All ETDs from UAB

Advisory Committee Chair

Ragib Hasan

Advisory Committee Members

Alan Sprague

Purushotham Bangalore

Anthony Skjellum

Allen Johnston

Document Type

Dissertation

Date of Award

2017

Degree Name by School

Doctor of Philosophy (PhD) College of Arts and Sciences

Abstract

Cloud computing has become the dominant computing paradigm in recent years. As cloud computing technologies evolved, researchers have explored the possibility of building clouds utilizing ubiquitous mobile computing devices and extend the domain of cloud services from traditional SAAS, PAAS, or IAAS models to distributed storage, computation, and sensing as a service models. Given the number of unused mobile devices in our environment and the invention of low cost and powerful IoT devices, we hypothesize that building a mobile cloud will be economically feasible and will provide cloud services at cheaper rates than existing cloud computing models. We prove our hypothesis by developing a prototype of mobile cloud and run sample tasks on our cloud to show that the cost of computation in our proposed mobile cloud is lesser than existing and conventional cloud computing models. Conventional cloud computing platforms are also unable to provide cloud-based services in the absence of a continuous internet connectivity. Delay Tolerant Network (DTN) based approaches have considered the use of mobile devices for message forwarding over Wifi or Bluetooth networks to efficiently deliver contents from a source node to a destination node in a delay tolerant environment. Hence, we integrate the contextual information of mobile devices and the end-users into a DTN enabled mobile cloud computing framework for providing cloud-based services in a network-disconnected area. Finally, with the arrival of multiple mobile cloud providers with different business policies, it has become increasingly difficult for end users to select the best cloud platform which suits the requirement specifications for the intended service. We hypothesize that a user might need to use resources from multiple providers for deploying his service to be able to find the best match with the service requirements. Therefore, we propose a middleware that selects a set of cloud service providers based on users’ specifications and performs the necessary task scheduling and resource management associated with service deployment for end users.

Share

COinS