A comparison of software architectures for data-oriented mobile applications

Mobile applications have empowered and extended the usability of mobile devices far beyond merely supporting voice communication. Mobile applications which rely on remote data sources and databases are particularly challenging given the need to engage in complex business logic and transmit data through wireless media. Previous research has looked at a variety of approaches to address these problems. Two software architectural approaches have emerged as the primary ways to develop mobile applications accessing remote data sources: client-agent-server and client-intercept-server. Both make use of middleware with the main difference being in the way agent components are used. In this work, we compare the performance of the two approaches under different scenarios. Statistical analysis shows that the client-agent-approach performs better. The results of this research provide useful guidelines for the development of mobile applications needing to connect to remote databases or data sources.