The network switch is the system as far as the traffic is concerned. Every camera frame, every reader event, every intrusion signal goes through it. The wrong switch becomes the limiting factor on the install’s reliability; the right switch with a managed configuration becomes invisible infrastructure that does its job for ten years. Pick managed switches at every layer, size the PoE budget against the actual load, segregate traffic with VLANs, harden the configuration at commissioning, and the network layer holds up for the institution’s planning horizon.
Switch class by role
When the rule applies
Every switch on the project. The role determines the class: access-layer switches in IDFs serve devices, distribution switches aggregate traffic between IDFs, core switches anchor the network. The institutional default is enterprise-grade managed switches at every layer; consumer or SOHO switches do not appear on production installs.
The spec
Field note
PoE budget sizing
When the rule applies
Every PoE-capable switch on the install. The PoE budget is the sum of every port’s draw across the switch; exceed it and the switch cuts power to lower-priority ports in port order until the budget balances. Design at 75 percent of the published budget and the system has headroom for inrush, for incidental loads, and for the next device the institution adds.
The spec
Worked example
VLAN segmentation
When the rule applies
Every managed switch on the install. VLAN segmentation isolates traffic types onto separate broadcast domains, providing both security and performance benefits.
The spec
Field note
Configuration baseline
When the rule applies
Every switch starts from a documented configuration baseline at commissioning. The baseline is the hardening standard the institution expects on every device on its network.
The spec
Field note
Redundancy and high availability
When the rule applies
Switches in head-end, recording, and critical IDF locations on projects where the institutional design calls for high availability. The two main redundancy patterns are stacking (multiple physical switches operating as one logical switch) and MC-LAG (multi-chassis link aggregation for redundant uplinks).
The two patterns
Stacking
Two or more switches connected with a stacking cable, presenting as a single logical switch with one IP address, one configuration, and shared uplinks. Stack supports hot-swap of failed stack member. Used at the access layer where one failed switch should not bring down the IDF.
MC-LAG (Multi-Chassis Link Aggregation)
Two independent switches sharing an LACP-bonded link to a single downstream device. The downstream device sees a single LAG even though the two uplinks terminate on two different switches. Used between distribution and access where uplink redundancy matters but full stacking is not warranted.
The spec
Industrial and outdoor switches
When the rule applies
Switches installed in environments outside the typical office or equipment-room conditions: outdoor IDFs, transit stations, industrial buildings, parking structures, environmental cabinets.
The spec
Field note
Network management and monitoring
When the rule applies
Every managed switch on the install. The switches integrate with the institutional Network Management System (NMS) for monitoring, alerting, configuration backup, and capacity planning.