In an active network, users can insert customized active codes into activenodes, to execute. Thus it needs more resources than those required by conventional networks, andthese resources must be effectively monitored ...
详细信息
In an active network, users can insert customized active codes into activenodes, to execute. Thus it needs more resources than those required by conventional networks, andthese resources must be effectively monitored and managed. Management policies in existing OSs aretoo complicated to apply to simple active packets. In this paper, we present new resourcesmanagement policies that are mainly adoped to manage CPU, storage and transmission ***, we use SPF algorithm to schedule and process active packets, and import an interval queuemethod to allocate transmission bandwidth, and use feedback mechanism to control congestion . At thesame time, we design some experiments on prototype systems with and without resources managementpolicies respectively. The experiments results show that management policies presented by us caneffectively manage resources in active nodes and can improve the performance of active networks.
The dispensing mechanism of active code is the key technology in activenetwork. Conventional capsule and programmable switch approaches have their own shortcomings. DCCANmechanism presented in this paper overcomes the...
详细信息
The dispensing mechanism of active code is the key technology in activenetwork. Conventional capsule and programmable switch approaches have their own shortcomings. DCCANmechanism presented in this paper overcomes their shortcomings. In this paper, capsule andprogrammable switch, approaches are introduced and their shortcomings are analyzed. The principle ofDCCAN mechanism is described. The theoretical analyses in transmission bandwidth based on DCCANmechanism and capsule approach are described, and key factors which affect the transmissionbandwidth based on DCCAN mechanism are also discussed. At the same time, the theoretical analyses inlatency based on DCCAN mechanism and capsule approach are described. The using condition of DCCANmechanism is also discussed.
暂无评论