【摘要】 在虚拟机全系统在线迁移中,由于镜像文件数据量巨大,对整个迁移过程的效率有着关键的影响,因此优化迁移时间,成为虚拟机迁移技术的研究热点.对于以分布式存储系统(其中较为常见的是Ceph块设备)作为镜像文件存储方式的虚拟机进行迁移时,镜像文件需要经过源存储节点到源计算节点,再到目的计算节点,最后到目的存储节点.这种方式忽略了底层存储系统特点可以给迁移带来的好处,针对上述问题,提出了一种Ceph块设备跨集群迁移算法,采用源存储节点并行向目的存储节点迁移数据的方式,利用了存储节点的计算和网络能力.实验表明,该算法加快了迁移速度,同时适当增加存储节点数目能进一步提升算法效率.更多还原

【Abstract】 The migration time of huge image files is vital to the efficiency of the whole-system on-line migration of a virtual machine. Therefore, optimizing the migration time has become a hot research area in virtual-machine migration technology. For virtual machines based on a distributed storage system, in which the more common one is Ceph block device, image data must go from the source storage nodes to the source client nodes, then to the destination client nodes, and finally to the destination storage nodes. This way ignores the benefit from the storage system to the migration. Given the above problems, a migration algorithm for Ceph block device cross clusters is effective. Image data go from the source storage nodes to the destination storage nodes in parallel, which uses the network of storage nodes.The result of the experiment shows that this algorithm shortens the migration time,and a few more storage nodes can improve efficiency of this algorithm.更多还原