This document specifies TCP Friendly Rate Control (TFRC), a congestion control mechanism for unicast flows in a best-effort Internet environment. TFRC is reasonably fair when competing with TCP flows, but has lower throughput variation, making it suitable for applications like streaming media where a smooth sending rate is important. It obsoletes RFC 3448 and updates RFC 4342. TFRC uses a throughput equation based on loss event rate and round-trip time to determine the allowed sending rate. It is designed to be fair with TCP, but responds slower to bandwidth changes. TFRC is suitable for applications requiring smooth throughput, such as telephony or streaming media, but not for those needing rapid data transfer. TFRC is implemented as a receiver-based mechanism, with the receiver calculating the loss event rate. This is suitable for applications where the sender is a large server and the receiver has more computational resources. It can also be implemented in sender-based variants, as allowed in DCCP's CCID 3. The document outlines the protocol mechanism, including the TCP throughput equation, packet contents, data sender and receiver protocols, and implementation issues. It specifies how the loss event rate is calculated, how feedback packets are used to adjust sending rates, and how oscillations are reduced. The document also discusses the calculation of the loss event rate, the size of loss intervals, and history discounting to improve response to congestion changes. The document provides detailed specifications for TFRC, including the calculation of the loss event rate, the handling of feedback packets, and the adjustment of sending rates based on network conditions. It also discusses the implementation of TFRC in various scenarios and the differences from RFC 3448. The document includes security considerations, acknowledgments, and references.This document specifies TCP Friendly Rate Control (TFRC), a congestion control mechanism for unicast flows in a best-effort Internet environment. TFRC is reasonably fair when competing with TCP flows, but has lower throughput variation, making it suitable for applications like streaming media where a smooth sending rate is important. It obsoletes RFC 3448 and updates RFC 4342. TFRC uses a throughput equation based on loss event rate and round-trip time to determine the allowed sending rate. It is designed to be fair with TCP, but responds slower to bandwidth changes. TFRC is suitable for applications requiring smooth throughput, such as telephony or streaming media, but not for those needing rapid data transfer. TFRC is implemented as a receiver-based mechanism, with the receiver calculating the loss event rate. This is suitable for applications where the sender is a large server and the receiver has more computational resources. It can also be implemented in sender-based variants, as allowed in DCCP's CCID 3. The document outlines the protocol mechanism, including the TCP throughput equation, packet contents, data sender and receiver protocols, and implementation issues. It specifies how the loss event rate is calculated, how feedback packets are used to adjust sending rates, and how oscillations are reduced. The document also discusses the calculation of the loss event rate, the size of loss intervals, and history discounting to improve response to congestion changes. The document provides detailed specifications for TFRC, including the calculation of the loss event rate, the handling of feedback packets, and the adjustment of sending rates based on network conditions. It also discusses the implementation of TFRC in various scenarios and the differences from RFC 3448. The document includes security considerations, acknowledgments, and references.