Network Topologies: Ring vs. Star
When it comes to network design, it is important to understand the benefits and limitations of different topologies, particularly the use of ring and star topologies.
Ring Topology
Within a ring, each switch connects to adjacent switches on either side creating a ring or loop. This provides the benefit of adding redundancy as there are two paths to each switch but has the drawback of requiring a protocol such as Spanning Tree Protocol (STP) or Ethernet Protection Switching Rings (EPSR) to prevent broadcast storms from occurring.
In addition, when a change or failure occurs in the network, switches must flush their internal tables and relearn the topology of the network which can result in disruption to services such as video and IP telephony. This disruption is particularly noticeable with STP because the protocol takes some time to recover. Ring protocols like EPSR are specifically designed to provide fast recovery to limit disruption for real-time services such as voice and video.
Star Topology
Within a star topology, the core switch connects to every other switch creating a star. In order to deliver redundancy within a star topology, two links to each switch are required which are then aggregated using link aggregation. As the core switch represents a single point of failure, it is important that a second switch be added and ideally, this would use a chassis stacking technology like Virtual Chassis Stacking (VCS) to enable both switches to be managed as a single logical switch.
Star topologies have the benefit of no disruption to service in the event of a failure but require more interfaces than a ring topology. Depending on a number of factors including speed and type of fiber used, there may be a cost impact.
Ring vs Star: which is better?
It is important to understand that it is not a simple case of one topology being better but ensuring the solution meets the needs of the client and their applications. Generally speaking, a star topology will deliver a more resilient network than a ring topology but does require more interfaces. As there is dedicated bandwidth to each switch, the bandwidth requirements of the links within a star topology will be lower than that of a ring. This means a star topology while having more interfaces can actually result in lower cost solution than a ring topology.
It is also important to understand that a star topology can be delivered over cabling infrastructure that is deployed as a ring to provide the resilience of a star topology with the path diversity of a ring. This is illustrated in the diagram shown below.
Where a ring topology is used, it is important that an advanced protection protocol like EPSR is used for rapid recovery to ensure minimal disruption to services in the event of a link or device failure.
