Performance Issues in Optical Burst/Packet Switching
Towards Digital Optical Networks
Agile bandwidth management techniques in slotted all-optical packet switched networks
Computer Networks: The International Journal of Computer and Telecommunications Networking
Reservation techniques in an OpMiGua node
ONDM'07 Proceedings of the 11th international IFIP TC6 conference on Optical network design and modeling
Research in optical burst switching within the e-Photon/ONe network of excellence
Optical Switching and Networking
Design and test of a software defined hybrid network architecture
Proceedings of the first edition workshop on High performance and programmable networking
Hi-index | 0.00 |
We show that our proposed optical packet switched (OPS) hybrid network model supports both high throughput efficiency and guaranteed service transport (GST) with no packet loss and constant delay. The network comprises a wavelength routed optical network (WRON) enabling GST, and a statistically multiplexed (SM) network enabling high throughput. In the WRON, packets follow a fixed wavelength path and forwarding is based on the packets wavelength, while in the SM network, packets are switched according to header information. High reliability of GST packet forwarding is possible, because the forwarding does not depend on the operation of the packet switch. A novel node design supporting three Quality of Service (QoS) classes is proposed and described in detail. It facilitates full sharing of the link bandwidth by segregating GST packets and SM packets using polarization time division multiplexing (PTDM). SM packets are differentiated into two specified sub-QoS classes employing a novel buffer reservation technique and asynchronous buffer scheduling algorithm AIP3Q. We present a detailed analysis of the node throughput performance, demonstrating that buffering resource requirements are reduced in the optical packet switch, because processing and buffering of GST packets is avoided. The SM packet QoS differentiation permits a further reduction of buffer resources