基于向量时钟的动态自适应同步策略研究  

A Study of Dynamic Adaptive Synchronization Strategies Based on Vector Clocks

作　　者：王艺遥 齐和平[1] 李学艳[1] 毛玉辉 柴彪 WANG Yiyao;QI Heping;LI Xueyan;MAO Yuhui;CHAI Biao(North Automatic Control Technology Institute,Taiyuan 030006,China)

机构地区：[1]北方自动控制技术研究所,太原030006

出　　处：《火力与指挥控制》2022年第7期108-112,117,共6页Fire Control & Command Control

基　　金：“十三五”共用技术装备预研基金资助项目(41401040103)。

摘　　要：针对分布式仿真系统信息交互的性能优化,研究时钟不一致、传输时延等引起的网络拥塞、因果倒序、时钟错乱等问题,是保证分布式仿真系统可靠运行的重要因素。以多服务器架构的分布式仿真系统为研究对象,考虑服务器与仿真成员之间的信息传输特性,根据向量时钟和传统的保守策略、乐观策略,提出一种新的动态自适应有界时间窗算法。算法正确处理系统中事件消息的因果逻辑,同时保证同步策略的可重复性,通过实验验证了算法的可行性。In view of the performance optimization of information interaction of distributed simulation system,it is an important factor to ensure the reliable operation of distributed simulation system to study the problems of network congestion,causal reverse sequence and clock disorder caused by clock inconsistency and transmission delay etc. Taking the distributed simulation system of multi-server architecture as the research object,considering the information transmission characteristics between servers and simulation members,and according to vector clock and traditional conservative strategy,optimistic strategy,a new dynamic adaptive bounded time window algorithm is proposed. The causal logic of event messages in the system is handled by the algorithm,while the repeatability of synchronization strategies is ensured. Finally,the feasibility and performance of the algorithm are verified by experiments.

关 键 词：分布式仿真 保守策略 乐观策略 向量时钟 动态自适应策略 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]