Research On The Models Of Scheduling And Market Transactions Of Cloud Service Resources

With the development and expansion of the informatics service market, and the informatics service requirements of the users increase day by day. The traditional resource organization and management has not been able to meet the needs of users in the cloud services market.On the other hand, the informatics service providers cannot afford the high cost of service, which prompt the emergence and the development of the service concept of "cloud computing"."Cloud computing" as a new form of infrastructure, software applications and platform of resource service, which can transfer the market service resources from the local to the network environment, and is provided to the users rapidly. Cloud users purchase, payment and customize cloud services, according to their personalized requirements, which can effectively reduce the investment and management cost. But at the same time, the service operation is not easy for most users, even for the computer professionals, that is also a time-consuming and labor-intensive thing, and cannot effectively guarantee the quality of service. Therefore, it can provide effective optimization management method in cloud service market, it can promise every link of cloud services markets orderly.So it is a problem to be solved immediately, which has a very important practical significance.According to the above problem,we use the basic market economic optimization theory, game theory, and the theory of multi-objective optimization to solve the problem. We analysis the problem by the related knowledge and methods of the specific cloud services market. First of all, we analyses the market mechanism of cloud services, from the perspective of the market competition mechanism, the resource scheduling mechanism, trading mechanism and pricing mechanism. Then we construct the optimization scheduling model of the cloud services resource, the trading model of the cloud service resource, the pricing model of the cloud service resource. The main research contents are as follows:(1) To the resource scheduling optimization problems in cloud services market, according to the highly relevant optimization distribution, we choose the service cost, the service time, the service Qos attribute value, the user's satisfaction as scheduling optimization index. We build the cloud service resources optimization scheduling model. During the process, in order to determine the different preferences of the users, we use the subjective and objective methods. We empower service resource attribute according to the different preferences of the users. Though this way, the resource scheduling cloud be more scientific and be more appropriate to reflect users, preferences. Take advantage of the utility function, we select the optimal service resource candidate sets. We use the service resources- tasks- providers to encode. We make the resources, tasks, and service providers together, which can provide the basis methods to complete the users requirements. Finally through the simulation, we analysis the effectiveness of the model.(2) We make the pricing strategy in the cloud services market. We do it respectively from the following two aspects: Firstly, according to user's utility maximization, we confirm the service contract and on-demand pricing strategy of the market. and then we research the changes of users' utility in the pricing strategy, according to the market under the two kinds of mechanism that was completely monopoly and competition introduced. Secondly, according to the market profit maximization, we determine the optimal market pricing strategy on the basis of the contract and on-demand pricing.Lastly, according to the actual market, there are plenty of idle resources in some time. At this point, we should do the spot instance pricing strategy for these idle resources. The pricing is to be regarded as a incentive mechanism to adjust users' behavior in the market, so as to realize the market equilibrium and the resource used effectively.(3) According to the service market crowded phenomenon, we set up the crowded pricing model. First of all, we set the congestion fee according to the priority and the cost of current,and then from two angles of the static and dynamic to analyze the rationality in decrease market pressure about the congestion pricing. Finally according to the game theory in information economics, we do the non-cooperative games between the users and the users, and the games between the users and the cloud agents. Above these two kinds of games are to set the congestion charge and the priorities policy of users in order to maximize the effectiveness of the market resources.(4) Based on the trade problems of the resource market service, this paper constructs the cloud services market trading system, which is the system by the cloud agents to coordinate the entire trade process. The cloud agents control the configuration of resources and the balance between supply and demand in market. And we put forward the multiple instances trading decision model, which is under the condition of the budget constraints, time constraints and budget and time constraints.