Resource Management Policy Research Foundation Cloud Bank Model

As an emerging IT service delivering model, Cloud computing can convert all kinds of computing resource into "products" which could meet demands from various types of customers. In commercial circumstances, to guarantee customers’ demands and experiences of services is very vital, and is also the key and core of the progress of Cloud computing. In order to guarantee QoS of customer, a Cloud platform needs to efficiently manage various computing resource in background. This procedure of resource management mainly includes:scheduling and allocating infrastructures among users’tasks, computing load balance over all compute nodes, and making end users to adjust their usage of resources automatically by certain outside methods such as price lever.By applying a big circumstance of Cloud Bank model, this thesis introduces theories and principles of microeconomics to manage resource outside and inside of a Cloud environment.The principal innovations in this thesis are:1. Cloud Bank has a life cycle, and the system needs to design different resource pricing policies in its various phases during the life cycle. This thesis constructs a dynamic game model and designs a general pricing policy, and gives the initial resource price in the initial phase of Cloud Bank via this model.2. this thesis extends2×2Pareto optimal theory to the MxN dimensional Cloud environment, thus infrastructure allocation in distributed environment could combine both efficient and impartial.3. the resource scheduling strategy and pricing policy are completely based on users’ QoS.