CAJ | 학술논문

为使实时测控系统满足未来多元化任务实时数据处理需求，设计了一个基于集群技术的实时测控数据处理系统。为解决集群系统的强实时性和高可用性问题，结合实时数据处理系统的特点，研究了实时集群体系结构、粒度划分和负载均衡方法。采用冗余部件、异常处理和向前恢复机制，消除了系统中的单一故障点；实现任务级粒度划分，优化了控制体系结构；采用基于任务分配表的动态负载均衡和人机交互管理相结合的方法，有效地解决了任务分配和故障迁移的实时性问题，提高了系统的可用性。
위사실시측공계통만족미래다원화임무실시수거처리수구，설계료일개기우집군기술적실시측공수거처리계통。위해결집군계통적강실시성화고가용성문제，결합실시수거처리계통적특점，연구료실시집군체계결구、립도화분화부재균형방법。채용용여부건、이상처리화향전회복궤제，소제료계통중적단일고장점；실현임무급립도화분，우화료공제체계결구；채용기우임무분배표적동태부재균형화인궤교호관리상결합적방법，유효지해결료임무분배화고장천이적실시성문제，제고료계통적가용성。
To meet the future demand for diversified real-time data processing tasks,a real-time measurement and control data processing system based on cluster technology has been designed. In order to solve the strong real-time performance and high availability of cluster system,combining the characteristics of real-time data processing system,the real-time cluster architecture,granularity division and load balancing method are studied. Using redundant components,exception handling and recovery mechanism,the single point of failure in the system is eliminated. The control system structure is optimized by implementing the task level granularity division. The dynamic load balancing based on task allocation table and the human-computer interaction management,effectively solve the real-time problem of the task allocation and fault migration,improving the availability of the system.