Network Troubleshooting 24% Lesson 7 of 9

Lesson 5.4.2 — Troubleshooting Wired & Network Performance Issues

Avatar Of Asad IjazAsad Ijaz ·Sep 21, 2026 ·5 min read
78% through domain
Illustration Of A Highway Narrowing To One Lane Alongside A Dedicated Fast Lane Merging Ahead Of Traffic

Domain 5.0 | Network Troubleshooting — 24% of exam

Learning Objectives

By the end of this lesson, you will be able to:

  • Distinguish a bandwidth bottleneck from general network congestion
  • Explain the performance impact of a duplex mismatch beyond just rising collision counters
  • Describe MTU mismatch and how it causes fragmentation or dropped packets
  • Explain how QoS misconfiguration affects latency-sensitive traffic during periods of congestion

Key Terms

TermDefinition
Bandwidth BottleneckA single link or device with lower capacity than the rest of the path, constraining overall throughput
Network CongestionToo much traffic competing for available capacity, causing queuing delay and drops
Duplex MismatchTwo connected devices operating with different duplex settings, causing collisions and severely degraded throughput
MTU MismatchDevices along a path configured with different maximum transmission unit sizes, causing fragmentation or dropped packets
QoS (Quality of Service)A set of mechanisms that prioritize certain traffic types, typically latency-sensitive ones, over others during congestion

Explanation

From Wireless to Wired Performance

The previous lesson covered performance issues specific to wireless networks. This lesson rounds out objective 5.4 by covering performance problems that apply to wired and general network infrastructure, regardless of the wireless-specific factors covered previously.

Bandwidth Bottlenecks and Network Congestion

These two terms describe genuinely different situations, even though both produce a similar user-facing symptom of “things are slow”:

  • A bandwidth bottleneck is a single link or device somewhere along a path with meaningfully lower capacity than the rest of the path — a 1 Gbps uplink connecting two otherwise 10 Gbps segments, for instance. Traffic through that path can never exceed what the bottleneck link allows, regardless of how much capacity exists everywhere else. Link aggregation, covered earlier in this course, is one common remedy specifically for this kind of constraint.
  • Network congestion is a broader condition where the total traffic demand exceeds available capacity at some point in the network, causing queuing delay and, eventually, dropped packets as buffers fill up. Congestion can occur even without any single bottleneck link, simply from too much aggregate demand at a shared point.
Diagram Comparing A Single Narrow Link Constraining Traffic Flow Against Multiple Wide Links Converging Into A Congested Junction Point
How A Single Constraining Link Differs From Overall Traffic Demand Exceeding Available Capacity

The distinction matters for the fix: a bottleneck is solved by upgrading or aggregating the specific constraining link, while congestion may require broader capacity planning, traffic shaping, or addressing the actual source of excess demand.

Duplex Mismatch Revisited from a Performance Angle

Rising collision counters, covered earlier in this module, are one visible symptom of a duplex mismatch — but the actual performance impact goes well beyond an occasional collision. A genuine duplex mismatch, where one device is configured for full duplex and the other for half duplex, produces a specific kind of collision called a late collision, and the resulting performance degradation is often severe — throughput dropping to a small fraction of the link’s rated capacity, not just a minor dip. This is exactly why a duplex mismatch is worth ruling out early when a specific link performs dramatically worse than its rated speed would suggest, rather than assuming a more exotic cause first.

MTU Mismatch

MTU (Maximum Transmission Unit) defines the largest packet size a given link or device will handle without fragmenting it. An MTU mismatch occurs when devices along a single path are configured with different MTU values — a packet sized correctly for one segment can turn out to be too large for another segment further along the path.

Diagram Showing An Oversized Packet Either Being Fragmented Or Dropped When It Exceeds A Downstream Device'S Mtu, Depending On The Don'T-Fragment Flag
How A Packet Exceeding A Downstream Device’S Mtu Gets Fragmented Or Dropped Depending On The Don’T-Fragment Flag

When this happens, one of two things occurs: the oversized packet gets fragmented into smaller pieces to fit, adding overhead and reducing effective throughput, or — if the packet has its “don’t fragment” flag set — it gets dropped outright, with the sender typically expected to receive an error message indicating the packet needed to be smaller. This produces a genuinely distinctive symptom pattern: small packets flow through the path just fine, while larger packets specifically fail or perform poorly, which is a strong, specific clue pointing toward an MTU mismatch rather than a more general connectivity problem.

QoS Misconfiguration

QoS (Quality of Service) mechanisms exist to prioritize certain traffic types — typically latency-sensitive traffic like voice and video — ahead of less time-sensitive traffic like file transfers or general web browsing, specifically during periods when available bandwidth is genuinely constrained. When QoS is missing or misconfigured, latency-sensitive applications can suffer from jitter, delay, and dropped packets during congestion, even when the network has technically sufficient total bandwidth for all the traffic combined — because without prioritization, that bandwidth gets allocated on a first-come, first-served basis rather than according to which traffic actually needs low latency most.

Diagram Comparing Voice Traffic Stuck In A Mixed Queue Without Qos Against Voice Traffic Prioritized In A Fast Lane With Qos Enabled
How Qos Prioritizes Latency-Sensitive Traffic Ahead Of Other Traffic Specifically During Congestion

This is exactly why a voice call can sound choppy on a network with seemingly plenty of overall bandwidth available — the total capacity might genuinely be sufficient, but without QoS actively prioritizing the voice traffic during a momentary spike in competing demand, that voice traffic gets stuck waiting in the same queue as everything else.

Recognition-Level Verification Concepts

A few patterns are worth recognizing on sight:

  • A single link performing far below the surrounding network’s capacity, with the rest of the path performing fine, points toward a bandwidth bottleneck at that specific link.
  • General slowness affecting many different destinations simultaneously, without one obvious constraining link, points toward broader network congestion.
  • Throughput on a specific link dropping to a small fraction of its rated speed, alongside rising collision counters, points toward a duplex mismatch rather than a minor cabling issue.
  • Small packets succeeding while larger ones specifically fail or fragment along the same path points toward an MTU mismatch.
  • Voice or video quality suffering specifically during periods of network congestion, despite adequate total bandwidth, points toward missing or misconfigured QoS.

Common Exam Traps

  • A bandwidth bottleneck and network congestion are related but distinct problems, requiring different fixes. A bottleneck is solved by upgrading one specific link; congestion may require broader capacity or traffic management changes.
  • A duplex mismatch’s performance impact is often severe, not minor. Don’t underestimate how dramatically throughput can drop from a duplex mismatch compared to a routine cabling issue.
  • An MTU mismatch’s distinctive symptom is size-dependent failure — small packets working while large ones fail is a strong, specific signal, not a vague “sometimes it works” pattern.
  • QoS problems don’t require insufficient total bandwidth to occur. The network can have technically adequate capacity overall and still suffer real latency-sensitive traffic problems if that capacity isn’t being prioritized correctly during brief congestion spikes.
  • These four issue categories can look similar at first (“the network feels slow”) but have genuinely different root causes and fixes — resist treating general performance complaints as a single undifferentiated troubleshooting category.

Lesson 5.4.2 Practice Quiz — Wired & Network Performance Issues

17 questions covering bandwidth bottlenecks, congestion, duplex mismatch, MTU mismatch, and QoS misconfiguration.

N10-009 · Domain 5.4
Question 1Plain
What is a bandwidth bottleneck?
A bandwidth bottleneck is a single constraining link or device limiting overall throughput along a path.
Question 2Plain
What does MTU define?
MTU (Maximum Transmission Unit) defines the largest packet size a link or device will handle without fragmenting it.
Question 3Plain
What does QoS do?
QoS prioritizes latency-sensitive traffic ahead of other traffic specifically during periods of congestion — it doesn't increase total bandwidth.
Question 4Choose Two
Which two statements correctly distinguish a bandwidth bottleneck from network congestion? (Choose two.)
A bottleneck is a single constraining link; congestion is broader excess demand — the two are related but distinct concepts requiring different fixes.
Question 5Choose Two
Which two statements about duplex mismatch are correct? (Choose two.)
Duplex mismatch produces late collisions and can cause a genuinely severe throughput drop — its impact is significant, not negligible.
Question 6Choose Two
Which two statements about MTU mismatch are correct? (Choose two.)
The distinctive MTU mismatch pattern is size-dependent: larger packets fail or fragment, while smaller packets continue to succeed on the same path.
Question 7Scenario
A single 1 Gbps link connects two otherwise 10 Gbps network segments, and overall throughput between them never exceeds 1 Gbps. What does this describe?
A single lower-capacity link constraining an otherwise higher-capacity path is exactly a bandwidth bottleneck.
Question 8Scenario
Users across the network report general slowness reaching many different destinations, with no single obviously constraining link identified. What does this suggest?
Broad slowness across many destinations, with no single constraining link, points toward general network congestion rather than a localized bottleneck.
Question 9Scenario
A specific link's throughput drops to a small fraction of its rated speed, and its interface counters show rising collisions. What is the likely cause?
Severely degraded throughput alongside rising collisions is the classic duplex mismatch pattern.
Question 10Scenario
Small packets travel a network path without any issue, but larger packets specifically fail or get fragmented along the same path. What does this pattern suggest?
Size-dependent failure — small packets fine, large ones failing — is the distinctive signature of an MTU mismatch.
Question 11Scenario
A VoIP call becomes choppy during a brief period of network congestion, even though the network has technically sufficient total bandwidth for all combined traffic. What does this suggest?
Latency-sensitive traffic suffering during congestion despite adequate total bandwidth is a classic sign of missing or misconfigured QoS prioritization.
Question 12Exhibit
Based on this throughput test, what is limiting the connection?
Throughput Test: Segment A: 10 Gbps capable Uplink between A and B: 1 Gbps Segment B: 10 Gbps capable Measured throughput A-to-B: 0.94 Gbps
Throughput capped almost exactly at the 1 Gbps uplink's capacity, despite 10 Gbps segments on either side, is a textbook bandwidth bottleneck.
Question 13Exhibit
Based on this monitoring data, what is occurring?
Network Monitoring — Peak Hours (9-10 AM) Multiple core links: 85-95% utilization simultaneously Queuing delay: Elevated across several unrelated paths Packet drops: Increasing during this window only
Elevated queuing and drops across multiple unrelated links simultaneously, specifically during peak hours, is exactly network congestion rather than a single bottleneck.
Question 14Exhibit
Based on this interface log, what issue is indicated?
Interface Gi0/9 Log: Late Collisions: 3,204 and climbing Local Duplex: Full Remote Duplex: Half Throughput: 8 Mbps on a 100 Mbps link
Late collisions, mismatched duplex settings, and throughput reduced to a small fraction of the rated 100 Mbps together confirm a duplex mismatch.
Question 15Exhibit
Based on this ping test, what issue is indicated?
Ping Test (varying packet size, Don't Fragment flag set): Size 100 bytes: SUCCESS Size 500 bytes: SUCCESS Size 1500 bytes: FAILED — "Packet needs to be fragmented but DF set"
Small packets succeeding while a larger packet fails with an explicit fragmentation-needed error is a textbook MTU mismatch.
Question 16Exhibit
Based on this QoS configuration, what problem exists?
QoS Configuration Review: Voice VLAN traffic: No DSCP marking applied Priority queue configuration: Not configured Result: Voice traffic treated identically to bulk file transfer traffic
No DSCP marking or priority queuing for voice traffic, leaving it treated identically to bulk traffic, is exactly a missing QoS configuration.
Question 17Exhibit
Based on this VoIP quality log, what is the likely root cause?
VoIP Quality Log: Jitter spikes: Correlate exactly with network congestion peaks (9-10 AM daily) Total link bandwidth: Sufficient for all combined traffic (verified) QoS: Not configured on this link
Jitter correlating with congestion despite verified sufficient total bandwidth, combined with no QoS configured, points directly at missing QoS prioritization as the root cause.
📝

Summary

A bandwidth bottleneck is a single constraining link limiting overall throughput; network congestion is broader excess demand exceeding available capacity — each requires a different fix.

A duplex mismatch produces severe throughput degradation through late collisions, going well beyond the occasional collision counter increment that might first draw attention to it.

An MTU mismatch causes fragmentation or outright packet drops specifically for larger packets, producing a distinctive size-dependent failure pattern.

QoS misconfiguration allows latency-sensitive traffic to suffer during congestion even when total bandwidth is technically sufficient, since that capacity isn't being allocated according to actual latency sensitivity.

This lesson completes N10-009 objective 5.4 (Performance Issues) at 2/2 lessons; the next lessons move into objective 5.5, troubleshooting tools and protocols — the final objective of the course.

Avatar Of Asad Ijaz

Lead Networking Architect and Editor at NetworkUstad. BS in Computer Networks and Security, CCNP and CCNA certified, with 11+ years of experience in enterprise network design, implementation, and troubleshooting. Writes practical tutorials on routing, IPv4 management, network automation, and security fundamentals.