【摘要】 跨广域网的虚拟机实时迁移是多数据中心云计算环境的重要技术支撑。当前跨广域网的虚拟机实时迁移受到带宽小和无共享存储的限制而面临着技术挑战,如镜像数据迁移的安全性和一致性问题。为此,该文提出基于哈希图(HashGraph)的跨数据中心虚拟机实时迁移方法,运用去中心化的思想,实现数据中心之间可靠和高效的镜像信息分布式共享。通过HashGraph中Merkle DAG存储结构,改善了重复数据删除在跨数据中心迁移虚拟机镜像时的缺陷。与现有方法相比,该文方法缩短了总迁移时间。更多还原

【Abstract】 Live migration of Virtual Machines(VMs) across WANs is an important support for multi-datacenter cloud computing environments. The current live migration of VMs across WANs faces many technical challenges due to the limitations of small bandwidth and no shared storage, such as ensuring the security and consistency of image data migration. Therefore, a method for VM live migration across data centers based on HashGraph is proposed in this paper, decentralized ideas are used to achieve reliable and efficient image distribution between data centers. The Merkle DAG of HashGraph improves the deficiencies of deduplication when migrating images across data centers. Compared with existing methods, it can reduce total migration time.更多还原