基于生物启发计算的数据密集型服务供应机制研究

Nowadays, data-intensive service provision, which is characterized by the large of scale of services and data and also the high-dimensions of QoS, are widespread in the area of scientific computing. However, most of the existing works failed to take into account the characteristics of data-intensive services and the effect of the big data sets on the whole performance of service provision, and they also seldom considered the efficiency of service selection and the reliability of composite services. There are many new challenges for service provision, especially in terms of autonomy, scalability, adaptability, and robustness. To tackle these problems, this project tries to utilize the features of biological systems to develop service provision mechanisms based on the bio-inspired computing, in order to optimize the whole cost while the performance has also been guaranteed. First, the QoS model of data sets will be developed to optimize the data access and selection of data replica. Then, by combining the scalability and the adaptability of the biological systems, the efficient service selection algorithms and the adaptable service composition strategies will be proposed respectively. Finally, an autonomic and dynamic negotiation mechanism will be investigated to guarantee the QoS attributes of the composite services. In summary, our project will deliver novel and effective service provision solutions for complex workflow applications in data-intensive scientific computing. This research will make significant contributions in both theories and applications of services computing.

当前数据密集型服务供应已广泛存在于不同领域的科学计算中，其具有规模大、QoS维度高等特点。但现有工作并未充分考虑数据密集型服务的特点，忽略了大数据集对服务供应性能的影响，且极少考虑数据密集场合下的效率和服务质量，从而使得服务供应面临自主性，可扩展性，可适应性和鲁棒性等方面的新挑战。针对以上问题，本项目从数据密集型服务供应面临的挑战入手，在确保性能的前提下以尽量降低服务供应的开销为目标，研究基于生物启发计算的数据密集型服务供应机制。通过建立数据集的QoS模型，优化服务对数据的访问和数据副本选择；通过结合生物系统的可扩展性和自主性，实现并行化的高效服务选择；通过结合生物系统的自适应性，实现动态自适应的服务组合，并提出自主动态的协商机制，有效保证组合服务的QoS属性。本项目的研究工作将为数据密集的复杂科学计算工作流的高效执行提供新颖且有效的解决方案，在服务计算领域具有重要的理论意义和应用价值。

本项目从数据密集型服务供应的特点，即规模大，组合环境动态多变，QoS维度高和数据主导出发，研究数据密集型场合下的服务供应的效率和服务质量，充分考虑服务组合在自主性、可扩展性、可适应性和鲁棒性方面面临的新挑战。首先研究数据集的规模，位置和开销对整个数据密集型服务供应过程性能的影响，建立优化数据访问和数据副本选择模型，为数据密集型服务选择提供有效支持。在此基础上针对数据密集型服务的特点，以满足全局和局部QoS约束条件为前提，建立单目标和多目标服务选择优化模型，并结合生物系统的可扩展性，设计有效且高效的服务选择算法。接下来研究自适应的服务组合策略，结合生物系统的自主性和自适应性，设计动态自适应的服务组合算法。最后研究数据密集型服务组合过程中的协商机制，以确保组合服务满足服务等级协议中的QoS约束条件。研究成果将促进Web服务选择和服务组合方法的丰富和发展，为数据密集的复杂科学计算工作流的高效执行提供新颖且有效的解决方案，并且将进一步发掘生物启发计算在解决数据密集型问题上的潜能。