The paper introduces a novel congestion control protocol, eXplicit Control Protocol (XCP), designed to address the inefficiencies and instability of TCP in high bandwidth-delay product environments. XCP generalizes the Explicit Congestion Notification (ECN) proposal by introducing a more flexible and analytically tractable framework that decouples utilization control from fairness control. This decoupling allows for better service differentiation and improved performance in dynamic environments with varying traffic demands and RTTs. XCP's design is based on a control theory framework, which models the protocol as a linear feedback system with delay. The stability analysis shows that XCP is stable regardless of link capacity, round-trip delay, and number of sources, provided certain parameters are set correctly. Extensive packet-level simulations demonstrate that XCP outperforms TCP in both conventional and high bandwidth-delay environments, achieving high utilization, small queue sizes, and near-zero packet drops. XCP also exhibits better fairness, higher utilization, and smaller queue sizes compared to TCP, even in dynamic environments with many short web-like flows. The paper also discusses the practical implementation of XCP, noting that it requires minimal per-flow state in routers and few CPU cycles per packet, making it suitable for high-speed routers. Additionally, XCP facilitates distinguishing between error losses and congestion losses, which is useful for wireless environments, and enables the detection of misbehaving sources. The protocol's performance incentives suggest potential deployment paths, including coexistence with TCP in the same network.The paper introduces a novel congestion control protocol, eXplicit Control Protocol (XCP), designed to address the inefficiencies and instability of TCP in high bandwidth-delay product environments. XCP generalizes the Explicit Congestion Notification (ECN) proposal by introducing a more flexible and analytically tractable framework that decouples utilization control from fairness control. This decoupling allows for better service differentiation and improved performance in dynamic environments with varying traffic demands and RTTs. XCP's design is based on a control theory framework, which models the protocol as a linear feedback system with delay. The stability analysis shows that XCP is stable regardless of link capacity, round-trip delay, and number of sources, provided certain parameters are set correctly. Extensive packet-level simulations demonstrate that XCP outperforms TCP in both conventional and high bandwidth-delay environments, achieving high utilization, small queue sizes, and near-zero packet drops. XCP also exhibits better fairness, higher utilization, and smaller queue sizes compared to TCP, even in dynamic environments with many short web-like flows. The paper also discusses the practical implementation of XCP, noting that it requires minimal per-flow state in routers and few CPU cycles per packet, making it suitable for high-speed routers. Additionally, XCP facilitates distinguishing between error losses and congestion losses, which is useful for wireless environments, and enables the detection of misbehaving sources. The protocol's performance incentives suggest potential deployment paths, including coexistence with TCP in the same network.