Congestive collapse
NAS Recovery - Congestive collapse
Congestive collapse refers to a condition in which a packet switched network experiences difficulty in communication due to congestion. When congestion collapse occurs, many data packets get lost and the resultant throughput becomes negligible.
How Does It Occur?
Congestion usually occurs when more packets are sent than that could be handled by intermediate routers. In such situation, these routers discard many such packets and expect the end points to retransmit the unsent information. By resending the packets, in fact they double the data rate and aggravate the situation. As a result, the entire network gets pushed into a congestion collapse.
Where Does It Occur?
Choke points in a network are more prone to congestion collapse. Choke points are those parts of the network where the total incoming bandwidth to a node exceeds the outgoing bandwidth. Most common choke points are the connection points between a local area network (LAN) and a wide area network (WAN). A DSL modem is the most common example of a small network choke point.
How Can It Be Prevented?
- The following strategies will help prevent congestion collapse:
- A mechanism to reorder or drop those packets under overload.
- An end-to-end flow control mechanisms designed at the end points can proactively respond to congestion and behave appropriately.
- A ‘slow start’ strategy will ensure that new connections do not further congest the traffic before the existing congestion is cleared.
- Random early detection (RED), in which packets are dropped randomly in a network when congestion collapse is expected.
- Fair queuing at choke points help avoid congestion by letting a small number of connections pass through them.
Ideally, network backbone should be provisioned with enough bandwidth to keep congestion at the periphery.
