QoS-aware hierarchical token bucket (QHTB) queuing disciplines for QoS-guaranteed Diffserv provisioning with optimized bandwidth utilization and priority-based preemption

In most commercial IP routers, the traffic engineering is not fully supporting the optimized bandwidth sharing for QoS-guaranteed DiffServ provisioning with optimized bandwidth sharing that dynamically redistributes any unused bandwidth to backlogged flows and provides fast fault restoration with priority-based preemption. In this paper, we propose a QoS-aware hierarchical token bucket (QHTB) queuing discipline for QoS-guaranteed DiffServ provisioning. Each end-to-end aggregated traffic flow that includes realtime traffic and non-realtime traffic is described by traffic parameters, and link bandwidth is allocated according to the traffic parameters. If extra available bandwidth is available, it is re-distributed according to the predefined rates and backlog status. As a prototyping of the proposed architecture, we modified the hierarchical token bucket (HTB) kernel module in Linux IP/MPLS router. The performances of the proposed QoS-aware HTB (QHTB) in various traffic conditions have been analyzed with a series of systematic experiments. From the experimental results, we verified that the proposed QHTB provides optimized bandwidth utilization with guaranteed QoS provisioning for time-sensitive realtime service classes, while optimized bandwidth sharing among throughput-sensitive non-realtime service classes. Performance analysis of fast fault restorations with priority-based preemption is also provided that gives us some guidelines to efficiently schedule bandwidth sharing for i) QoS-guaranteed differentiated realtime service provisioning, ii) optimized bandwidth sharing with dynamic reallocation of unused bandwidth, and iii) fast fault restoration with priority-based preemption.