The first and most simple consideration of network design is the quality of the equipment used. 'Industrial Ethernet' switches have high MTBFs and better build quality than off the shelf consumer switches and are often much more suited to mission-critical applications. Whilst the cost of Industrial Ethernet may be a little higher in the beginning, the total cost of a complete IP CCTV system may be greatly reduced if repeated site visits to rectify network faults are rendered unnecessary.
Classic star topology
A typical Ethernet network employs a star topology as detailed in Figure 1 below. Using off-the-shelf switches it is possible to build networks such as these quite cost-effectively.
However, it is often desirable or essential (look in the tender specification!) to build a degree of additional reliability into the network. This increased reliability is required in an IP surveillance system where down-time of the CCTV system is deemed unacceptable. This is especially true of critical security applications in airports, on the roads and on the rail network.
There are several mechanisms for providing redundancy in a network, the most common being LACP, RSTP and ring redundancy. All switches must support the required redundancy capability and be connected to each other in an appropriate configuration. Due to their simplicity and ease of expansion, redundant rings are becoming the most popular redundancy solution for IP video surveillance systems.
Redundant Ring topology
Using specialist Ethernet products - often referred to as 'Industrial Ethernet switches' - it is possible to build a ring topology that can suffer a single point failure (cable broken or connector unplugged) and continue to operate as if nothing had changed. Comparing Figure 1 and Figure 2 below, the difference between a star and a ring topology is clearly visible. It is not possible to build a redundant ring with off the shelf unmanaged switches - this can cause 'broadcast storms' resulting in network failure.
How does ring redundancy work?
Consider Figure 2. The network traffic from camera A follows the default path through the network from switch 2 to switch 1 to switch 4 to the destination - the server. If a cable is broken or a connector un-plugged along this default route, the switches will automatically reconfigure themselves to transmit data the alternative way around the ring - switch 2 to switch 3 to switch 4 to server. This network reconfiguration can happen in as little as 300mS. Poorly designed networks can have a much higher re-configuration time with video disappearing from the screen for 10s of seconds rather than less than 1 second.
Redundant Ring topology with dual attached server
Figure 2 shows a network with redundancy at the 'network layer' but no redundancy at the 'device layer'. If the single link from the server to the network were broken, all IP cameras would be lost. Given that we are investing time to build a redundant ring, the next logical step is to dual attach the server to the network removing the single point of failure associated with the server. Dual attachment of Ethernet devices is not always a trivial process. Amplicon's technical team can help with such requirements. Figure 3 shows the topology with dual attached server.
Redundant Ring topology with all devices dual attached
In highly critical IP surveillance applications, it may be necessary to dual attach all devices to the network. This gives the maximum reliability but also generates maximum cost and complexity.
Amplicon can advise on all aspects of security automation solutions. If you would like more information on our IP video surveillance or IP CCTV systems, please call sales on 01273 570 220 or email security@amplicon.co.uk.
Back to top 
