Posted on Leave a comment

How to Daisy Chain PoE Switches Properly?


As the number of IoT devices increases, more network ports are required and we need to expand our network accordingly. When you connect multiple switches, you can get more ports that can be used, which means more devices can be utilized over the same network. The only question is How? There are a lot of ways to connect PoE switches. You can daisy chain them, star link them, cascade them, cluster them, or stack them. In today’s post, we will teach you how to daisy chain multiple PoE switches properly.

What is Daisy Chain Topology?


Daisy chain is a type of network topology in which PoE switches are connected in a linear structure. In a daisy chain topology, the communication is typically uni-directional, which means data transmission can only flow in one direction. It’s relatively easy to install and configure as it only requires a single cable to connect all PoE switches. This simplicity makes it a popular choice for small networks. Moreover, daisy chain is a cost-effective solution as it requires minimal hardware investment. It cleans up space by reducing the complexity with limited wiring. Besides, the daisy chain system is highly scalable, meaning that it can be extended as per the network requirements.



Why Daisy Chain Network Fails?


However, as the network size grows, the daisy chain topology becomes increasingly complex to manage, and the chances of a single-point failure significantly increase. A single node failure or interruption can break the entire data transmission chain, leading to system failures and downtime. In addition, security could be a concern in a daisy chain topology, as data transmission is ubiquitous, and any device can access the network by simply plugging it into a free port. Moreover, if the core switch wants to communicate with the PoE switch located at the end of the chain, the data has to pass through each intermediate node until it reaches its destination. Consequently, the data transmission speed can significantly slow down due to latency and packet drops.



How to Daisy Chain PoE Switches Properly?


In the daisy chain topology, there are two types of networks: linear daisy chain and ring daisy chain. A linear daisy chain is a series of switches connected one after the other. But once one PoE switch fails, the others will also be affected. In a ring daisy chain, PoE switches are connected in a closed loop (switch A→switch B→switch C→switch D→switch A), which can be more resilient since data can flow in both directions around the loop, providing multiple pathways between connected devices, which enables quick failover in the case of link failure. In the linear daisy chain, if switch B fails, all the succeeding switches will be affected, and you need to take down the entire system to perform maintenance. However, by creating a redundant fiber link between switch D and switch A in the ring daisy chain using a pair of BiDi SFP transceivers, even if the second PoE switch fails, data will simultaneously be re-directed to the other route (switch A→switch D→switch C). Unlike common SFP transceivers, BiDi SFP transceivers only have one port which uses an integral WDM coupler to transmit and receive signals over a single strand fiber, which allows the data to flow the other way.



How to Solve the Loop Issues in Daisy Chain?


However, setting up a ring daisy chain requires a bit more configuration and can be more challenging to manage. And when a network is daisy-chained, it’ll introduce loop issues that can lead to network performance problems. A loop is when data transmits around the network repeatedly. If a PoE switch fails to route data properly or if there is a misconfiguration, it can cause a loop that results in duplicate packets, network congestion and poor performance. To prevent loop issues in a ring daisy chain network, it is essential to properly configure each switch’s routing information to ensure that packets are routed efficiently and without creating a loop. The use of STP or equivalent protocols like ERPS can be effective in preventing loop issues by blocking redundant paths.



ERPS creates a logical ring topology that allows for data to be transmitted in both directions without forming a physical loop. It works by creating two logical rings, one known as the primary ring and the other as the secondary ring. Each ring has a set of dedicated forwarding and blocking ports that ensure the packet is forwarded only in the correct direction. In case of a link failure, ERPS detects the failure and switches the active forwarding path to the secondary ring, avoiding any loops that may have been caused due to the failure. Compared with STP and RSTP (recovery time≅2s), ERPS also has a faster convergence time of under 50 msec, which greatly improves network resilience.



Fastcabling has also launched a series of L2+ industrial fast ring managed network switches that support a fast recovery time of less than 20 msec, making them ideal for industries such as surveillance systems, automated production lines, mining, oil and gas, and waste management. Moreover, with ERPS, the network topology can be easily and rapidly changed to accommodate growth and change in network needs. It also enhances network security and reduces the risk of cyber-attacks by ensuring data integrity and confidentiality.



Another solution is to use Token Ring. Token Ring is a communication mechanism that creates a ring topology for data signals to flow in one direction at a time. A token is passed around the ring, giving the device holding the token the right to send data. If no device is ready to send data or if the data is already being transmitted, the token is passed along to the next device in the ring. This mechanism avoids the creation of loops because the tokenized communication process only allows one device to transmit data at any given time.



Final Thoughts


In conclusion, daisy-chaining multiple PoE switches can be a cost-effective way to extend your network, but there are also some problems like single-point failures, loop issues, and reduced performance. But such problems can be properly settled if you use the right products (ERPS-supported ones) and the right topologies as per your network requirements.

Leave a Reply