基于深度哈希算法的云计算虚拟机迁移模型

Cloud Virtual Machine Migration Model Based on Deep Hash Algorithm

云计算中跨数据中心虚拟机迁移存在带宽小与无共享存储功能问题,导致迁移过程数据的安全性受到威胁。为降低带宽开销,提升抵御攻击能力,研究基于深度哈希算法的云计算虚拟机迁移模型。利用深度哈希算法获取需要迁移虚拟机基本镜像的类似程度,根据类似程度构建哈希图,依据哈希图构建云计算虚拟机迁移模型;通过上述模型中迁移代理主机实现虚拟机迁移时的信息传递;存储单元利用哈希图存储虚拟机基本镜像及镜像间的相似度;通过迁移单元迁移存储单元中需要迁移的基本镜像数据块,完成虚拟机迁移。实验证明,上述模型在不同负载时跨数据虚拟机迁移时间最短,带宽开销最低,并具备较优的抵御攻击能力,其SLA违反率既低又平稳。

Small bandwidth and unshared storage affect the migration performance of cross-data center virtual machines in cloud computing. In order to reduce the bandwidth overhead and improve the ability to resist attacks, the cloud computing virtual machine migration model based on a deep hash algorithm is studied in this work. According to the principle of the deep hash algorithm, the similarity of the basic image of the virtual machine to be migrated was obtained. The hash graph was built by similarity. Concurrently, a cloud computing virtual machine migration model was constructed based on a Hash graph. The information transfer during virtual machine migration was completed via the migration agent host in the above model. The storage unit was used to store the basic image of the virtual machine and the similarity between the images through the Hash graph. The virtual machine migration was completed by migrating the basic image data blocks in the storage unit according to the migration unit. The experimental results show that the model has a short migration time across data virtual machines, low bandwidth overhead, excellent attack resistance and a low and stable SLA violation rate(under different loads).