With the beginning of the new millennium, the explosive growth of Internet users and Internet-related services continues to promote research for constructing high-speed optical systems and networks. Thus, the control of bandwidth in an efficient and s...
With the beginning of the new millennium, the explosive growth of Internet users and Internet-related services continues to promote research for constructing high-speed optical systems and networks. Thus, the control of bandwidth in an efficient and scalable way is eventually important to realize such a vision. Optical burst switching (OBS) has been proposed as a new switching paradigm for optical network. Compared with optical circuit switching, it provides efficiency and scalability by statistical multiplexing of bursts. In case of comparing with optical packet switching, it is easier to implement than OPS because of less stringent requirements in processing control signaling and achieving synchronization.
OBS networks usually employ one-way reservation by sending a burst control packet (BCP) with a specific offset time, before transmitting each data burst frame (BDF). due to such a property, burst-contentions when multiple bursts contend with the same wavelength for the same output link simultaneously in a switch, are occurred and lead to burst losses, eventually degrading the Quality of Service (QoS). There have been several studies in order to minimize the losses.
The nodes in OBS networks should conduct the scheduling of transit traffic and departure traffic(input traffic) simultaneously. The transit traffic has to pass for transmission to its destination node at this node. The departure traffic is generated at each node. Due to the feature of one-way reservation in OBS, the transit traffic and departure traffic unavoidably compete to reserve resource of same output port. This contention leads to data burst losses. If transit traffic which is spent resource passing over several nodes is dropped, the waste rate of network is increased dramatically. Otherwise, if departure traffic is dropped, the waste rate of network isn’t increased. If departure traffic doesn’t be transmitted well to guarantee transmission of transit traffic, the queue size and delay of departure traffic is increased significantly.
Therefore, in this paper, we propose the input traffic flow control (ITFC) algorithm considering drop rate of transit traffic and queue size of departure traffic. The proposed ITFC algorithm decides inflow of departure traffic considering drop rate of transit traffic firstly. If queue size of departure traffic excesses predefined value, ITFC algorithm decides proper inflow of departure traffic considering queue size of departure traffic and drop rate of transit traffic simultaneously. We use the OPNET simulator to evaluate the performance of the proposed ITFC algorithm and conventional departure burst shaping algorithm in terms of resource utilization and drop rate. We demonstrate the performance of the proposed ITFC algorithm is more better than existing one.