Quality of service differentiation, teletraffic analysis and network layer packet redundancy in optical packet switched networks

摘要: Optical Packet Switching (OPS) has emerged as a promising candidate for the next-generation Wavelength Division Multiplexed (WDM) based alloptical network. By enabling packet switching in optical domain, OPS networks can provide cost-efficient and transparent transport services to higher layers. However, commercial deployment of requires not only maturation several key technologies, but also thorough investigation number networking challenges related OPS, since are fundamentally different from today’s store-and-forward networks. This thesis addresses latter issue by considering following three issues:· Quality Service (QoS) differentiation at WDM layer, with focus on loss rate (PLR) delay-jitter differentiation.· Teletraffic analysis networks.· How combat using network layer redundancy.First, crucial is due contention. Contention occurs when destined wavelength currently occupied another packet. Several approaches such have been proposed recent literature, e.g. utilizing conversion, buffering, deflection routing or traffic shaping.This considers novel approach OPS: The Network Layer Redundancy Scheme (NLPRS) allows redundancy packets be injected into network, thus reconstruction lost data egress node. Results show that NLPRS able reduce end-to-end PLR orders magnitude an asynchronous ring without conversion.Another QoS layer. Due lack random access memory, existing schemes suitable point-to-point architecture feasible use Hence, new utilize needed.A preemption scheme, Preemptive Drop Policy (PDP), bufferless OPS. With PDP, high priority arrivals allowed preempt take over busy low case results lower compared traffic. PDP extended Adaptive (APDP), which provides absolute guarantees ivtraffic measurement probability parameter adjustment.An access-restriction Allocation algorithm (WA), studied. In WA, full range output converters, certain wavelengths fibre exclusively reserved traffic.When (with respect PLR) introduced it shown average throughput decreases, often referred penalty introducing differentiation. main cause this because resources must used non-optimal manner employing isolate service classes. thesis, found For slotted stays same after introduction differentiation.An evaluation framework quantifying proposed. Using framework, fundamental including evaluated. It preemptive techniques result lowest penalty, followed dropping techniques. because, preemption, dropped port congested. access-restriction, highly strained, statistically dropping, independently state port.A scheme isolates classes ensuring transmitted time-slot does reduced isolated.QoS share pernode (SPN) contention resolution pool consisting Tunable Converters (TWCs) Fibre Delay Lines (FDLs) particular, may differentiated switch architecture.Analytical models some derived, providing explicit PLR. addition, analytical regarding derived both particularly useful studying effects nonuniform Furthermore, Erlang Engset model switch. Regarding model, how blocking evaluated vusing either calls cleared (LCC) overflow (OFL) model. all models, time-, call- congestion derived. A numerical presented reveals there small, non-negligible, deviation between observed probabilities, depends input/output fibres system load.