Research On Container Cloud Resource Placement And Migration Strategy In Multi-tenant Environment

With the vigorous development of container cloud platforms,the number of container services hosted by the cloud center is increasing,and the scale of tenants is also increasing.Tenants are entities that pay to the cloud center and publishers of cloud tasks.It is particularly important to ensure quality of service(Qo S)for tenants in the multi-tenant container cloud platform.At the same time,the dynamic change of node load will lead to container migration.The traditional single-target migration strategy can not meet the differential needs of a large number of tenants.Based on the above problems,this thesis focuses on the placement and migration strategy of container cloud resources in multi-tenant environments,to ensure the Qo S of tenants and reduce the operation cost of the cloud center.The main work of this thesis is as follows:1.Most of the container allocation polices of the current container cloud platform are modeled from the perspective of tasks or containers,which can not guarantee the Qo S of tenants.Therefore,a cost-effective container resource placement strategy is proposed.From the perspective of tenants,the strategy constructs the tenant Qo S model and tenant default cost model according to tenant level,and constructs the resource idle cost model of the container cloud platform.The improved whale optimization algorithm(WOA)is used to solve the container set to be placed with the minimum tenant default cost as the goal,and the WOA and greedy algorithm are used to solve the optimal container placement scheme with the minimum resource idle cost as the goal.Finally,simulation experiments show that the proposed strategy can improve the resource utilization of the cloud platform and reduce the operation cost of the cloud center while ensuring the Qo S of tenants.2.The dynamic change of node load will affect the stable operation of the platform.Container migration technology is usually used to balance the system load,and container migration will seriously affect the service quality of tenants.Based on this,a container migration strategy based on a multi-objective selection algorithm(CMS-MCSA)is proposed.The strategy uses the node load detection algorithm to screen the nodes to be migrated,defines the container’s favor and rejection of node resources,constructs a multi-objective container selection evaluation model,and obtains the migration container set and target node-set by solving the objective function of the model.The migration strategy proposed in this thesis is compared with the Mu strategy and MMT strategy.The results show that the CMS-MCSA strategy can reduce the SLA default rate of container migration while ensuring fewer migration times and lower migration times,and has better cluster load balance.The research work shows that the container resource placement strategy and CMS-MCSA container migration strategy proposed in this thesis can not only effectively reduce the tenant SLA default rate,but also improve the resource utilization of the container cloud platform,which has good practical significance.