Abstract: A thought of parallel computing is applied to IM information reversion by putting forward a parallel computing model based on tuplespace. Thinking size and picking it automatically, the model realizes dynamic load balancing. It achieves stability of reversion, through an improved task-taking algorithm. Theoretic analysis and experiments demonstrate that the model can improve the performance effectively.

Key words: presence and instant messaging, information reversion, parallel, tuplespace

摘要： 在即时通信信息还原过程中，引入分布式并发处理的思想，提出基于元组空间的并行处理模型。该模型使用细化任务粒度和主动提取任务相结合的方法，实现自适应的动态负载均衡。通过对任务提取算法的改进，分析还原的可靠性。理论分析和实验测试证明，该模型有效地提高了信息还原的速度。

关键词: 即时通信, 信息还原, 并行化, 元组空间

CLC Number: 

• TP393.02