Having in mind that real-time streams tolerate some deadline misses according to (m,k)-firm constraints [1], this paper presents a solution that consists in integrating (m,k)-firm temporal requirements into the guaranteed–rate quality of service such as IntServ Model to provide both bandwidth and delay guarantees as required by real-time applications. Analytic study using Network Calculus gives the upper bound on delay guaranteed by the proposed approach and shows better behavior compared to the classical IntServ QoS model.