Star Topology

Part of Networking

The dominant network architecture: every device connects to a central switch or hub.

Why This Matters

The star topology — where every device has its own direct connection to a central switch — is the architecture used in virtually every modern Ethernet network. Understanding it deeply means understanding not just where cables go, but why this arrangement is better than alternatives, what its failure modes are, and how to design star-topology networks that are reliable, scalable, and maintainable.

Star topology replaced earlier bus and ring topologies not just because it is more reliable, but because it enables centralized management and diagnostic visibility that other topologies cannot provide. When something goes wrong in a star topology, the problem is usually localized to one link — one device or one cable — and the switch’s port statistics often reveal exactly what is wrong without requiring physical inspection.

Why Star Topology Dominates

The bus topology (all devices sharing one cable) has three critical weaknesses: any break in the cable disrupts all devices simultaneously; all devices share a collision domain, limiting throughput as device count grows; and adding or removing a device requires modifying the shared cable.

Star topology eliminates all three weaknesses. Each device has its own dedicated cable to the central switch, so a cable break affects only that device. Switches give each port its own collision domain, and modern switches run all ports in full-duplex mode, eliminating collisions entirely. Adding a device requires only plugging in a new cable to an available switch port.

The central switch also provides management capabilities impossible in a bus or ring: per-port traffic statistics, port enable/disable, VLAN assignment, and link speed/duplex configuration. These capabilities make the switch the operational center of the network — the place where problems are first detected and many solutions are applied.

Designing a Star Topology Network

The design process for a star-topology network begins with a clear picture of what must be connected and where connections terminate.

Start with a site plan. Mark every location where a network connection is needed: workstations, printers, servers, network cameras, door access controllers, wireless access points. Each of these locations needs at least one cable run to a wiring closet.

Identify where wiring closets will be located. A wiring closet is where cables from multiple locations terminate, connecting through a patch panel to switches. The rule of thumb is that no cable should exceed 90 meters in length (leaving 10 meters for patch cables at each end, for a total channel length of 100 meters). This limits the service radius of each wiring closet. Buildings with a small footprint can be served by a single wiring closet; larger buildings require multiple wiring closets connected by uplinks.

Plan switch capacity. Count the number of ports needed in each wiring closet. Add 20-30% spare capacity for future growth. Select switches that provide enough ports, with enough uplink capacity to connect to other switches if needed.

Wiring Closets and Patch Panels

A wiring closet (also called a telecommunications room or MDF/IDF) houses the patch panels, switches, and other active equipment at the center of the star. It needs: adequate space for current equipment plus future expansion, electrical power with surge protection and UPS (Uninterruptible Power Supply), cooling (active equipment generates heat), physical security (unauthorized access to the wiring closet can compromise the entire network), and cable management.

Patch panels are passive devices that terminate the permanent cable runs from device locations. Each port on the patch panel corresponds to one cable run; the far end of that cable is a wall plate in the device location. Short patch cables connect patch panel ports to switch ports, making it easy to change which device connects to which switch port without touching the permanent cable runs.

Cable management — cable trays, cable ties, and labeled paths — keeps the wiring closet organized. An organized wiring closet enables quick identification of specific cables, easier troubleshooting, and cleaner airflow for cooling. An unorganized wiring closet (a “cable spaghetti” situation) makes changes extremely difficult and troubleshooting nearly impossible.

Uplinks and Cascaded Stars

A single switch serving a small network is a simple star. Larger networks use cascaded stars: multiple edge switches, each serving a cluster of devices, connected by uplinks to a higher-level switch. This hierarchical structure maintains the star topology properties at each level while scaling to large numbers of devices.

Uplinks (connections between switches) should use faster cables or fiber, because they carry traffic from multiple downstream devices. If each access switch serves 24 devices at 1 Gbps, its uplink to the distribution switch should provide at least 1 Gbps — preferably more. Oversubscribed uplinks create bottlenecks.

Link aggregation (LACP, 802.3ad) combines multiple physical links into one logical link, multiplying bandwidth and providing redundancy. Two 1 Gbps uplinks aggregated provide 2 Gbps bandwidth and failover if one link fails. This is the standard approach for uplinks that need more bandwidth than a single cable provides.

Redundancy and Spanning Tree

A star topology has a single point of failure at the center: if the switch fails, all connected devices lose connectivity. For networks where this is unacceptable, redundancy requires a second switch.

Dual-homing connects critical devices to two separate switches. If one switch fails, the device’s connection to the other switch remains active. The device uses both connections simultaneously (with link aggregation) or keeps one in standby (with Spanning Tree).

Spanning Tree Protocol (STP) prevents loops in redundant topologies. A loop in the network (where traffic can circulate indefinitely) creates a broadcast storm — broadcast frames circulate forever, consuming all bandwidth. STP detects loops and blocks redundant links, creating a loop-free logical topology. When a primary link fails, STP unblocks the previously blocked redundant link, restoring connectivity.

Modern variants like Rapid Spanning Tree Protocol (RSTP) converge much faster than the original STP (seconds versus potentially minutes), making failover practically transparent to users. MSTP (Multiple Spanning Tree Protocol) allows different VLANs to use different spanning tree topologies, enabling load balancing across redundant links.

Fault Isolation and Diagnostics

The star topology’s greatest operational advantage is fault isolation. When a device loses connectivity, the problem is almost certainly in its dedicated link to the switch — one cable or one switch port. Diagnosing a bus network problem could require inspecting every device on the segment; diagnosing a star network problem requires checking exactly one cable and one switch port.

Switch port statistics (accessible through the switch’s management interface) show per-port frame counts, error rates, and link status. A port showing high CRC error rates has a cable or connector problem on that link. A port showing no received frames but an active link LED indicates the connected device is not transmitting. A port that flaps (repeatedly goes up and down) indicates an intermittent cable fault.

The management port on the switch (typically accessible via web interface or SSH) provides all this diagnostic information. Making a habit of checking switch port statistics when investigating problems — before pulling cables or rebooting devices — often reveals the exact fault immediately, saving significant troubleshooting time.