Date of Award
Master of Science (MS)
Dr. Haiying Shen, Committee Chair
Dr. Rong Ge
Dr. Walter Ligon
The exceptional rise in usages of mobile devices such as smartphones and tablets has contributed to a massive increase in wireless network trac both Cellular (3G/4G/LTE) and WiFi. The unprecedented growth in wireless network trac not only strain the battery of the mobile devices but also bogs down the last-hop wireless access links. Interestingly, a signicant part of this data trac exhibits high level of redundancy in them due to repeated access of popular contents in the web. Hence, a good amount of research both in academia and in industries has studied, analyzed and designed diverse systems that attempt to eliminate redundancy in the network trac. Several of the existing Trac Redundancy Elimination (TRE) solutions either does not improve last-hop wireless access links or involves inecient use of compute resources from resource-constrained mobile devices. In this research, we propose TailoredRE, a personalized cloud-based trac redundancy elimination system. The main objective of TailoredRE is to tailor TRE mechanism such that TRE is performed against selected applications rather than application agnostically, thus improving eciency by avoiding caching of unnecessary data chunks. In our system, we leverage the rich resources of the cloud to conduct TRE by ooading most of the operational cost from the smartphones or mobile devices to its clones (proxies) available in the cloud. We cluster the multiple individual user clones in the cloud based on the factors of connectedness among users such as usage of similar applications, common interests in specic web contents etc., to improve the eciency of caching in the cloud. This thesis encompasses motivation, system design along with detailed analysis of the results obtained through simulation and real implementation of TailoredRE system.
Soundararaj, Vivekgautham, "TailoredRE: A Personalized Cloud-based Traffic Redundancy Elimination for Smartphones" (2016). All Theses. 2387.