What is a star network?
A star network is a local area network (LAN) topology in which all nodes -- personal computers (PCs), workstations or other devices -- are directly connected to a common central computer that is often referred to as a hub. Therefore, a star network is often referred to as a hub-and-spoke network topology.
Every workstation connected to the hub is indirectly connected to each workstation using the hub as an intermediary device. Star networks are typically deployed at the access layer of enterprise networks. The access layer uses a centralized network switch to connect all endpoints to the rest of the LAN.
This graphic shows a star network consisting of a central hub. Each workstation is shown as a PC or laptop with the central hub interconnecting all devices. The lines that interconnect endpoint PCs and laptops are the spokes of the star topology. In the real world, a star network can consist of either wired or wireless connections. In this specific case, each spoke is connected to a hub by a wired connection.
How does a star network work?
Because a star network uses a centralized hub, that hub is responsible for controlling communications between devices. However, there are different ways that a central hub can manage these communications. For example, an Ethernet hub is a network device that simply listens to a communication destined for a device on a different spoke and then retransmits -- or broadcasts -- the message out to all spokes. This is the simplest form of a network hub, as it only must repeat the message to all other connected spokes. But this method can become inefficient quickly, as each communication is sent out to all spokes, as opposed to only the spoke the message was intended for.
If too many devices begin communicating on a network hub, the amount of broadcast traffic can quickly reduce network throughput. It also places the Ethernet hub where it physically looks like a star network topology but operates like a traditional bus network topology.
An Ethernet switch, on the other hand, may look like an Ethernet hub from a physical cabling perspective, but it is far more sophisticated when it comes to how the centralized device handles the transmission of communications to the intended spoke device. Ethernet switches eliminate the need to broadcast communications out to all spokes on the star network. Instead, the Ethernet switch maintains a media access control (MAC) address table. This table statically or dynamically maps the physical MAC address to the port or spoke where the spoke endpoint resides. Therefore, if an Ethernet switch knows the MAC address and specific spoke that the MAC address lives on, it can use this information to send a communication directly out to a single spoke, as opposed to broadcasting the communication to all spokes -- unnecessarily using network bandwidth. Ultimately, an Ethernet switch accomplishes the same goal as an Ethernet hub with the added benefit of better network transport efficiency.
Comparing star network topologies with other topologies
The star network topology works well when workstations are deployed randomly throughout a building or facility. With the hub-and-spoke design, it is easy to add or remove workstations, as all cabling is pulled and connected to a central hub.
From a cabling perspective, if the workstations are reasonably close to the vertices of a convex polygon and the system requirements are modest, the ring network topology may serve the intended purpose at a lower cost than the star network topology. If the workstations lie nearly along a straight line, the bus network topology may be best.
In a star network, a cable failure on a single spoke will only affect the spoke endpoint that it links to the central computer. All the other workstations will continue to function normally with the exception that they will not be able to communicate with the device that resides on the failed spoke. In other words, a star network is powerful from this perspective, as a failure on one spoke does not affect communications of other spokes that are in a functioning state. However, the caveat is that, if the central hub were to fail, all spokes on the star network also fail. If any workstation goes down, none of the other workstations will be affected. Therefore, if network redundancy is required, a mesh network topology may be a preferable option.