Measuring Network Performance (CCST Networking)

Network Performance: Understanding Key Metrics

Introduction to Network Performance

• Kevin Wallace introduces the topic of network performance, outlining the key areas of focus for the video.

• The discussion will cover terminology such as bandwidth vs. throughput, latency, and delay components.

• A practical demonstration using a tool called iperf will be included to measure network throughput.

Measuring Network Throughput

• To manage a network effectively, measuring its performance is essential; speed test websites like speedtest.net can provide basic metrics.

• Example measurements show download speeds of over 801 Mbps and upload speeds of over 894 Mbps, highlighting actual throughput.

• Latency is introduced as the total delay experienced during data transmission, with specific examples showing 19 ms for downloading and 20 ms for uploading.

Bandwidth vs. Throughput

• Bandwidth refers to the theoretical maximum capacity of a link, while throughput indicates the actual amount of data transmitted over time.

• An analogy compares bandwidth to road lanes: more lanes (higher bandwidth) allow more cars (data packets), but real-world conditions affect throughput.

• Factors like packet loss or errors can reduce throughput; construction on roads serves as an analogy for network congestion affecting data flow.

Understanding Latency and Delay Components

• Latency encompasses all delays in packet transmission from one point to another in a network; it consists of various delay components.

• Propagation delay is defined as the time taken for signals (light/electricity) to travel through different mediums; it's limited by the speed of light.

• For example, a fiber optic connection from London to Los Angeles has an inherent propagation delay of about 59 milliseconds.

Types of Delays Affecting Network Performance

• Transmission delay occurs when packets exit an interface like a router; this varies based on interface speed and packet size.

• Kevin mentions historical context regarding slow interfaces but does not elaborate further in this segment.

Understanding Network Delays and Throughput Measurement

Types of Network Delays

• The speaker discusses the impact of a 56 kilobit per second circuit, noting that sending a 1500 byte packet takes 214 milliseconds, which significantly affects voice quality in applications like VoIP.

• Emphasizes that while modern networks with higher speed links may mitigate some issues, overall latency is still influenced by various delays including propagation delay and transmission delay.

• Introduces queuing delay, explaining how routers or switches buffer packets when they receive data faster than they can transmit it. Quality of Service (QoS) can prioritize certain traffic types within these queues.

• Defines queuing delay as the time a packet spends in a sub-queue before being transmitted based on available bandwidth and QoS settings.

• Mentions processing delay, where routers must check packets for errors and consult routing tables before forwarding them, contributing to overall latency.

Accessing Course Content

• The speaker promotes an all-access pass for courses at less than a dollar a day, providing streaming access to current and pre-release content.

• Highlights upcoming course updates including CCST Networking and AWS Certified Cloud Practitioner among others, emphasizing the value of continuous learning through subscription.

• Discusses pricing strategy aimed at making training more accessible compared to traditional one-time course fees.

Measuring Network Throughput with iPerf

• Introduces iPerf as a free open-source utility for measuring throughput between devices on a network. It can be downloaded from iperf.fr and works across multiple platforms.

• Describes setting up iPerf with one device acting as the server and another as the client to generate traffic for performance measurement during troubleshooting scenarios.

• Explains running tests on different wireless bands (2.4 GHz vs. 5 GHz), aiming to identify performance discrepancies affecting wireless clients' speeds.

• Prepares to demonstrate iPerf in action using Mac OS commands to run the utility effectively from its directory.

Testing Network Throughput with iPerf

Setting Up the Test

• The speaker discusses various options for testing network performance, focusing on a basic throughput test between two PCs using iPerf.

• To set up the server, the command ./iperf3 -s is used, which listens on Port 5201 for incoming connections. The default test duration is 10 seconds but can be adjusted with the -T option.

Running the Client Test

• On PC2 (the client), the command ./iperf3 -c 172.16.105.160 is executed to connect to the server and run a 10-second test.

• The output displays real-time data transfer rates, indicating that throughput measures actual data transmitted rather than theoretical bandwidth; initial results show fluctuations in speed.

Analyzing Throughput Results

• Initial throughput averages around 430 megabits per second while connected to a 5 GHz wireless band.

• A switch to a crowded 2.4 GHz band prompts another test to assess any decline in performance.

Observing Performance Decline

• Results from switching bands reveal a significant drop in throughput from approximately 430 megabits per second down to about 80 megabits per second, highlighting how network conditions affect performance.

• This example illustrates how iPerf can effectively diagnose network issues by comparing different configurations and their impact on throughput.