A cross-layer design framework, in the sense of maximizing sum-throughput, is proposed for a multi-service network of heterogeneous traffic. Considering different prescribed target packet loss rates and maximum automatic repeat request (ARQ) delays for the services in the data link control (DLC)-layer, we derive the respective optimum signal-to-interference-plus-noise ratio (SINR)-targets and variable spreading factor (VSF)s of the services in the physical (PHY)-layer, analytically as functions of the multi-user interference (MUI) and fading fluctuations. A multi-dimensional Markov chain is employed to model the packet traffic characteristics in the network. Performance of the proposed optimized cross-layer system is demonstrated and compared to state-of-the-art universal mobile telecommunication system (UMTS) standard, with conventional power, rate, and error control, for various practical system settings using theoretical and simulation results. Considerable improvement in sum-throughput performance of the multi-service mobile communication network is achieved through joint optimization of PHY-layer and DLC-layer parameters, whilst satisfying unique quality of service (QoS) constraints of different traffic types.Index Terms-Cross-layer design, dynamic power control, adaptive multi-rate transmission, packet error rate, automatic repeat request, multi-service, multi-dimensional Markov chain.