Study And Application Of Red Algorithm In Obs Edge Node

With the emerging data services on the Internet explosive growing, wavelength division multiplexing technology has been developed rapidly to meet the people's demand for bandwidth. However, the inadequate optical switching ability of the current optical networks becomes the bottleneck for furtherly developing the performance of optical networks. Optical burst switching (OBS) technique is a practical switching technique, which is also a transitional technique between the optical circuit switching (OCS) and optical packet switching (OPS). In this dissertation, the queue management technique in OBS edge node is studied. And the performance of random early detection (RED) and two variants algorithms are simulated and discussed. Besides, the basic functions of OBS edge node are implemented by ARM9 platform. The author's works done in this dissertation are as follows:Firstly, the Active Queue Management (AQM) technique is adopted by the OBS edge node. Moreover, RED and two self-tuning RED (W-ARED and M-ARED) are simulated by OPNET concerning the performance of thoughtput of OBS networks. Simulation results show that RED, W-ARED (self-tuning the Wp parameter in RED) and M-ARED (self-tuning the maxp parameter in RED) improve the network performance.Secondly, the back door function is added into Linux kernel which can be preparation of flexibility in the relevant functional modules loaded into kernel. In addition, the OBS layer is added between the physical layer and network layer.Finally, the frameworks of Burst contrl packet (BCP) and Burst data packet (BDP) are designed. And corresponding data structures are analyzed. And the basic functions of OBS edge node such as the sending and receiving of BCP and BDP are implemented by ARM9.