This paper presents a simple analytic model of TCP throughput as a function of loss rate and round trip time (RTT), along with empirical validation. Unlike previous models, this model accounts for both TCP's fast retransmit and timeout mechanisms, which are crucial for accurate throughput prediction. The authors' measurements show that timeout events are more frequent than fast retransmit events in most TCP traces, highlighting the importance of including timeout effects in modeling. The model is validated against real-world TCP traces, demonstrating its ability to accurately predict throughput across a wide range of loss rates. The model also incorporates the impact of receiver-side window limitations, which can significantly affect throughput. The paper also extends the model to include timeout events and receiver-limited windows, providing a more comprehensive characterization of TCP behavior. The final model, referred to as the "full model," is shown to be more accurate than previous models, especially in cases where timeouts dominate. The paper concludes that timeouts have a significant impact on TCP performance and that the proposed model effectively captures this impact. The model is also used to derive a simplified approximation of TCP throughput, which can be applied in protocols designed to ensure TCP-friendliness. The authors acknowledge the contributions of various researchers and institutions in providing the data used in their analysis.This paper presents a simple analytic model of TCP throughput as a function of loss rate and round trip time (RTT), along with empirical validation. Unlike previous models, this model accounts for both TCP's fast retransmit and timeout mechanisms, which are crucial for accurate throughput prediction. The authors' measurements show that timeout events are more frequent than fast retransmit events in most TCP traces, highlighting the importance of including timeout effects in modeling. The model is validated against real-world TCP traces, demonstrating its ability to accurately predict throughput across a wide range of loss rates. The model also incorporates the impact of receiver-side window limitations, which can significantly affect throughput. The paper also extends the model to include timeout events and receiver-limited windows, providing a more comprehensive characterization of TCP behavior. The final model, referred to as the "full model," is shown to be more accurate than previous models, especially in cases where timeouts dominate. The paper concludes that timeouts have a significant impact on TCP performance and that the proposed model effectively captures this impact. The model is also used to derive a simplified approximation of TCP throughput, which can be applied in protocols designed to ensure TCP-friendliness. The authors acknowledge the contributions of various researchers and institutions in providing the data used in their analysis.