一种基于移动Agent的云端计算任务安全分割与分配算法

Security segmentation and allocation algorithm based on mobile Agent for cloud computing task

为了让在云端计算环境中运行的计算任务机密和隐私得到保障,避免恶意节点或竞争对手窥视任务的内部结构、系统逻辑和任务目标,必须构建一种全新的基于移动Agent的云端计算任务安全分割与分配算法。这种算法不仅要结合集群服务器节点的计算能力和特征,还要将用户终端节点的相关方面也考虑进去,把整个大的任务合理地分割成若干子任务,并将移动Agent作为这些子任务的数据结构和代码构成的载体;然后把这些数据和代码分配到与之对应的合理的任务执行节点上进行工作。通过将实验原型系统和该算法结合起来进行性能分析与探索,证明该算法对执行子任务的终端节点安全性的保障具有较强的有效性,即便有恶意节点窥探到了分配给该任务执行节点的数据和代码,甚至联合起来对系统进行攻击,也不会对这个任务的总体执行思路和整体目标的安全性造成威胁。

To guarantee the confidentiality and privacy of computing tasks operated in cloud computing environment,and avoid the malicious nodes or the competitors peering the internal structure,system logic and task objects,a new security segmentation and allocation algorithm of cloud computing task based on mobile Agent needs to be constructed. This algorithm combines with the computing ability and characteristics of cluster server nodes,and the relevant aspects of the user terminal nodes are taken into account. The entire task is segmented into some sub-task effectively,then the mobile Agent is taken as the carrier of data structure and code structure of the sub-task. The data and codes are allocated to reasonable task execution nodes corresponding to them. The performance is analyzed and explored by combining with experimental prototype system and this algorithm. It proved that the algorithm has better validity to guarantee the terminal nodes security of executing sub-task. When the malicious nodes peered the data and codes allocated to the task execution nodes,even the malicious nodes are combined together to attack on the system,it can not threaten the overall execution thought of the task and the security of the entire object.