Development of new scheduling strategies for wireless application offloading

Smartphones are no longer used only for voice communication; instead, they are used for web surfing, GPS navigation, acquiring and watching videos and photos, gaming, and many other purposes. As consequence, these systems consume more power and shorten the battery life. Even though the battery technology has been continuously improving, it has not been able to keep up with the rapid growth of power consumption of these mobile devices. As a result, energy consumption has become a primary constraint for battery-powered mobile systems. On the other hand, mobiles terminals (MTs) have limited computation resources and thus there is an increasing gap between the demand for complex applications and the availability of the required resources for executing such applications in mobile devices. Cloud computing is a flexible and costeffective concept that allows MTs to have access to larger computational resources than those available in typical terminals. These computational resources, such as processing, memory and storage, are located in remote devices (i.e. servers) and the terminals access them via mobile wireless channels. Cloud computing may extend the battery life by migrating the energy-intensive parts of the computation to the remote server. On the other hand, small cells deployments can be seen as an opportunity to offer low-cost solutions for cloud computing services if the small cells are equipped with some enhanced computational and storage capabilities. In a multiuser scenario, the available resources must be shared among the different users, including the radio resources required for the communication between the MT and the small cell in the uplink and downlink, and the processing resources located at the remote processor. The main objective of this master thesis is to develop new scheduling strategies for this scenario in order to improve the quality of service (QoS) perceived by the different users in terms of delay, while achieving a target energy saving. In this sense, we have considered two different approaches: three decoupled schedulers, each one managing the uplink, processing and downlink resources independently, and a single scheduler which allocates jointly the communication and computation resources.