随机自治可生存调度算法研究

生存性是指信息系统在面临攻击、失效等灾难情况下仍能继续提供服务的能力。它强调以任务为本的理念，是信息安全保障技术体系的一个重要环节。其目标是在保证基本服务安全的情况下对非基本服务的快速恢复。因此，减少基本服务集合是平衡生存能力和性能代价的途径之一。本文提出的随机自治可生存调度(SASS)算法在细粒度无集中控制 制的基础上实现服务漂移的动态调度，它使用扩充了的TCP协议实现广域网环境自治的细粒度连接迁移机制。基于该机制给出了在生存性意义上的服务器状态转换图，并描述了随机自治可生存调度算法。

On the Algorithm for Stochastic Autonomous Scheduling Survivability

Survivability is the ability for an information system to continue providing services under the disaster situation such as being attacked or malfunctioning and so on. It emphasizes the idea of mission critical, and is the important component of the technical framework for information assurance. Its goal is to recover ordinary services rapidly assuring the essential services are secure. So minimizing the essential service set is one of the ways of trading performance cost for survivability. The SASS algorithm proposed here realizes the dynamic scheduling of service roaming based on fine granularity without centralized control. With the expanded TCP protocol, it implements the autonomous fine-grained connection migration mechanism for WANs. A server state transition chart of survivability is given based on this mechanism, and the SASS algorithm is described.