A Proposal to add Explicit Congestion Notification (ECN) to IP

A Proposal to add Explicit Congestion Notification (ECN) to IP

January 1999 | K. Ramakrishnan, S. Floyd
This document proposes adding Explicit Congestion Notification (ECN) to the Internet Protocol (IP). It outlines the need for ECN to improve congestion control by allowing routers to notify end systems of congestion without dropping packets. Current TCP relies on packet drops to indicate congestion, but this can lead to inefficiencies and unfairness. ECN enables routers to set a Congestion Experienced (CE) bit in packet headers, signaling congestion to end systems. This allows transport protocols to adjust their behavior without relying solely on packet drops. The proposal includes two bits in the IP header: the ECN-Capable Transport (ECT) bit and the CE bit. The ECT bit indicates that a transport protocol is ECN-capable, while the CE bit indicates congestion. TCP must be modified to support ECN, including negotiation during setup, new flags in the TCP header (ECN-Echo and Congestion Window Reduced), and adjustments to congestion control algorithms. ECN-capable TCP reduces its congestion window upon receiving a CE packet, similar to a dropped packet. The document discusses the importance of active queue management (e.g., RED) in conjunction with ECN to detect congestion before buffer overflow. It also addresses compatibility issues, incremental deployment, and the need for routers and end systems to support ECN. The use of ECN is expected to reduce unnecessary queueing delays and improve network performance. However, challenges remain, including ensuring that non-compliant end systems do not undermine ECN's effectiveness. The document also considers the interaction of ECN with other protocols and technologies, such as IPsec and tunnels. Finally, it highlights the benefits of ECN for real-time and delay-sensitive traffic, and the need for further research and standardization.This document proposes adding Explicit Congestion Notification (ECN) to the Internet Protocol (IP). It outlines the need for ECN to improve congestion control by allowing routers to notify end systems of congestion without dropping packets. Current TCP relies on packet drops to indicate congestion, but this can lead to inefficiencies and unfairness. ECN enables routers to set a Congestion Experienced (CE) bit in packet headers, signaling congestion to end systems. This allows transport protocols to adjust their behavior without relying solely on packet drops. The proposal includes two bits in the IP header: the ECN-Capable Transport (ECT) bit and the CE bit. The ECT bit indicates that a transport protocol is ECN-capable, while the CE bit indicates congestion. TCP must be modified to support ECN, including negotiation during setup, new flags in the TCP header (ECN-Echo and Congestion Window Reduced), and adjustments to congestion control algorithms. ECN-capable TCP reduces its congestion window upon receiving a CE packet, similar to a dropped packet. The document discusses the importance of active queue management (e.g., RED) in conjunction with ECN to detect congestion before buffer overflow. It also addresses compatibility issues, incremental deployment, and the need for routers and end systems to support ECN. The use of ECN is expected to reduce unnecessary queueing delays and improve network performance. However, challenges remain, including ensuring that non-compliant end systems do not undermine ECN's effectiveness. The document also considers the interaction of ECN with other protocols and technologies, such as IPsec and tunnels. Finally, it highlights the benefits of ECN for real-time and delay-sensitive traffic, and the need for further research and standardization.
Reach us at info@futurestudyspace.com