The Research Of Cloud Data Migration Policy Based On Swarm Intelligence Algorithm

Cloud computing has drawn more and more attention. At the same time the cluster and the amount of data started to surge ahead. All trades have emerged a large-scale cloud datacenter. The users’ requirements for bandwidth is also increasing with the demand of data transmission emerged the explosive growth trend in the cloud datacenter. The soaring amount of information makes the network bandwidth become a serious bottleneck problem. The data migration is a key part of ensuring the efficient operation, smooth upgrade and update system of the cloud data center, it is also occupy a pivotal position in the field of cloud computing. Efficiency and reliability of data migration can directly affect the performance of cloud data center. Data migration strategy is an important prerequisite for the implementation of data migration and more stable and efficient operation of the system to ensure the stability of the future. Good migration strategy not only save migration costs but also maintain and manage cloud data center better. When the load of the server is too much, the data traffic can be shared to the appropriate node device according to the dynamic migration strategy to realize load balancing. When some data is visited by a large amount, the network bandwidth congestion would lead to the cloud data center could’t provide efficient services. At this point, it is important to design an efficient dynamic data migration strategy for solving the problem.A part of the researchers found a suitable target location to meet the demand of migration and realized the load balance or optimized the operation cost of the cloud data center through dynamic data migration strategy. In this process, the location of the transfer destination is randomly selected, as long as the server can accommodate the data to be migrated, the data can be achieved migration, but the efficiency is not high, and from a long-term perspective, migration cost may also be great. In order to ensure the efficiency of migration, there are still a part of the researchers to achieve data migration through the optimal selection strategy, but only to the performance as the only target for cloud data center. It can temporarily achieve load balancing, but the strategy doesn’t consider migration cost and bandwidth pressure, it is still not able to solve the bandwidth bottleneck problem fundamentally.In this paper, we focus on the study of the maximum performance and save the cost of migration based on the target to find a destination server to achieve the migration of data through the dynamic data migration strategy. For this purpose, a novel heuristic algorithm IB-FA is proposed, which is based on the firefly algorithm to achieve the dynamic migration strategy of cloud data. The dynamic data migration strategy for load balancing is transformed into a multi-objective optimization problem with performance, migration cost and bandwidth constraints. In order to solve the multi-objective optimization problem, the IB-FA architecture of the cloud environment is proposed to adapt to the dynamic changes. In the contrast experiments of random data migration strategy and optimal data migration strategy the experimenta l results show that the IB-FA algorithm is able to find a more suitable for data migration position. The IB-FA algorithm disperses the I / O operations at the same time it can obviously optimizes the speed of data access and effectively improve the bandwidth utilization. So as to solve the bandwidth bottleneck in cloud data center.