Congestion

Congestion, in the context of networks, refers to a network state where a node or link carries so much data that it may deteriorate network service quality, resulting in queuing delay, frame or data packet loss and the blocking of new connections. In a congested network, response time slows with reduced network throughput. Congestion occurs when bandwidth is insufficient and network data traffic exceeds capacity.

Data packet loss from congestion is partially countered by aggressive network protocol retransmission, which maintains a network congestion state after reducing the initial data load. This can create two stable states under the same data traffic load - one dealing with the initial load and the other maintaining reduced network throughput.

In modern networks, avoiding congestive collapse involves the application of network congestion avoidance techniques along with congestion control, such as:

• Exponential backoff protocols that use algorithm feedback to decrease data packet throughput to acceptable rates
• Priority techniques to allow only critical data stream transmission
• Allocation of appropriate network resources in anticipation of required increases in data packet throughput

Congestion has been described as a fundamental effect of limited network resources, especially router processing time and link throughput. Traffic directing processes, performed by routers on the Internet and other networks, use a microprocessor. Cumulative router processing time greatly impacts network congestion. In fact, intermediate routers may actually discard data packets when they exceed its handling capability. When this occurs, additional data packets may be sent to make up for unreceived packets, which exacerbates the problem. Network congestion often leads to congestion collapse.

Avoiding network congestion and collapse requires two major components:

• Routers capable of reordering or dropping data packets when received rates reach critical levels
• Flow control mechanisms that respond appropriately when data flow rates reach critical levels

This definition was written in the context of Networks