Date of Award
Doctor of Philosophy (PhD)
Smart devices have become an essential component in the life of mankind. The quick rise of smartphones, IoTs, and wearable devices enabled applications that were not possible few years ago, e.g., health monitoring and online banking. Meanwhile, smart sensing laid the infrastructure for smart homes and smart cities. The intrusive nature of smart devices granted access to huge amounts of raw data. Researchers seized the moment with complex algorithms and data models to process the data over the cloud and extract as much information as possible. However, the pace and amount of data generation, in addition to, networking protocols transmitting data to cloud servers failed short in touching more than 20% of what was generated on the edge of the network. On the other hand, smart devices carry a large set of resources, e.g., CPU, memory, and camera, that sit idle most of the time. Studies showed that for plenty of the time resources are either idle, e.g., sleeping and eating, or underutilized, e.g. inertial sensors during phone calls. These findings articulate a problem in processing large data sets, while having idle resources in the close proximity. In this dissertation, we propose harvesting underutilized edge resources then use them in processing the huge data generated, and currently wasted, through applications running at the edge of the network.
We propose flipping the concept of cloud computing, instead of sending massive amounts of data for processing over the cloud, we distribute lightweight applications to process data on users' smart devices. We envision this approach to enhance the network's bandwidth, grant access to larger datasets, provide low latency responses, and more importantly involve up-to-date user's contextual information in processing. However, such benefits come with a set of challenges: How to locate suitable resources? How to match resources with data providers? How to inform resources what to do? and When? How to orchestrate applications' execution on multiple devices? and How to communicate between devices on the edge?
Communication between devices at the edge has different parameters in terms of device mobility, topology, and data rate. Standard protocols, e.g., Wi-Fi or Bluetooth, were not designed for edge computing, hence, does not offer a perfect match. Edge computing requires a lightweight protocol that provides quick device discovery, decent data rate, and multicasting to devices in the proximity. Bluetooth features wide acceptance within the IoT community, however, the low data rate and unicast communication limits its use on the edge. Despite being the most suitable communication protocol for edge computing and unlike other protocols, Bluetooth has a closed source code that blocks lower layer in front of all forms of research study, enhancement, and customization. Hence, we offer an open source version of Bluetooth and then customize it for edge computing applications.
In this dissertation, we propose Leveraging Resources on Anonymous Mobile Edge Nodes (LAMEN), a three-tier framework where edge devices are clustered by proximities. On having an application to execute, LAMEN clusters discover and allocate resources, share application's executable with resources, and estimate incentives for each participating resource. In a cluster, a single head node, i.e., mediator, is responsible for resource discovery and allocation. Mediators orchestrate cluster resources and present them as a virtually large homogeneous resource. For example, two devices each offering either a camera or a speaker are presented outside the cluster as a single device with both camera and speaker, this can be extended to any combination of resources. Then, mediator handles applications' distribution within a cluster as needed. Also, we provide a communication protocol that is customizable to the edge environment and application's need. Pushing lightweight applications that end devices can execute over their locally generated data have the following benefits: First, avoid sharing user data with cloud server, which is a privacy concern for many of them; Second, introduce mediators as a local cloud controller closer to the edge; Third, hide the user's identity behind mediators; and Finally, enhance bandwidth utilization by keeping raw data at the edge and transmitting processed information. Our evaluation shows an optimized resource lookup and application assignment schemes. In addition to, scalability in handling networks with large number of devices. In order to overcome the communication challenges, we provide an open source communication protocol that we customize for edge computing applications, however, it can be used beyond the scope of LAMEN. Finally, we present three applications to show how LAMEN enables various application domains on the edge of the network.
In summary, we propose a framework to orchestrate underutilized resources at the edge of the network towards processing data that are generated in their proximity. Using the approaches explained later in the dissertation, we show how LAMEN enhances the performance of applications and enables a new set of applications that were not feasible.
"Leveraging Resources on Anonymous Mobile Edge Nodes"
(2018). Doctor of Philosophy (PhD), Dissertation, Computer Science, Old Dominion University, DOI: 10.25777/v7rq-tz62