Physical Network Topologies

While bus and star topologies dominated early Ethernet networks, a fundamentally different design emerged from IBM's research laboratories: ring topology. In this architecture, every node connects to exactly two other nodes, forming a closed circular path through which data flows in an orderly, predictable manner.

Ring topology introduced concepts foreign to contention-based networks: deterministic access, guaranteed maximum latency, and fair bandwidth distribution. These properties made ring networks the choice for applications requiring predictable performance—manufacturing control systems, financial trading floors, and mission-critical enterprise networks.

Ring topology arranges nodes in a closed loop where each device connects to exactly two neighbors. Data travels around this ring, passing through each node in sequence until reaching its destination or returning to its origin.

Physical vs. Logical Ring:

It's crucial to distinguish between physical ring and logical ring implementations:

Ring Network Components:

The token is a special frame that circulates continuously around the ring when no data is being transmitted. Only the station holding the token may transmit data—this simple rule eliminates collisions entirely and provides deterministic access.

Token Structure (IEEE 802.5 Token Ring):

| Starting Delimiter | Access Control | Ending Delimiter |
|       1 byte       |    1 byte      |      1 byte      |

The Access Control byte contains:

• Priority bits (3) — Token priority level
• Token bit (1) — 0 = free token, 1 = busy token (frame)
• Monitor bit (1) — Used by active monitor for loop prevention
• Reservation bits (3) — For priority reservation

Token Passing Operation:

Source Stripping vs. Destination Stripping:

Token Ring uses source stripping—the originating station removes its own frame after it circulates back. This provides:

• Acknowledgment — Source sees if destination successfully received the frame
• Loop prevention — Frames can't circulate indefinitely
• Bandwidth efficiency — Only one frame per token cycle

Some ring protocols use destination stripping where the recipient removes the frame. This allows spatial reuse—the same ring segment can carry another frame simultaneously—but complicates acknowledgment.

Ring networks have precise timing requirements that directly affect performance. Understanding these parameters reveals why rings behave fundamentally differently from contention-based networks.

Key Timing Parameters:

Ring Latency Calculation:

The ring must be able to contain at least one complete token (24 bits for Token Ring). Ring latency comes from:

1. Propagation delay — Signal travel time through cables (approximately 5 μs per kilometer)
2. Station latency — Processing delay at each station (typically 2.5 bit times per station)

Ring Latency (bits) = (Cable Length × Propagation Delay × Bit Rate) + (Stations × Station Latency)

Example Calculation:

For a 16 Mbps Token Ring with 50 stations and 500 meters total cable:

• Cable delay: 500m × 5μs/km = 2.5 μs = 40 bits at 16 Mbps
• Station delay: 50 stations × 2.5 bits = 125 bits
• Total ring latency: 165 bits

Since the token is only 24 bits, the ring can hold multiple tokens—but Token Ring uses a single-token protocol.

Efficiency Analysis:

Ring efficiency depends on the ratio of frame size to ring latency. With:

• Large frames → High efficiency (most of the time spent sending useful data)
• Small frames → Lower efficiency (overhead for token passing dominates)

Ring Efficiency Formula:

Efficiency = Frame Time / (Frame Time + Ring Latency)

For a 4000-bit frame on our example ring:

• Frame Time: 4000 bits / 16 Mbps = 250 μs = 4000 bit times
• Ring Latency: 165 bit times
• Efficiency: 4000 / (4000 + 165) = 96%

Unlike CSMA/CD where efficiency degrades as stations increase (more collisions), ring efficiency is independent of station count—only ring latency matters.

Rings can be designed for data flow in one direction (unidirectional) or both directions (bidirectional). Each approach offers different trade-offs in performance, fault tolerance, and complexity.

Unidirectional Ring:

The simpler design—data flows in one direction only (typically clockwise when viewed from above).

Bidirectional Ring:

Data can flow in either direction, with intelligent routing choosing the shorter path.

Dual Ring Architecture (Counter-Rotating):

FDDI (Fiber Distributed Data Interface) uses a dual-ring structure with counter-rotating rings:

• Primary Ring — Carries normal data traffic (clockwise)
• Secondary Ring — Rotates counter-clockwise; used for redundancy

Under normal operation, only the primary ring carries data. When a fault occurs:

1. Stations adjacent to the fault detect the break
2. They 'wrap' the primary ring around to the secondary ring
3. Data that would have crossed the break now loops back on the secondary ring
4. The logical ring is maintained, though with increased latency

Ring topology faces a fundamental challenge: since every station is an active repeater, any station failure could break the ring. Ring networks developed sophisticated fault detection and recovery mechanisms to address this vulnerability.

Types of Ring Faults:

Active Monitor Functions:

One station on the ring is elected as the Active Monitor (AM), responsible for:

1. Ring Recovery — Purging the ring and issuing a new token after faults
2. Token Management — Ensuring exactly one token circulates
3. Circulating Frame Detection — Setting and checking the monitor bit to detect orphaned frames
4. Ring Latency Buffer — Providing a 24-bit delay if needed to ensure the ring can hold a complete token
5. Neighbor Notification — Periodically announcing its presence so standby monitors can take over if needed

Standby Monitor Functions:

All other stations serve as Standby Monitors (SM), which:

1. Monitor the AM — Detect if the active monitor fails
2. Participate in Elections — Highest MAC address wins AM election
3. Report Errors — Notify network management of detected problems

MAU Bypass Mechanism:

In Token Ring's star-wired physical implementation, the MAU (Multistation Access Unit) provides automatic bypass:

• Each MAU port contains a bypass relay
• When a station is active, it keeps the relay energized, connecting the station to the ring
• If the station powers off or fails, the relay de-energizes
• Spring-loaded contacts close, bypassing the station and maintaining ring continuity

This mechanism provides approximately 5 nines (99.999%) of ring availability for station failures in Token Ring networks.

Two major ring topology implementations dominated enterprise networking before Ethernet's triumph: IBM's Token Ring and ANSI's FDDI (Fiber Distributed Data Interface). Understanding both reveals how ring concepts scaled from departmental networks to campus backbones.

Token Ring Specifics:

Developed by IBM and standardized as IEEE 802.5, Token Ring was the dominant IBM-shop LAN technology:

• Deterministic Performance — Guaranteed access made it suitable for manufacturing and financial applications
• IBM Integration — Tight integration with SNA (Systems Network Architecture) and mainframe environments
• Higher Cost — Token Ring NICs and MAUs cost 2-3x Ethernet equivalents
• Priority System — Built-in QoS with 8 priority levels
• Extensive Management — Ring poll, soft errors, and detailed diagnostics

FDDI Specifics:

FDDI extended ring concepts to higher speeds and longer distances:

• 100 Mbps Speed — 10x faster than Token Ring, equivalent to Fast Ethernet
• Fiber Optic Medium — Immunity to electrical interference, longer distances
• Dual Ring Design — Inherent redundancy for backbone reliability
• Timed Token Rotation — More flexible than Token Ring's simple priority
• CDDI Variant — FDDI over copper (Category 5) for shorter runs

Ring topology offered unique characteristics that made it ideal for certain applications while limiting its broader adoption. Understanding these trade-offs explains both its historical importance and its eventual displacement.

Why Ethernet Won:

Despite ring topology's technical merits, Ethernet's star-switched architecture ultimately dominated:

Full-duplex switched Ethernet achieved what ring topology promised—collision-free, deterministic operation—but with simpler protocols, lower costs, and continuously increasing speeds.

While Token Ring and FDDI installations are rare today, ring topology concepts persist in modern networking and continue to influence network design.

Where Rings Still Exist:

Concepts That Transferred:

Ring topology's innovations influenced broader networking:

• Token Bucket/Leaky Bucket — QoS algorithms derived from token concepts regulate traffic flow
• Election Protocols — Active monitor election influenced spanning tree root election and other distributed algorithms
• Deterministic Networking — Time-Sensitive Networking (TSN) in modern Ethernet echoes ring's bounded-latency goals
• Fault Detection — Ring's comprehensive error detection and reporting set expectations for network diagnostics

Historical Timeline:

We have explored ring topology comprehensively—from its circular structure through token passing mechanics, timing analysis, fault tolerance, and historical significance. Let's consolidate the essential knowledge:

What's Next:

Having mastered the three foundational topologies—bus, star, and ring—we'll explore mesh topology, an architecture designed for maximum redundancy and fault tolerance. Mesh topology is the foundation of the Internet's routing infrastructure and modern data center fabrics.