This paper introduces TCP Vegas, an implementation of TCP that significantly improves throughput and reduces losses compared to the Reno distribution of BSD Unix. The key techniques employed by Vegas include a more accurate retransmission mechanism, a proactive congestion avoidance mechanism, and a modified slow-start mechanism. The paper presents experimental results from both simulations and measurements on the Internet, demonstrating that Vegas achieves between 37% and 71% better throughput with one-fifth to one-half the losses compared to Reno. The retransmission mechanism reduces the time to detect lost packets, the congestion avoidance mechanism prevents unnecessary congestion, and the modified slow-start mechanism avoids packet losses during initial connection setup. The performance evaluation shows that Vegas does not negatively impact other TCP connections and can adapt to changing network conditions, leading to efficient bandwidth utilization.This paper introduces TCP Vegas, an implementation of TCP that significantly improves throughput and reduces losses compared to the Reno distribution of BSD Unix. The key techniques employed by Vegas include a more accurate retransmission mechanism, a proactive congestion avoidance mechanism, and a modified slow-start mechanism. The paper presents experimental results from both simulations and measurements on the Internet, demonstrating that Vegas achieves between 37% and 71% better throughput with one-fifth to one-half the losses compared to Reno. The retransmission mechanism reduces the time to detect lost packets, the congestion avoidance mechanism prevents unnecessary congestion, and the modified slow-start mechanism avoids packet losses during initial connection setup. The performance evaluation shows that Vegas does not negatively impact other TCP connections and can adapt to changing network conditions, leading to efficient bandwidth utilization.