Fig. 1. Illustration of various burst scheduling scenarios. (a) LAUC-VF unnecessarily denies reservation to the new burst. (b) The new burst is scheduled by rescheduling burst 3. (c) Two rescheduling operations are required to schedule the new burst.

Fig. 2. Illustration of the main concept of Ordered Scheduling.

Fig. 3. Illustration of the bookkeeping for a typical time slot (with the slot size exaggerated and the unshown bursts occupying the whole slot) for the two implementations of the admission control test. For the basic version, N_occupied = 16. For the enhanced version, N_total = 16 and there are two entries in each of heads and ends.

Fig. 4. Illustration of the matching operation when some bursts fall entirely within the slot. (a) Actual burst pattern; (b) burst pattern seen by the algorithm.

Fig. 5. Timing diagram for scheduling a burst.

Fig. 6. Example of control packet formats. (a) Normal header packet format; (b) control packet formats in Ordered Scheduling.

Fig. 7. Illustration of the queueing model approach to analysing OBS. (a) OBS with advanced reservation; (b) corresponding queueing model.

Fig. 8. Probability distribution of packet length.

Fig. 9. Simulated OBS topology for a single core node.

Fig. 10. Burst dropping performance under different traffic loading.

Fig. 11. Effect of traffic composition on burst dropping performance. (a) Overall performance; (b) performance of the two classes.

Fig. 12. Overall dropping performance with different number of QoS classes.

Fig. 13. Burst dropping performance with different buffer depth. (a) Single QoS class; (b) two QoS classes.

Fig. 14. Burst dropping performance of basic Ordered Scheduling with different slot size.

Fig. 15. Burst dropping performance with different number of wavelengths per link.

Fig. 16. Simulation topology for whole network.

Fig. 17. Burst dropping performance for whole network under different traffic loading.