Optimizing DaaS Web Service Based Data Mashups

Data Mashup is a special class of mashup application that combines information on the fly from multiple data sources to respond to transient business needs. In this paper, we propose two optimization algorithms to optimize Data Mashups. The first allows for selecting the minimum number of services required in the data mashup. The second exploits the services´ constraints on inputs and outputs to filter out superfluous calls to component services in the data mashup. These two algorithms are evaluated and tested in the healthcare application domain, and the reported results are very promising.