Congestion Control Algorithms in Computer Networks - The Logic Tutorial

Congestion Control Algorithms in Computer Networks

Congestion control algorithms are techniques used in computer networks to manage and mitigate network congestion, which occurs when the demand for network resources exceeds the available capacity. These algorithms aim to prevent or alleviate congestion by adjusting the rate at which data is transmitted through the network. Here are some common congestion control algorithms:

1. TCP Congestion Control Algorithms

   • Slow Start: Initially, the sender's congestion window (cwnd) is set to a small value, and it gradually increases the transmission rate until it encounters congestion or packet loss.
   • Congestion Avoidance: After reaching a certain threshold, the sender increases the cwnd more gradually to avoid causing congestion.
   • Fast Retransmit: If the sender receives three duplicate acknowledgments (ACKs), it assumes packet loss and retransmits the missing packet without waiting for a retransmission timeout.
   • Fast Recovery: After Fast Retransmit, the sender enters the Fast Recovery phase, where it adjusts the cwnd to account for the packet loss and continues transmission.

2. Network-Assisted Congestion Control

   • Explicit Congestion Notification (ECN): Routers mark packets as experiencing congestion instead of dropping them, allowing the sender to adjust its transmission rate accordingly.

3. Active Queue Management (AQM) Algorithms

   • Random Early Detection (RED): Routers monitor the average queue length and proactively drop or mark packets before the queue becomes full, providing early congestion signals to senders.
   • Controlled Delay (CoDel): Monitors the queue delay and drops packets if the delay exceeds a specific threshold, providing a more responsive congestion signal.

4. Delay-based Congestion Control Algorithms

   • TCP Vegas: Adjusts the sending rate based on the estimated round-trip time (RTT) of the connection, aiming to maintain a small but consistent queue at the bottleneck link.
   • LEDBAT (Low Extra Delay Background Transport): Designed for background traffic, it adjusts the sending rate based on one-way delay measurements to minimize the impact on other traffic.

5. Bandwidth Estimation Algorithms

   • TCP Westwood: Estimates the available bandwidth based on the acknowledgment rate and adjusts the sending rate accordingly.
   • TCP BBR (Bottleneck Bandwidth and Round-trip propagation time): Estimates the bottleneck bandwidth and round-trip propagation time to adjust the sending rate.

6. Multipath Congestion Control Algorithms

   • MPTCP (Multipath TCP): Splits data across multiple paths or subflows, distributing the traffic and providing congestion control for each subflow.

These congestion control algorithms are designed to strike a balance between maximizing network utilization and avoiding congestion collapse. The choice of algorithm depends on factors such as network conditions, application requirements, and the trade-offs between responsiveness, fairness, and throughput.