A new architecture for scalable anonymous communication system（SACS） was proposed. The users were divided into several subgroups managed by different sub-blenders, and all sub-blenders were managed by the main-blend...A new architecture for scalable anonymous communication system（SACS） was proposed. The users were divided into several subgroups managed by different sub-blenders, and all sub-blenders were managed by the main-blender using two layers management scheme. The identity information of members are distributed on different sub-blenders, which makes each member keep much less information and network overload greatly reduce. The anonymity and the overhead of the new scheme were analyzed and compared with that of Crowds, which shows the cost of storage and network overhead for the new scheme largely decreases while the anonymity is little degraded. The experiment results also show that the new system architecture is well scalable. The ratio of management cost of SACS to that of Crowds is about 1：25 while the value of P（1｜H1＋） only increases by 0.001-0.020, which shows that SACS keeps almost the same anonymity with Crowds.展开更多
基金Projects(60403032) supported by the National Natural Science Foundation of ChinaProject (NCET-05-0683) supported by the New Century Excellent Talents in University, ChinaProject(IRT0661) supported by Changjiang Scholars and Innovative Research Team in University, China
文摘A new architecture for scalable anonymous communication system（SACS） was proposed. The users were divided into several subgroups managed by different sub-blenders, and all sub-blenders were managed by the main-blender using two layers management scheme. The identity information of members are distributed on different sub-blenders, which makes each member keep much less information and network overload greatly reduce. The anonymity and the overhead of the new scheme were analyzed and compared with that of Crowds, which shows the cost of storage and network overhead for the new scheme largely decreases while the anonymity is little degraded. The experiment results also show that the new system architecture is well scalable. The ratio of management cost of SACS to that of Crowds is about 1：25 while the value of P（1｜H1＋） only increases by 0.001-0.020, which shows that SACS keeps almost the same anonymity with Crowds.
