QoS Provisioning power of optimal aggregate-flow scheduling
Huan Ren, Ki‐Hong Park · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2001
Providing scalable QoS-sensitive services to applications with varying degrees of elasticity using aggregate-flow scheduling is a challenging problem. Our previous work advanced a theoretical framework where an optimal differentiated services provisioning problem is formulated and solved to yield solutions for optimal per-hop behavior(PHB) and absolute QoS shaping using end-to-end control. In this paper, we build on our previous work and achieve the following innovations. First, we generalize the mean square error criterion employed in the previous work for measuring goodness of aggregate-flow scheduling by introducing a scaling function which is applied to the TOS field label value in the IP header. This allows the service provider to configure the PHB so as to export customized QoS differentiation---essential when shaping end-to-end absolute QoS over per-hop relative QoS---commensurate with the QoS profiles of its user base. Second, we put the theoretical framework and analysis ``to the test'' by presenting a comprehensive performance evaluation study of differentiated services as affected by optimal aggregate-flow per-hop control. In particular, we show that end-to-end QoS shaping power degrades gracefully as the number of service classes at routers in a multi-hop path is decreased, which admits heterogeneity in router configurations and incremental deployment. Collectively, our results demonstrate that relative QoS can be effectively harnessed to achieve absolute QoS in WAN environments if the PHB implements optimal aggregate-flow scheduling in the generalized framework.