPH / KMH) | KINETIC ENERGY     | MODERN ASTM RATING     |
+-------------------+-------------------+-------------------+--------------------+------------------------+
| K4                | 15,000 lbs        | 30 mph (48.3 km/h)| 451,000 ft-lbs     | ASTM M30-P1            |
|                   | (6,800 kg)        |                   | (611 kJ)           |                        |
+-------------------+-------------------+-------------------+--------------------+------------------------+
| K8                | 15,000 lbs        | 40 mph (64.4 km/h)| 802,000 ft-lbs     | ASTM M40-P1            |
|                   | (6,800 kg)        |                   | (1,087 kJ)         |                        |
+-------------------+-------------------+-------------------+--------------------+------------------------+
| K12               | 15,000 lbs        | 50 mph (80.5 km/h)| 1,253,000 ft-lbs   | ASTM M50-P1            |
|                   | (6,800 kg)        |                   | (1,699 kJ)         |                        |
+-------------------+-------------------+-------------------+--------------------+------------------------+

---

Barrier Topologies: Fixed vs. Active Systems

Perimeter barrier engineering deploys a complementary mixture of passive (fixed) barriers along continuous boundary lines and active (operable) barriers at authorized vehicular checkpoints.

1. Fixed (Passive) Barriers

• Deep-Mount Stationary Bollards: Heavy structural steel pipes (often 8- to 12-inch nominal diameter, Schedule 80 or heavier) embedded $3\text{–}5\text{ ft}$ ($0.9\text{–}1.5\text{ m}$) in a reinforced concrete foundation, frequently a continuous grade beam. The pipe core is filled with concrete, sometimes with a rebar cage. Clear spacing between bollards is typically kept to about $4\text{ ft}$ ($1.2\text{ m}$) or less so passenger cars cannot slip between them. The crash rating applies only to the foundation design that was tested.
• Shallow-Mount Bollards: Developed specifically for urban data centers where underground utility ducts, fiber banks, stormwater culverts, or transit tunnels prevent deep excavation. Shallow-mount bollards require only $8\text{ to }12\text{ inches}$ ($200\text{–}300\text{ mm}$) of excavation depth, utilizing heavy interconnected subgrade steel plates that distribute crash loads across a broad concrete pad through tensile anchoring.
• Crash-Rated Cable Fencing: Engineered high-tensile steel wire ropes (typically three to four 0.75- to 1.0-inch pre-stretched galvanized cables) threaded through structural tubular posts and tied into massive concrete terminal deadman anchors. These provide cost-effective ASTM M30/M40-rated continuous protection across long rural perimeter boundaries.

2. Active (Operable) Barriers

• Active Wedge Barriers: High-strength structural steel plates hinged at ground level that raise into a triangular wedge geometry to block incoming traffic lanes. Operated by high-pressure electro-hydraulic pumping units (HPU) or electromechanical actuators. Many wedge barriers carry M50-P1 ratings and are built for continuous heavy-duty cycling at high-traffic security checkpoints.
• Retractable Rising Bollards: Cylindrical steel bollards that lower flush into underground subgrade casings to admit authorized vehicles. They preserve an open architectural aesthetic while delivering M30/M50 crash resistance.
• Emergency Fast Operation (EFO): Active barriers at high-security checkpoints should include an Emergency Fast Operation (EFO) circuit. Under standard operation, rising or lowering takes $3\text{–}5\text{ seconds}$; when triggered by an EFO panic push-button at the guard booth or automated perimeter breach alert, the barrier deploys from hydraulic or pneumatic accumulators in roughly 1 to 2 seconds (per manufacturer), preventing an adversary from tailgating an authorized vehicle through the gate.
• Drop-Arm Crash Beams and Cantilever Slide Gates: Overhead or horizontal crash-rated sliding gates spanning multi-lane driveways, reinforced with high-tensile internal steel cables and locking-pin receivers anchored to concrete buttress towers.

---

Perimeter Fencing Specifications and Clear Zones

While crash bollards mitigate vehicular threats, perimeter fencing deters and delays pedestrian intruders, trespassers, and saboteurs.

+---------------------------------------------------------------------------------------------------------+
|                               PERIMETER FENCING SPECIFICATION MATRIX                                    |
+-----------------------+-----------------------------+---------------------------------------------------+
| COMPONENT             | MINIMUM SPECIFICATION       | TECHNICAL JUSTIFICATION                           |
+-----------------------+-----------------------------+---------------------------------------------------+
| Fence Fabric Height   | 8 ft (2.4 m) minimum        | Prevents unassisted jumping / vaulting            |
+-----------------------+-----------------------------+---------------------------------------------------+
| Outrigger Extension   | 1 ft (0.3 m) at 45° angle   | Adds vertical delay; angles toward threat side    |
|                       | (3 strands barbed / razor)  | to force climber backward                         |
+-----------------------+-----------------------------+---------------------------------------------------+
| Overall Barrier Height| About 9 ft (2.7 m) or more  | Common high-security target (DoD baseline: 8 ft)  |
+-----------------------+-----------------------------+---------------------------------------------------+
| Fabric Material       | 9-gauge galvanized chain-   | Resists standard manual hand wire cutters;        |
|                       | link OR 358 welded mesh     | 358 mesh eliminates finger/toeholds               |
+-----------------------+-----------------------------+---------------------------------------------------+
| Bottom Anchoring      | Buried 12 in (300 mm) OR    | Prevents intruder from using a car jack or lever  |
|                       | secured to concrete sill    | to lift fence bottom and crawl underneath         |
+-----------------------+-----------------------------+---------------------------------------------------+
| Post Foundation       | Continuous grade beam or    | Resists wind loading and provides structural      |
|                       | concrete piers below frost  | stability for fence-mounted PIDS cables           |
+-----------------------+-----------------------------+---------------------------------------------------+

The "358" High-Security Welded Wire Mesh

In modern data center design, standard chain-link fence is frequently replaced with 358 anti-climb welded wire mesh (also known as "prison mesh"). The designation "358" originates from its precise physical dimensions:

• $3\text{ inches}$ ($76.2\text{ mm}$) horizontal wire spacing
• $0.5\text{ inches}$ ($12.7\text{ mm}$) vertical wire spacing
• $8\text{ gauge}$ ($4.0\text{ mm}$) high-tensile steel wire

The narrow $0.5\text{ inch}$ vertical aperture leaves no practical finger or toe holds for climbing and makes it hard for the jaws of standard manual bolt cutters to grip a wire. Furthermore, 358 mesh provides clear optical transparency for exterior surveillance cameras.

Perimeter Clear Zone Geometry

Perimeter fencing should be bounded on both sides by unencumbered clear zones, consistent with Crime Prevention Through Environmental Design (CPTED) principles. The widths below follow U.S. Army physical security guidance (FM 3-19.30); sites adapt them to their parcel:

1. Inner Clear Zone (Inside the Fence Line):
   • Width: $30\text{ ft}$ ($9\text{ m}$) or more.
   • Location: Between the inside face of the fence and any interior perimeter patrol road, secondary fence, or facility structures.
   • Function: Provides an unobstructed visual assessment lane for roving security patrols and pan-tilt-zoom (PTZ) surveillance cameras. Ensures that any intruder who scales the fence lands in a wide, well-lit, open assessment zone with no cover.
2. Outer Clear Zone (Outside the Fence Line):
   • Width: $20\text{ ft}$ ($6\text{ m}$) or more.
   • Location: Between the exterior face of the fence and the property boundary, adjacent brush, forest, or public parking.
   • Function: Strips away natural concealment, eliminating hiding spots for an adversary conducting pre-attack surveillance. Eliminates trees or utility poles that could act as "natural ladders" to bridge over the fence.
3. Vegetation and Grading Governance: Ground within both clear zones should be graded with gentle slopes and planted with low, non-concealing ground cover kept short (for example, under about $100\text{ mm}$ / $4\text{ in}$). Trees near the fence are removed or pruned high so they cannot conceal intruders or serve as climbing aids.

---

Perimeter Intrusion Detection Systems (PIDS)

Deploying high-security fencing without automated sensing violates the fundamental delay-detection principle. Perimeter Intrusion Detection Systems (PIDS) convert the perimeter boundary into an active sensory envelope.

+---------------------------------------------------------------------------------------------------------+
|                                 PIDS SENSOR TECHNOLOGY COMPARISON                                       |
+-----------------------+-----------------------------+---------------------------------------------------+
| SENSOR TECHNOLOGY     | OPERATIONAL MECHANISM       | PRIMARY DETECTION ATTRIBUTES & LIMITATIONS        |
+-----------------------+-----------------------------+---------------------------------------------------+
| Fence-Mounted Fiber-  | Phase-sensitive OTDR or     | • Detects cut, climb, lift, or ladder placement   |
| Optic Vibration Cable | interferometry senses light | • Immune to EMI and radio frequency interference  |
|                       | changes under strain        | • Software filters wind/rain environmental noise  |
+-----------------------+-----------------------------+---------------------------------------------------+
| Active Infrared (AIR) | Pulsed multi-beam infrared  | • Forms an invisible multi-beam optical fence     |
| Beams                 | lasers between transmitter  | • Requires line-of-sight; affected by dense fog,  |
|                       | and receiver towers         |   heavy blowing snow, and dust storms             |
+-----------------------+-----------------------------+---------------------------------------------------+
| Bistatic Microwave    | Elliptical volumetric RF    | • High volumetric coverage across clear zone      |
| Sensors               | field between transmitter   | • Detects crawling, rolling, and running intruders|
|                       | and receiver units          | • Susceptible to standing water / weed growth     |
+-----------------------+-----------------------------+---------------------------------------------------+
| Buried Seismic / Ported| Ported coax sets up a buried| • Covert, invisible surface presence              |
| Coaxial Cables        | RF field; geophones sense   | • Detects people crossing; seismic types detect   |
|                       | footfall vibration          |   footfalls and digging; needs stable soil        |
+-----------------------+-----------------------------+---------------------------------------------------+
| Thermal Cameras with  | Long-wave infrared (LWIR)   | • Operates in 0-lux total darkness and smoke      |
| Video Analytics       | analytics classify humans,  | • Automatically slews PTZ cameras to target       |
|                       | animals, and vehicles       | • Provides real-time visual assessment to the SOC |
+-----------------------+-----------------------------+---------------------------------------------------+

Sensor Fusion and Nuisance Alarm Rate (NAR) Mitigation

A major operational failure of standalone PIDS is a high Nuisance Alarm Rate (NAR) caused by wind buffeting fences, heavy rain, or wildlife, which leads to operator alarm fatigue and eventual system deactivation.

Good practice uses multi-sensor fusion: an alarm is only escalated to the SOC if two independent sensory technologies correlate within a specified spatial and temporal window (e.g., a fence vibration sensor alert confirmed by an optical beam break or thermal analytics target classification).

---

Practical Design Scenario: Designing an ASTM M50-P1 Vehicle Checkpoint

Scenario Parameters

A newly planned Class F4 cloud data center campus is situated on a semi-rural parcel adjacent to a county highway with a posted speed limit of 55 mph ($88.5\text{ km/h}$). The site plan features a straight, $400\text{ ft}$ ($122\text{ m}$) entry driveway leading directly from the highway to the main vehicle security gatehouse.

Directly behind the gatehouse, at a distance of only $25\text{ ft}$ ($7.6\text{ m}$), sits the facility's primary outdoor bulk diesel fuel storage depot ($60,000\text{ gallons}$ of fuel supply for backup generators).

Engineering Threat Assessment

1. Kinetic Energy Calculation: An attacking medium-duty truck ($15,000\text{ lbs} / 6,800\text{ kg}$) entering the straight driveway can accelerate to at least 50 mph ($80.5\text{ km/h} = 22.35\text{ m/s}$) prior to impacting the barrier gate: $m = \frac{15,000\text{ lbs}}{32.174\text{ ft/s}^2} = 466.2\text{ slugs} \quad (6,800\text{ kg})$ $v = 50\text{ mph} = 73.33\text{ ft/s} \quad (22.35\text{ m/s})$ $KE = \frac{1}{2} m v^2 = \frac{1}{2} \times 466.2 \times (73.33)^2 = 1,253,500\text{ ft-lbs}$ $KE_{metric} = \frac{1}{2} \times 6,800\text{ kg} \times (22.35\text{ m/s})^2 \approx 1.70\text{ MegaJoules (MJ)}$
2. Barrier Penetration Analysis: If the engineering team specifies a standard ASTM M50-P3 crash gate:
   • An M50-P3 rating permits the vehicle cargo bed to travel up to $98.4\text{ ft}$ ($30.0\text{ m}$) past the barrier before coming to rest.
   • Because the bulk diesel fuel depot is only $25\text{ ft}$ ($7.6\text{ m}$) behind the gate line, a P3 barrier could let the truck travel as much as $73.4\text{ ft}$ beyond the tanks' position, so it could strike and rupture the tanks, causing a major fuel spill and fire.
3. Remediation Specification:
   • The gate barrier should be specified as ASTM F2656 M50-P1 (maximum penetration $\le 1.0\text{ m} / \le 3.3\text{ ft}$). A P2 rating (up to $7.0\text{ m}$) would leave only about $0.6\text{ m}$ of margin before the tanks.
   • An active wedge barrier with an Emergency Fast Operation (EFO) cycle of about 1 to 2 seconds is deployed across the inbound traffic lane.
   • Flanking the active wedge barrier, deep-mount fixed steel bollards rated M50-P1 are embedded into a continuous reinforced concrete grade beam on $3.5\text{ ft}$ centers.
4. Roadway Geometry Mitigation (Chicane Design): Rather than maintaining the straight $400\text{ ft}$ drag-strip approach, the civil design is modified to incorporate a curvilinear serpentine chicane with heavy concrete curb islands, tight $90^\circ$ radius bends, and passive decorative boulder obstacles. This forces approaching vehicles to reduce speed to under $18\text{ mph}$ ($29\text{ km/h}$), reducing the maximum achievable kinetic energy by over 87% prior to barrier impact: $KE_{18mph} = KE_{50mph} \times \left(\frac{18}{50}\right)^2 = 1.70\text{ MJ} \times 0.1296 \approx 0.22\text{ MJ}$

---

Common Exam Traps & DCDC Pitfalls

• Conflating P1 and P3 Penetration Ratings: On certification exams, candidates often assume P3 is superior to P1 because "3 is higher than 1." In ASTM F2656, P1 is the most stringent penetration rating ($\le 1.0\text{ m}$), whereas P3 allows up to $30.0\text{ m}$ ($98.4\text{ ft}$) of penetration. Specifying P3 where assets sit close behind the barrier is a serious design flaw.
• Ignoring Standoff Scaling Physics: Forgetting that blast pressure falls non-linearly (roughly $1/R^3$ in the near field). Doubling standoff distance does not simply halve blast pressure; near the charge it can cut overpressure by up to a factor of about 8.
• Treating Bollards as Isolated Posts: Installing crash bollards on foundations that differ from the crash-tested design, such as shallow, isolated post holes in place of the tested grade beam or footing. A crash rating applies only to the tested configuration; an under-designed footing lets a $15{,}000\text{ lb}$ truck lever the post out of the ground.
• Vegetation Inside Clear Zones: Permitting decorative landscaping or mature trees inside the clear zones (about 30 ft inside and 20 ft outside the fence). Trees obstruct automated CCTV analytics, introduce thermal shadows, and act as natural scaling ladders over fences.