针对智慧协同网络提出一种服务内容在传输路径上的缓存分配策略。该策略根据服务内容的流行度部署其在传输路径上的缓存位置,以求充分、高效地发挥网络缓存作用,进而提升网络的总体性能。所提分配策略分别在5层树型拓扑和由279个节点组成的真实网络拓扑中进行了性能测试。结果显示,该策略在所测的性能参数中表现出色,就平均服务获取距离而言,较命名数据网络(NDN,named data nerworking)所使用的LCE(leave copy everywhere)策略,其性能提高20%以上。
A cache allocation policy(CA) of service contents along delivery paths was proposed for the smart collaborative network. CA allocates on-path caches to service contents based on their popularity, expecting to fully and efficiently utilize the cache resource along the delivery path and further improving the network performance. The performance of proposed policy CA was evaluated under a 5-layer tree topology and a 279-node real network topology. The results show that CA outperforms others in terms of tested performance indexes. Compared with LCE(leave copy everywhere) used by named data networking(NDN), CA reduces the distance that users fetch service contents over 20%.