摘要
针对典型的云平台下虚拟化系统的特点,提出了一种结合选择性日志的准同步检查点算法VM_QSC:保持不同虚拟机节点固有的优化检查点周期,通过物理节点Hypervisor选择性地进行虚拟机的消息日志的稳定存储,在全局监控节点维护虚拟机一致线信息,保持全局的一致性。与传统的准同步检查点和同步检查点相比,该算法维持了虚拟机检查点设置的自主性,并显著降低了虚拟化系统的容错开销,可以有效应用于云计算环境下的虚拟资源管理和动态迁移。
Considering the characteristic of the virtual machines based on cloud platform,a Quasi-synchronous checkpointing with selective message logging algorithm for virtual machine (for short,VM_QSC) was presented.The algorithm keeps the inherent optimized checkpoint interval of the VM nodes,selectively stores stable optimistic message log,and maintains the consistency of the whole VM systems by hypervisor on the physical machines.Performance evaluation and simulation result show that in contrast with the traditional communication induced checkpointing and coordinated checkpointing,VM_QSC maintains the autonomy checkpointing,and saves communication and storage cost.It adapts for the cloud platform to manage the virtual resource and migrate the virtual machine
出处
《计算机科学》
CSCD
北大核心
2014年第5期46-49,共4页
Computer Science
基金
国家"八六三"高技术研究发展计划重大项目课题(2013AA01A205)
国家自然科学基金青年项目(61003047)
中央高校基本科研业务费专项资金(HIT.NSRIF.2014)资助
关键词
准同步检查点
选择性日志
卷回恢复
虚拟机
云计算
Quasi-synchronous checkpointing
Selective message logging
Rollback recovery
Virtual machine
Cloud computing