| Summary: | 博士 === 元智大學 === 資訊工程學系 === 106 === The development of early Cloud market has occurred largely from within a provider-driven approach, that is, the Cloud service providers generally predefine a common Cloud service without providing consumers with special customized requirements. Consumers can only buy the standardized Cloud services in the market and are at the end of the value chain. To establish a more open Cloud market ecosystem, the Cloud market has been shifting the services emphasis from reducing costs to focusing more on the customers demand, and the goal of market is not only to improve the benefit of the Cloud provider but also to satisfy the diverse need of the Cloud consumers. To that end, an efficient and autonomous negotiation of Service Level Agreement (SLA) is essential to resolve the conflicts between Cloud providers and Cloud consumers.
In this thesis, an agent-based fuzzy constraint-directed negotiation (AFCN) model is proposed for Service Level Agreements in the Cloud Computing. The proposed AFCN model supports a many-to-many bargaining negotiation infrastructure and provides a fully distributed and autonomous approach that does not require a third-party agent to coordinate the negotiation process. During the course of negotiation, each agent is endowed with beliefs about the market condition and its opponent’s behavior, with intentions to guide agent’s behavior, which represent the goal the agent would like to achieve.
The novelty of the proposed model is to add the fuzzy membership function to exchange information for representing the imprecise QoS preferences (e.g., task completion time, availability, and price etc.) that must be satisfied. This added information sharing is of critical importance for the effectiveness of distributed coordination because it not only reveals the opponent’s behavior preference, but also can specify the possibilities prescribing the extent to which the feasible solutions are suitable for agent’s behavior. Thus, the AFCN not only can improve the performance of negotiation, but also enforce a global consistent negotiation behavior through the iterative exchange of offers and counter-offers. Moreover, the AFCN model can flexibly adopt different behavioral strategies such as the competitive, win-win, and collaborative strategies for dynamic market environments which enable an agent to reach an agreement benefit all participants without reducing any of an agent’s desires.
To analyze the negotiation behavior and performance of AFCN, we used the CloudSim platform to simulate the Cloud market environment. The experimental results demonstrate that the proposed AFCN model outperforms the other agent-based approaches in terms of the level of satisfaction, the ratio of successful negotiation, the average revenue of PAs, the buying price of unit Cloud resource, and the convergence speed in the Cloud Computing market.
Moreover, we also proposed a two-tiered AFCN model for the Inter-Cloud environment. The two-tiered AFCN model can utilize the belief and intention information from first tier agent to guide the second tier negotiation and generate a more favorable proposal based on the integrated solution of two tiers. The experimental results demonstrate the proposed two- tiered AFCN model outperforms other agent negotiation models and give full play to the efficiency and scalability of Inter-Cloud federation.
|
|---|
