RFC 5681

RFC 5681
Title	RFC 5681
Author	Sally Floyd; Van Jacobson
Published	September 2009
Series	Request for Comments
Number	5681
Status	Standards Track
Pages	60

RFC 5681 RFC 5681 is the IETF standards-track document that updates and clarifies congestion control behavior for the Transmission Control Protocol used across the Internet. It specifies algorithms, sender and receiver requirements, and interactions with earlier work to ensure interoperability among implementations from vendors, research labs, and standards bodies. The document builds on preceding studies and recommendations from organizations and events that shaped Internet architecture.

Background and Purpose

RFC 5681 derives from long-standing research and operational experience dating to experimental networks such as ARPANET, contributions from institutions like Xerox PARC, and standards efforts led by Internet Engineering Task Force, Internet Architecture Board, and working groups including TCPM Working Group. It codifies congestion avoidance and control principles originally motivated by congestion collapse incidents analyzed in reports connected to MIT, Stanford University, and the University of California, Berkeley research communities. The document aims to reconcile differing implementations from vendors such as Cisco Systems, Juniper Networks, and Sun Microsystems while referencing prior specifications including those from RFC 1122 and influential papers by researchers at Lawrence Berkeley National Laboratory and Bell Labs.

Congestion Control Algorithms Specified

The specification formalizes additive increase/multiplicative decrease (AIMD) behavior originally advocated in literature by researchers at MIT, Bell Labs, and Stanford University, and traces practical mechanisms to implementations from BSD derivatives and stacks in products by Microsoft and Apple Inc.. It defines slow start, congestion avoidance, fast retransmit, and fast recovery, aligning with algorithmic descriptions found in university theses from UC Berkeley and proceedings at conferences like SIGCOMM and ICNP. The text references experimental comparisons performed by teams at IETF meetings and academic venues such as IEEE INFOCOM.

TCP Reno and TCP NewReno Behavior

RFC 5681 clarifies behavior for established variants including TCP Reno and TCP NewReno, differentiating their fast retransmit and fast recovery interactions as implemented historically in FreeBSD, NetBSD, OpenBSD, Linux kernel, and stacks in appliances from Cisco Systems and Sun Microsystems. It prescribes retransmission timeout and duplicate acknowledgment handling consistent with analyses presented at ACM SIGCOMM 1988 and later RFCs influenced by researchers at UC Berkeley and Bell Labs. The document situates these behaviors relative to experimental proposals from research groups at University of Washington and Carnegie Mellon University.

Sender and Receiver Requirements

RFC 5681 stipulates sender-side requirements for congestion window management and receiver-side behaviors for acknowledgement generation, reflecting interoperability considerations between implementations from Microsoft, Apple Inc., Google, and open-source projects like the Linux kernel and FreeBSD. It specifies how receivers should generate acknowledgements to enable fast retransmit as observed in deployments by network operators including Sprint Corporation and AT&T. The requirements echo testing outcomes presented at events such as IETF meetings and workshops hosted by IAB and IETF Transport Area contributors.

Implementation Considerations and Interoperability

The document discusses implementation choices and trade-offs faced by developers at organizations such as Cisco Systems, Juniper Networks, Intel Corporation, and IBM when integrating congestion control into operating systems like Windows NT, macOS, and Linux kernel. It emphasizes interoperability testing practices promoted by groups like IETF and standards committees within IEEE and references empirical evaluations from research labs at Lawrence Berkeley National Laboratory and Bell Labs. Deployment considerations include interaction with middleboxes produced by vendors like Checkpoint Software Technologies and measurements reported by network measurement projects at RIPE NCC and ARIN.

Security and Misuse Considerations

RFC 5681 considers misuse scenarios and potential attacks relevant to implementers at institutions such as Google and Microsoft and operators at carriers like Verizon Communications. It addresses how crafted packet sequences could influence congestion control state, a topic investigated in academic research from Carnegie Mellon University and ETH Zurich and discussed at security conferences such as USENIX Security Symposium and NDSS. The section encourages implementers to consider robustness against middlebox interference from vendors like Akamai Technologies and F5 Networks.

History and Impact on Networking Standards

RFC 5681 consolidates decades of work from researchers and organizations including Bell Labs, MIT, Stanford University, UC Berkeley, and standards forums such as IETF and IAB. Its guidance influenced subsequent standards and implementations from industry players like Cisco Systems, Juniper Networks, Microsoft, and open-source communities around the Linux kernel and FreeBSD. The document served as a baseline for later congestion control research presented at SIGCOMM, IEEE INFOCOM, and ACM CoNEXT, and informed protocol extensions developed by teams at Google and Facebook.

Category:Internet protocols