如何从海量服务中快速准确的发现目标服务是云计算研究的重点和难点。然而，目前云计算环境下服务描述和发现主要遵循pay-before-you-go(PBYG)模式，难以及时描述动态异构服务信息和准确高效的发现合适服务。针对上述问题，项目拟构建pay-as-you-go(PAYG)模式的服务描述和发现机理。项目引入数据空间和数据集成新思想，首先吸收极松散结构模型优点建立服务描述模型，再从分析服务演化特点入手，研究该模型动态演化机理，进而研究渐进式服务描述方法；然后，研究模型查询代数，设计延迟计算、服务相似度等算法实现支持best-effort查询的服务发现，同时，探究语义映射规则渐进生成和语义标注方法，形成有效的PAYG模式服务语义动态添加机制，实现该模式语义服务发现；最后，构建评估机制验证所提出的理论和方法。项目将为云环境服务描述与发现研究提供新思路和理论依据，为实时服务发现提供支持。

How to discover the target service from the mass services quickly and accurately is the major problem in cloud computing. However, services description and discovery in cloud computing environment have mainly followed the pay-before-you-go (PBYG) fashion, and it is thus difficult to describe the heterogeneous and evolving service information timely, and to discover appropriate service accurately and efficiently. To address above problem, this research project, which is inspired by new ideas in the fields of dataspace and data integration, will focus on how to describe and discover services in a pay-as-you-go(PAYG) fashion. First of all, a very loosely structured data model is presented to describe services and the relationships among them, and then the dynamic evolution mechanism of this model is constructed by analyzing the characteristics of service evolution, after which, the incremental service description methods are presented. Secondly, the model's query algebra is studied, and then the ways to lazily compute and query this model are provided for supporting best-effort queries. Thirdly, semantic mapping rules are generated gradually, and then an approach to annotate semantic description is also proposed, and thus an effective mechanism based on PAYG fashion, which adds semantics to the query processing, is constructed. It is to support semantic service discovery. Finally, the validity of the work in this project is examined by theory and experiment. The project will not only provide new ideas and theoretical basis for service description and discovery in cloud environment, but also support the real-time service discovery.