Computer Science – 14.2 Circuit switching, packet switching | e-Consult
14.2 Circuit switching, packet switching (1 questions)
Congestion in a packet-switched network occurs when the network's capacity is exceeded, leading to increased delays and packet loss. This can happen when too many packets are being transmitted simultaneously, or when a particular link becomes overloaded. Congestion can propagate through the network, affecting performance across multiple links.
Two common congestion control mechanisms are:
- TCP Congestion Control (e.g., Slow Start, Congestion Avoidance): TCP uses a sophisticated congestion control algorithm to adjust the sending rate based on network conditions.
- Slow Start: The sender starts with a small congestion window and gradually increases it until congestion is detected.
- Congestion Avoidance: When congestion is detected, the sender reduces the congestion window and avoids aggressive increases.
Trade-offs: TCP congestion control can be effective at preventing congestion, but it can also reduce throughput, especially in high-latency networks. It requires accurate detection of congestion, which can be challenging.
- Explicit Congestion Notification (ECN): ECN allows routers to signal congestion to the sender without dropping packets. Routers mark packets with an ECN bit when they detect congestion. The sender then reduces its sending rate when it receives ECN signals.
Trade-offs: ECN requires support from both the sender and the network infrastructure. It can be more efficient than TCP congestion control in some situations, but it is more complex to implement.