Abstract: The paper proposes a new buffering and scheduling policy for data bursts (DBs) with fiber delay lines (FDL) , which is set at core nodes according to a specific configuration. The main idea is that when competing at a core node, the low-priority class data burst (DB) is buffered with FDL in order to forward the high-priority DB as early as possible, and the reserved wavelength channel for the low-priority DB will be re-assigned to the contending high-priority DB by sending a new burst control packet (BCP) , called the second BCP signaling, from the node to the downstream nodes. The effects of network traffic loads, the average length of DBs and the size of FDL buffer on DB’s loss probability and end-to-end delay are studied by an OBS network system simulation. Results show that the proposed scheduling policy can reduce obviously the DB loss probability and end-to-end delay for high-priority class DB. So it can improve significantly the end-to-end QoS for high-priority DBs.

Key words: optical burst switching, scheduling policy, fiber delay line, data burst, burst control packet