Year: 2019
Type: dissertation
Abstract: In the modern era of the Internet of Things and ubiquitous computing, autonomic management has become an essential component in the platforms because of its ability to govern resources adaptively from the sensor layer to the data center. In this thesis, we present an Autonomic Management System (AMS) that is designed to govern compute resources and network bandwidth for IoT applications. Security attacks such as economic Denial of Sustainability (EDoS) make the autonomic system vulnerable to wasting resources and also incur a cost for the application owners. We propose two frameworks using statistical techniques and Hierarchical Temporal Memory to thwart EDoS attacks for the AMS. The AMS has been implemented on the SAVI-IoT platform. The evaluation results demonstrate the system's ability to seamlessly adjust resources based on traffic workload. Further analysis proves its capability to reduce economic damage for the application owner and reduce resource wastage for the infrastructure.
Source: TSpace (University of Toronto)
Author: Rajsimman Ravichandiran
Institution: University of Toronto
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W3152724799
