This title slide diagram shows two desktop computers connected through a switch using Ethernet cables, with their hardware MAC addresses displayed above each device. This illustrates the fundamental role of switches in connecting devices at Layer 2 of the OSI model using MAC addressing.
Review: From Cables to Connections
Two key components:
Layer 2 Focus
Cables carry bits, but NICs and switches work with frames and MAC addresses.
This horizontal timeline shows the seven layers of the OSI model, with physical infrastructure layers (Physical, Data Link, Network, Transport) at the left and application-oriented layers at the right. An arrow labeled "Today!" points to the Data Link layer, indicating that this module focuses on Layer 2 switching and MAC addressing. The bracket on the left shows that Module 2 covered the Physical layer.
After completing this module, you will be able to:
This section introduces network interfaces and transceivers, the hardware endpoints that connect hosts to Ethernet networks and expose physical and data-link functions.
What is a Network Interface Card (NIC)?
NIC Responsibilities
The NIC diagram shows a computer with an integrated network interface card that contains the unique MAC address burned into its hardware. The NIC translates between the computer’s internal data and network signals transmitted through the RJ-45 port to the switch.
Key Point
Without a NIC, your computer is an island—no network access!
Figure: Examples of network interface cards with different ports and form factors.
Examples of different network interface cards showing various form factors and connector types.
NIC Features and Specifications
Auto-Negotiation
NICs automatically negotiate the best speed and duplex with the switch. Usually “just works”!
| NIC Type | Typical Use |
| 1 Gbps RJ-45 | Desktops, laptops |
| 10 Gbps RJ-45 | Workstations |
| 10 Gbps SFP+ | Servers |
| 25/100 Gbps | Data centers |
Dual-Port NICs
Servers use multiple NIC ports for:
Why Transceivers Matter
Without a transceiver, your switch can’t understand fiber optic light signals—it only speaks electrical!
Figure: Transceiver cutaway showing optical/electrical conversion components.
Key Components
Laser/LED: Converts electrical → light Photodetector: Converts light → electrical Circuit board: Manages the conversion
| Transceiver | Speed | Cable Type | Connector | Distance |
| SFP | 1 Gbps | MMF or SMF | LC | 550m (MMF) |
| 10km (SMF) | ||||
| SFP+ | 10 Gbps | MMF or SMF | LC | 300m (MMF) |
| 10–80km (SMF) | ||||
| QSFP | 40 Gbps | MMF or SMF | MPO/MTP | 100m (MMF) |
| 10km (SMF) | ||||
| QSFP28 | 100 Gbps | MMF or SMF | MPO/MTP | 100m (MMF) |
| 10km (SMF) | ||||
| GBIC | 1 Gbps | MMF or SMF | SC | 550m (MMF) |
| (legacy) | 10km (SMF) | |||
| SFP-T | 1 Gbps | Copper (UTP) | RJ-45 | 100m |
| SFP+ DAC | 10 Gbps | Twinax copper | Integrated | 3–7m |
MMF = Multimode Fiber SMF = Single-mode Fiber DAC = Direct Attach Cable.
Always match transceiver type to cable type and required distance.
Layer 1: Physical Infrastructure at the Green Dragon
The Green Dragon network infrastructure diagram illustrates a small business physical topology with a cellar server connected via multimode fiber to a core switch, three client devices connected via copper Cat6/Cat5e cables, and a single-mode fiber uplink to the ISP. This demonstrates the cable diversity typical in modern networks.
Layer 1 maps physical links between endpoints, switching infrastructure, and uplinks.
Mismatch Issues
Vendor Lock-In
Some manufacturers only accept their own branded transceivers. Third-party modules are cheaper but may not work.
Signal Strength Issues
Troubleshooting Tip: Transceivers
Transceiver compatibility
Fiber type (MMF vs SMF)
Cable distance vs. transceiver rating
Here we examine Ethernet frame structure and MAC addressing, which form the basis of Layer 2 switching decisions and local delivery.
The Ethernet frame structure shows how data is packaged for transmission at Layer 2. The preamble synchronizes receivers, source and destination MAC addresses identify communicating devices, the type field indicates the protocol (like IPv4), the payload carries actual data (46-1500 bytes), and the Frame Check Sequence detects transmission errors. Total maximum size is 1518 bytes for standard frames.
Header Fields
Preamble: Synchronization bits Dest/Src MAC: Who it’s going to/from Type: What protocol is inside
Trailer Field
FCS (Frame Check Sequence): Error detection—receiver recalculates to verify frame wasn’t corrupted.
MAC vs IP
MAC = permanent hardware address (Layer 2). IP = changeable logical address (Layer 3).
This diagram breaks down a MAC address (00:1A:2B:3C:4D:5E) into its two components: the first three bytes identify the manufacturer (Organizationally Unique Identifier or OUI), and the last three bytes are the unique device identifier assigned by that manufacturer. This 48-bit hexadecimal address is burned into the NIC hardware at the factory.
Common Formats
00:1A:2B:3C:4D:5E (colons) 00-1A-2B-3C-4D-5E (dashes) 001A.2B3C.4D5E (dots—Cisco)
This diagram shows a specific MAC address (00:50:56:3C:4D:5E) split into two segments: the Organizationally Unique Identifier (00:50:56) assigned by IEEE to identify the manufacturer (in this case VMware), and the Device ID (3C:4D:5E) which is unique to each NIC from that vendor. This hierarchical structure ensures globally unique hardware addresses.
Example OUIs
00:50:56 = VMware 00:0C:29 = VMware (alternate) 00:1A:A0 = Dell
Special MAC Addresses
Broadcast: FF:FF:FF:FF:FF:FF Goes to ALL devices on the network.
Multicast: Starts with 01:00:5E Goes to a group of devices.
Can MACs Be Changed?
Yes! Software can “spoof” a different MAC. Useful for troubleshooting, but also a security concern.
Case Study: Sam’s First NIC Installation
Case Study: Sam’s First NIC Installation
Samwise Gamgee is setting up the new admin burrow network for the Shire. He
purchases 10 Gbps NICs for the hobbit-hole workstations, but the switches only
have SFP ports (1 Gbps, not SFP+).
Sam also notices one workstation showing a MAC address of 00:00:00:00:00:00 in the network settings.
Review Questions
Will the 10 Gbps NICs work in the 1 Gbps switch ports?
What speed will the connection actually operate at?
What might cause a MAC address of all zeros?
Case Study Solution: Sam’s First NIC Installation
Solution: Sam’s First NIC Installation
Yes—10 Gbps NICs will auto-negotiate down to match the switch’s 1 Gbps capability.
The connection will operate at 1 Gbps—limited by the slower device (the switch).
All-zeros MAC address typically means:
Key Lessons
This section compares hub-, bridge-, and switch-based forwarding and explains how modern switches learn and forward traffic efficiently.
The Core Problem
How do we connect multiple devices efficiently without wasting bandwidth or causing collisions?
Why Hubs Are Obsolete
10 devices on a 100 Mbps hub = roughly 10 Mbps each (minus collision overhead). Terrible!
This diagram shows a hub at the center connected to five PCs, with all devices enclosed in a single red dashed circle representing one collision domain. In a hub-based network, all connected devices share the same bandwidth and must take turns transmitting to avoid collisions, making hubs inefficient for modern networks.
Hub = Layer 1
Hubs operate at the Physical layer—they don’t understand MAC addresses.
Bridges: Learning MAC Addresses
Bridge = Layer 2
Bridges read MAC addresses and make forwarding decisions—smarter than hubs!
This diagram depicts a bridge connecting two network segments, with two PCs on the left and two PCs on the right. Each segment is enclosed in its own collision domain (green dashed circles labeled CD 1 and CD 2), meaning collisions on one side do not affect the other. Bridges use MAC learning to forward frames only when necessary, reducing network congestion compared to hubs.
Selective Forwarding
Traffic between L1 and L2 stays on the left—the bridge doesn’t forward it right.
Bandwidth Advantage
A 24-port Gigabit switch = up to 24 Gbps total capacity (each port gets full 1 Gbps).
This diagram shows a switch with five connected PCs, where each device has its own individual collision domain (small blue dashed circles around each PC). Unlike hubs, switches eliminate collisions by providing dedicated bandwidth per port, allowing simultaneous full-duplex communication. This architecture maximizes network efficiency and throughput in modern LANs.
Switch = Layer 2
Like bridges, switches read MAC addresses—but with many more ports and better performance.
How Switches Learn and Forward
The Four Switch Actions
Learning: See source MAC → record which port it came from
Forwarding: Know destination MAC → send only to that port
Flooding: Unknown destination → send to ALL ports (except source)
Filtering: Same-segment traffic → don’t forward
This diagram depicts a switch with three connected PCs, showing the learned MAC address table that maps each device’s hardware address (AA:AA, BB:BB, CC:CC) to its corresponding port (P1, P2, P3). Switches build this table dynamically by examining source MAC addresses on incoming frames, enabling intelligent forwarding decisions rather than broadcasting all traffic like hubs.
MAC Address Table
The switch builds a table mapping MAC addresses to ports. This is how it knows “who is where.”
Layer 2: MAC Address Communication at the Green Dragon
This diagram illustrates a practical switching scenario where the Front Desk (MAC AA:AA:AA on Port 1) sends a frame to the Manager (MAC CC:CC:CC on Port 3). The switch examines the destination MAC address in the frame and forwards it only to Port 3, blocking it from Ports 2 and 4 (marked with X). This selective forwarding demonstrates how switches use MAC tables to direct traffic efficiently, conserving bandwidth compared to hubs that broadcast to all ports.
Layer 2 uses MAC addresses to forward frames. The switch sends traffic only to the destination port—not everywhere.
This section focuses on operational switch features and management choices used in real deployments.
Unmanaged Switch
Smart Switch
Managed Switch
Key Question
Do you need to separate traffic, monitor performance, or configure security? If yes → managed switch.
Layer 2 Switch (Standard)
This simple diagram shows a Layer 2 switch connected to a separate router, illustrating that traditional Layer 2 switches must forward traffic to an external router for inter-network routing decisions.
Layer 3 Switch (Multilayer)
This diagram shows a Layer 3 multilayer switch with the annotation "Routes internally!" above it, indicating that L3 switches can perform IP routing without needing an external router. This capability makes inter-VLAN routing much faster in enterprise networks.
When to Use Layer 3 Switches
Large networks with multiple VLANs benefit from Layer 3 switches—inter-VLAN traffic stays fast without bottlenecking through a router.
Switch Interface Configuration Basics
Access Methods
Common Settings
Speed/Duplex Mismatch
If one side is set to auto and the other is manually configured, they may negotiate incorrectly. Result: slow speeds, errors, packet loss.
Port Security
Limit which MAC addresses can connect to a port:
Case Study: The Green Dragon Inn Network
Case Study: The Green Dragon Inn Network
The Green Dragon Inn is expanding and needs a network for guest hobbits and
staff. Frodo suggests a cheap unmanaged switch. Gandalf recommends a managed
switch instead.
The network requirements:
Review Questions
Which switch type should they choose and why?
What feature would separate guest from staff traffic?
Why is remote management valuable for an inn?
Case Study Solution: The Green Dragon Inn Network
Solution: The Green Dragon Inn Network
Managed switch—unmanaged switches cannot separate traffic or be configured remotely.
VLANs (Virtual LANs) separate guest and staff traffic logically on the same physical switch.
Remote management benefits:
Key Lesson
The extra cost of managed switches pays off in flexibility and security. For any business network, managed is the right choice.
Advanced switching capabilities improve resiliency, bandwidth utilization, and loop prevention in larger networks.
Link Aggregation and NIC Teaming
Benefits
More bandwidth: 4 × 1G = 4 Gbps total Redundancy: If one link fails, others continue
This diagram shows link aggregation (LAG) where a server with multiple network interfaces connects to a switch using four bundled Ethernet cables, all labeled with the bracket "LAG". This configuration combines multiple physical links into one logical connection, providing higher bandwidth and redundancy. If one cable fails, traffic continues flowing through the remaining links. The LACP protocol (802.3ad) coordinates the aggregation between both endpoints.
Both Ends Must Match
LAG must be configured on both the switch AND the server/other switch.
Maximum Transmission Unit (MTU)
Critical Requirement
Every device in the path must support the same MTU! Mismatched MTU causes fragmentation or dropped packets.
When to Use Jumbo Frames
Spanning Tree Protocol: The Problem
Real Danger
An accidental loop can crash an entire network in under 30 seconds!
This diagram illustrates a broadcast storm caused by a network loop. Three switches are connected in a triangle topology with red arrows showing frames circulating endlessly around the loop, marked with an explosion symbol (X) in the center. Without Spanning Tree Protocol, broadcast frames entering this loop will replicate infinitely, overwhelming switch CPUs and making the network unresponsive.
Symptoms
All switch LEDs flashing rapidly, network unresponsive, high CPU on switches.
Spanning Tree Protocol: The Solution
STP Versions
802.1D (STP): Original, slow (30–50 sec) 802.1w (RSTP): Rapid, fast (1–2 sec) 802.1s (MSTP): Per-VLAN spanning trees
This diagram shows Spanning Tree Protocol (STP) preventing loops in a redundant topology. Three switches are connected in a triangle, with two links shown in green (active) and one link displayed with a red prohibition symbol labeled "BLOCKED". STP automatically identifies the redundant link and blocks it to prevent broadcast storms, while keeping it available as a backup if an active link fails.
Key Insight
STP provides redundancy WITHOUT loops—blocked ports wait as backups.
Power Budget
Switches have a total PoE power budget (e.g., 370W). Plan carefully—you can’t power unlimited devices!
| Standard | Power | Devices |
| 802.3af | 15.4W | IP phones |
| 802.3at (PoE+) | 30W | Cameras, APs |
| 802.3bt (PoE++) | 60–100W | PTZ cameras |
PoE Advantage
This simple diagram shows a PoE switch connected to a wireless access point with a single Ethernet cable labeled "Data + Power (one cable!)". This illustrates the key benefit of Power over Ethernet: eliminating the need for separate power adapters and electrical outlets at device locations.
Power over Ethernet: Green Dragon Power Budget
This diagram shows the Green Dragon pub’s PoE+ switch with a 120W power budget supporting four devices across ports P1–P4: a VoIP phone (7W), Main Hall wireless AP (25W), entry security camera (15W), and Garden AP (28W). The power budget bar displays 75W used (green) and 45W still available (gray), demonstrating the importance of tracking cumulative PoE consumption. Each device receives both data and power over a single cable, eliminating the need for separate power adapters.
PoE delivers data and power over one cable. Always track your power budget—the switch has limits!
Case Study: Bilbo’s Birthday Party Network Disaster
Case Study: Bilbo’s Birthday Party Network Disaster
Bilbo is hosting his 111th birthday party and needs a network for the event
planners. Shortly after setup, the entire network crashes—all switches show
frantically blinking lights, and no device can communicate.
Merry investigates and discovers someone connected both ends of a patch cable to the same switch to “make the cables neater.”
Review Questions
What has happened to the network?
What protocol should have prevented this?
What should they check on the switches?
Case Study Solution: Bilbo’s Birthday Party Network Disaster
Solution: Bilbo’s Birthday Party Network Disaster
A switching loop caused a broadcast storm—frames multiplied until they overwhelmed every switch.
Spanning Tree Protocol (STP) should block redundant paths automatically.
Immediate actions:
Key Lesson
Always use managed switches with STP enabled for any business network. A single accidental loop can take down everything!
The final section applies structured troubleshooting to interface and switch faults using indicators, commands, and counter analysis.
Hardware Failure and Port Status Indicators
LED Indicators
First Troubleshooting Step
Always look at the lights! LEDs tell you port status instantly without logging in.
Common Hardware Failures
No Link Light?
Check in order: cable → NIC → switch port → transceiver (if applicable).
Essential Commands
show interfaces Port status, speed, duplex, errors
show mac address-table Which MACs learned on which ports
show spanning-tree STP status, root bridge, blocked ports
show power inline PoE status and power consumption
Example Output
Switch# show interfaces Gi0/1 GigabitEthernet0/1 is up Speed: 1000 Mbps, Duplex: Full Input errors: 0, CRC: 0 Output errors: 0, Collisions: 0
Key Insight
These commands work on Cisco and many other managed switches. Learn them once, use them everywhere.
Error Types
CRC errors: Damaged frames—bad cable, NIC, or interference.
Collisions: Duplex mismatch or hub in path.
Runts: Frames too small (<64 bytes)—collision fragments.
Giants: Frames too large—MTU mismatch.
Input/Output Errors
Input errors: Problems receiving frames.
Output errors: Problems sending frames.
Discards: Frames dropped (buffer full, policy).
Interpretation
A few errors = normal. Rapidly increasing = active problem!
Pro Tip
Clear counters, wait 5 minutes, check again. If errors accumulate quickly, investigate that port.
MAC Address Table Troubleshooting
Command
show mac address-table
Example Output
VLAN MAC Address Port –– –––––- –– 1 00:1A:2B:3C:4D:5E Gi0/1 1 00:50:56:AA:BB:CC Gi0/2 1 FF:FF:FF:FF:FF:FF CPU
MAC Flapping
Same MAC appearing on multiple ports rapidly = possible loop or duplicate MAC address.
Common PoE Issues
Troubleshooting Steps
Verify switch supports PoE
Check total power budget remaining
Verify cable uses all 4 pairs
Check device PoE class requirements
Try known-good cable and port
| PoE Class | Max Power |
| Class 0 | 15.4W (default) |
| Class 1 | 4.0W |
| Class 2 | 7.0W |
| Class 3 | 15.4W |
| Class 4 | 30W (PoE+) |
Cable Matters!
PoE requires all 4 pairs. A cable that “works for data” may fail for PoE if pairs are damaged.
Key Concepts:
Configuration & Troubleshooting:
This module explored Layer 2 networking devices and protocols. You learned about NICs, transceivers, MAC addresses, Ethernet frames, and the evolution from hubs to bridges to switches. You also examined switch configuration, advanced features (VLANs, link aggregation, STP, PoE), and troubleshooting techniques. In the next module, we’ll move up to Layer 3 and explore IP addressing and subnetting.