Service provisioning from resource constrained mobile devices

The primary goal of this thesis is to investigate how the service-oriented architecture can be extended to mobile environments, facilitating resource constrained mobile devices to host services. With the advances in mobile devices and wireless communications, the opportunities to run resource-demanding applications on the mobile devices are growing. The increasing processing power, storage and support of multiple network interfaces enable the mobile devices to host services and participate in service discovery networks. These characteristics facilitate that mobile devices can act both as service consumers and service providers. Besides, roaming of mobile devices across wireless networks provides nomadic characteristics to the service providers to be available anytime and anywhere. As in web services architecture, service consumers and providers are loosely coupled and dynamically bound; this architecture is especially advantageous for the frequently changing mobile environments. -- Nevertheless, service provisioning from mobile devices is challenging, as the resources are far more constrained compared to the Internet servers that are originally targeted for web services technologies. Mobile devices have limited processing power and memory space; and suffer from lower bandwidths, higher error rates, and frequent disconnections compared to fixed networks. A few efforts have been taken to verify the feasibility of provisioning web services from mobile devices. However, they have not addressed the issue about how to host heavy-duty services on resource-constrained mobile devices with limited processing power, memory space, and transmission bandwidth. -- This thesis attempts to overcome the issues by introducing a framework which partitions the workload of resource-demanding services involving complex business processes and keeps the web service interfaces on mobile devices. The framework effectively leverages the capability of mobile devices, by offloading the partitioned computing load to resource-rich surrogates. Therefore, the mobile device works as the integration point with the support of surrogate nodes and other web services. The functions that require the resources of the mobile device or interaction with the mobile user are executed locally. -- The framework introduces an efficient partitioning approach for execution offloading from resource constrained mobile devices. The proposed approach considers the dynamic status of the resources of a mobile device to partition the tasks effectively. The framework provides a distributed platform for executing services, which adaptively offloads by considering both the interaction properties and the resource consumption of the tasks. The designed framework is analyzed using prototype experiments and the results have shown the effectiveness and efficiency of the approach in provisioning heavy-duty services from resource constrained mobile devices.