Network Performance and Bottlenecks
Network Speeds and Congestion
- Predefined Speeds: Networks operate at predefined speeds (e.g., 1 Gbit/s for a gigabit network).
- Congestion Issue: Occurs when multiple 1 Gbit links send traffic to the same destination, exceeding a single link's capacity.
- Buffer Overflow: Switches/routers have small buffers; excess packets may be discarded, leading to data loss.
Resolving Congestion
- Solutions: Increase network size/speed or reduce traffic.
- Misconception: "Network is slow" often indicates a bottleneck somewhere.
Identifying Bottlenecks
- Troubleshooting: Can be complex due to multiple possible causes:
- System bus speed
- CPU speed in switches/routers
- Storage drive speed (HDD vs SSD)
- Network speed variations
- Example: Database server delay led to 2-second transaction delays; resolved by configuration changes.
Network Usage Metrics
- Bandwidth Percentage: Indicates network use over time.
- Throughput: Amount of data moved through the network in a given timeframe.
- Monitoring Tools: Use switch statistics, SNMP, or NetFlow.
Latency and Measurement
- Latency: Delay between request and response; inherent in all connections.
- Detailed Measurement: Requires tools at each network link to capture packet traversal times.
Packet Loss and Contention
- Impact: Ideal is zero packet loss, but outages or bandwidth issues can cause discards.
- Corruption: Bad cables/wireless networks may corrupt data, causing retransmissions.
Real-Time Communication Challenges
- Voice and Video Sensitivity: Delay can disrupt voice calls or video streams.
- Jitter: Variation in packet arrival times; low jitter is preferred.
- High Jitter Effects: Causes delays, noise in calls, or video stutters.
Understanding these concepts and metrics is crucial for network administrators to maintain efficient and reliable network performance.