Chapter 4: Architecture Design

4.1 Typical System Topology

The typical underground parking surveillance system topology is organized as a five-layer hierarchy, with clearly defined data flows, redundancy paths, and power dependencies at each layer. The design prioritizes recording continuity and evidence integrity as the primary non-negotiable requirements, with all redundancy decisions made to protect these outcomes.

At the field layer, cameras are grouped into functional zones — entrance/exit, ramps, aisles, intersections, bay areas, lobbies, booths, and equipment rooms. Each zone connects to a dedicated PoE access switch housed in a local distribution cabinet. Access switches uplink via dual fiber paths to floor-level distribution switches, which connect via independent fiber routes to a dual-stack core switch in the main equipment room. This dual-path fiber design ensures that a single cable cut or switch failure cannot isolate an entire floor's cameras from the recording platform.

Figure 4.1: Typical System Topology — Five-Layer Architecture with Redundant Uplinks, HA Platform, and UPS Power Paths

Redundancy Design Principles

• Dual uplinks: Every access switch has two fiber uplinks to the distribution switch via physically separate cable routes (different conduit paths)
• Core redundancy: Core switches are deployed as a stack or MLAG pair with dual power supplies; failover time <50ms
• Recording redundancy: Either N+1 NVR configuration or VMS HA pair with shared storage; cameras automatically rebind to standby node on primary failure
• UPS coverage: Core switch, storage array, and entrance switches/cameras are all on UPS-backed circuits; minimum 30-minute runtime for core, 15 minutes for edge
• Time synchronization: Dual NTP sources with monitoring; all devices must maintain offset <500ms from NTP server

4.2 Core Functions and Performance Indicators

The following table defines the twelve core performance indicators for the surveillance system, each with its operational impact, implementation path, and acceptance verification method. These indicators form the basis of the acceptance test plan and should be measured during both commissioning and periodic O&M audits. Failure to meet any indicator in the "Critical" category constitutes a system defect that must be resolved before handover.

4.3 Device Connection Diagram

The device connection diagram illustrates the physical wiring path from each camera through the structured cabling system to the platform layer, including all intermediate components, connection types, and grounding requirements. Every connection in this path has a specific requirement that, if not met, creates a predictable failure mode — from water ingress at unsealed glands to recording gaps from undersized fiber uplinks.

Figure 4.2: Device Connection and Wiring Diagram — Camera to Platform Path with Grounding, UPS, and Integration Interfaces

Critical Wiring Requirements

• Drip loop: All camera cable entries must form a downward drip loop before entering the waterproof junction box to prevent water from running into the housing along the cable
• IP67 glands: Cable glands must match the cable outer diameter (OD) precisely; oversized glands allow moisture ingress even with IP67-rated housings
• Conduit fill: Metal conduit preferred in underground environments; fill ratio <40% to allow future cable additions and heat dissipation
• Fiber routing: Dual fiber uplinks (A/B paths) must be routed through physically separate conduits and cable trays to prevent a single incident from cutting both paths
• Grounding: All equipment cabinets must be bonded to the building equipotential grounding bus; ground resistance <4Ω; surge protection devices (SPD) installed at all outdoor cable entries
• Cable length: Cat6 copper runs must not exceed 90m (structured cabling) + 10m patch cords; longer runs require fiber media converters

4.4 Business Logic and Exception Handling

The system's business logic defines the normal data and control flows, the switching and recovery behaviors during failures, and the exception handling chains for the three most critical failure scenarios. Understanding these flows is essential for both commissioning verification and O&M response planning.

Normal Operation Flows

• Data flow: Main stream recorded continuously; sub stream for live preview; metadata (plate events, motion, analytics) indexed in VMS database
• Control flow: Alarms trigger camera pop-up in VMS, PTZ preset moves (if applicable), and priority recording mode activation
• Evidence flow: Operators search by time/camera/zone/plate; export evidence package with cryptographic hash; hash recorded in audit log