Unit
- Latency: The time taken for a packet to travel from source to destination, measured in milliseconds. It directly impacts user experience for real-time applications like VoIP and video conferencing. Low latency is critical for interactive services.
- Throughput: The rate at which data is successfully transmitted over a network, usually expressed in bits per second. It represents the actual data transfer capacity, which may be less than the theoretical bandwidth due to overhead and congestion.
- Jitter: The variation in packet arrival times. High jitter causes buffering and quality degradation in streaming media. Acceptable jitter depends on the application: voice requires under 30 ms.
- Packet Loss: The percentage of packets that fail to reach their destination. Even 1% loss can degrade TCP performance significantly due to retransmissions. Loss is measured by protocols like ICMP or UDP probes.
- Availability: The percentage of time the network is operational and accessible. Often expressed as nines (e.g., 99.9% availability allows about 8.7 hours downtime per year). Redundancy improves availability.
- Utilization: The percentage of available bandwidth that is in use. High utilization (above 70%) often leads to congestion and increased latency. Monitoring utilization helps in capacity planning.
- Mean Time Between Failures (MTBF): The average time a network component operates without failure. Higher MTBF indicates reliability. It is used in designing redundant architectures.
- Mean Time to Repair (MTTR): The average time taken to restore a failed component. Lower MTTR improves overall availability. Automation tools can reduce MTTR.
- Bandwidth: The maximum data transfer rate of a network path, often confused with throughput. It is a theoretical limit determined by physical media and encoding. Bandwidth is a key factor in capacity planning, but actual throughput is usually lower due to protocol overhead and congestion.
- Round Trip Time (RTT): The time for a signal to travel to a destination and back, commonly measured using ping. RTT includes propagation delay, processing delay, and queuing delay. It is a critical metric for TCP performance, as it influences retransmission timers and window scaling.
- Error Rate: The frequency of bit errors in data transmission, often expressed as Bit Error Rate (BER). High error rates cause retransmissions at higher layers, reducing effective throughput. It is influenced by signal-to-noise ratio and physical medium quality. Monitoring error rates helps identify faulty hardware or interference.
- Network Congestion: A state where demand for network resources exceeds capacity, leading to packet drops and increased latency. Congestion can be caused by traffic bursts or insufficient bandwidth. Techniques like traffic shaping and QoS help manage congestion. Proactive monitoring of congestion indicators prevents performance degradation.
Code Example
Python
from ping3 import ping
import statistics
host = '8.8.8.8'
rtts = []
for _ in range(5):
rtt = ping(host, unit='ms')
if rtt is not None:
rtts.append(rtt)
if rtts:
avg_rtt = statistics.mean(rtts)
print('Average RTT: {:.2f} ms'.format(avg_rtt))
else:
print('No response')This script pings a remote host five times, collects round-trip times (RTTs), and calculates the average. The ping3 library returns RTT in milliseconds. Use this for basic latency measurement.
Quick Revision
- Key network KPIs include latency, throughput, jitter, packet loss, availability, utilization, MTBF, and MTTR.
- Each KPI impacts application performance differently; optimization targets specific metrics.
- Tools like ping and iperf help measure these KPIs systematically.
- Baseline measurements are essential for identifying performance issues.