Doctoral Dissertation Oral Defense, Jiachen Wang

Title: Towards Efficient Resource Partitioning in Real-Time Wireless Networks

Ph.D. Candidate: Jiachen Wang

Major Advisor: Dr. Song Han

Associate Advisors: Dr. Shengli Zhou, Dr. Bing Wang

Date/Time: Wednesday, July 24th, 2024, 9 am

Location: HBL1947 room

Meeting link:
https://uconn-cmr.webex.com/meet/jiw19031

Abstract

The Internet of Things (IoT) has emerged as a transformative force in modern technology, enabling connectivity across everyday objects and significantly impacting various industries through the Industrial Internet of Things (IIoT). IIoT integrates industrial assets with information systems and business processes, streamlining manufacturing and optimizing industrial operations. Recent years have seen a shift towards real-time wireless networking (RTWN) technologies due to their deployment ease, mobility, and lower maintenance costs. Despite the widespread adoption of RTWN, challenges persist, particularly in network management. Current RTWN systems often struggle to balance network efficiency with real-time performance. Additionally, the dynamic nature of industrial networks further complicates management, necessitating prompt and adaptive reconfiguration to maintain reliable timing guarantees.

To address these challenges, we propose a series of partition-based resource management strategies for RTWN. The first approach focuses on minimizing end-to-end latency by creating adaptive time-dimension resource partitions, which align with different layers of the network routing graph. The second approach introduces a hierarchical resource partition framework, enhancing resource allocation efficiency and reducing reconfiguration overhead by constructing partitions subtree-by-subtree and limiting adjustments to nodes within these subtrees. To support multiple applications with diverse requirements on a shared RTWN infrastructure, we propose a resource virtualization framework. This framework allocates specific resources to each application, ensuring conflict resolution and isolation, thus maintaining seamless operation across applications. To further improve the applicability of the proposed real-time network virtualization framework on dynamic environments, we designed a flexibility-aware resource partitioning framework to construct partitions with predictable performance upon dynamics and maximize the availability.

Connect With Us