Cloud Services Brokerage for Mobile Ubiquitous Computing

Recently, companies are adopting Mobile Cloud Computing (MCC) to efficiently deliver enterprise services to users (or consumers) on their personalized devices. MCC is the facilitation of mobile devices (e.g., smartphones, tablets, notebooks, and smart watches) to access virtualized services such as software applications, servers, storage, and network services over the Internet. With the advancement and diversity of the mobile landscape, there has been a growing trend in consumer attitude where a single user owns multiple mobile devices. This paradigm of supporting a single user or consumer to access multiple services from n-devices is referred to as the Ubiquitous Cloud Computing (UCC) or the Personal Cloud Computing. In the UCC era, consumers expect to have application and data consistency across their multiple devices and in real time. However, this expectation can be hindered by the intermittent loss of connectivity in wireless networks, user mobility, and peak load demands. Hence, this dissertation presents an architectural framework called, Cloud Services Brokerage for Mobile Ubiquitous Cloud Computing (CSB-UCC), which ensures soft real-time and reliable services consumption on multiple devices of users. The CSB-UCC acts as an application middleware broker that connects the n-devices of users to the multi-cloud services. The designed system determines the multi-cloud services based on the user's subscriptions and the n-devices are determined through device registration on the broker. The preliminary evaluations of the designed system shows that the following are achieved: 1) high scalability through the adoption of a distributed architecture of the brokerage service, 2) providing soft real-time application synchronization for consistent user experience through an enhanced mobile-to-cloud proximity-based access technique, 3) reliable error recovery from system failure through transactional services re-assignment to active nodes, and 4) transparent audit trail through access-level and context-centric provenance