Abstract
Multicast can improve network performance by eliminating sending unnecessary duplicated flows in the data center networks (DCNs), thus it can significantly save network bandwidth and improve the network Quality of Service (QoS). However, the network multicast blocking causes the retransmission of a large number of data packets, and seriously influences the traffic efficiency of data center networks, especially for the multicast traffic in the fat-tree DCNs owing to multi-rooted tree structure. In this paper, we propose a novel multicast scheduling strategy to reduce the network multicast blocking. In order to decrease the operation time of the proposed algorithm, therefore, the remaining bandwidth the selected uplink connecting to available core switch should be close to and greater the three times than the bandwidth of multicast requests. Then the blocking probability of downlink at next time-slot is calculated by using markov chain theory. Furthermore, we select the downlink with minimum blocking probability as the optimal path at next time slot. In addition, theoretical analysis shows that the blocking probability of scheduling algorithm is close to zero and has lower time complexity. Simulation results verify the effectiveness of our proposed multicast scheduling algorithm.