Background and Motivation
Edge computing infrastructures are increasingly adopted to enable low latency, high availability, and localized data processing. Unlike cloud environments, edge infrastructures operate under resource-constrained conditions, including limited computational resources, unreliable network connections, intermittent power supply, and stringent data persistence requirements. Kubernetes (K8s) orchestrates application deployment through specific workload abstractions such as ReplicaSet, StatefulSet, DaemonSet, Jobs, and CronJobs. Among these, StatefulSet provides mechanisms for state management and persistence, ensuring ordered pod creation and unique identities for pods, along with manual provisioning of PersistentVolumes and headless services for networking.
However, Kubernetes’ existing mechanisms are primarily optimized for stable, high-resource cloud environments. Edge computing scenarios necessitate enhanced state resiliency through efficient replication, synchronization, and robust storage management strategies. This thesis will explore how Kubernetes-like platforms can improve state resiliency and persistent storage management at the Edge, addressing critical issues of resource scarcity, data integrity, replication efficiency, and network reliability.
Expected Outcomes
The thesis aims to propose and evaluate novel strategies for efficient state management and storage replication tailored for dynamic edge infrastructures. Key outcomes and research objectives include:
- An analysis of the limitations of existing Kubernetes StatefulSet and PersistentVolume approaches in edge computing scenarios.
- Development of a lightweight storage replication and synchronization mechanism suitable for resource-constrained edge devices, addressing unreliable connectivity and intermittent power.
- Strategies for placement and management of persistent volumes at the edge to minimize reliance on cloud connectivity, preserving the autonomy and responsiveness of edge applications.
- Robust data integrity and consistency mechanisms, incorporating lightweight verification and recovery procedures to prevent corruption and unauthorized data tampering.
- Implementation and evaluation of replication and synchronization methods via practical experiments on a testbed representing realistic edge conditions.
Requirements
- Strong knowledge of Kubernetes orchestration and containerized environments.
- Proficiency in distributed systems, data replication techniques, and fault tolerance mechanisms.
- Familiarity with edge computing constraints, including network reliability and resource limitations.
- Programming proficiency, particularly in Go or Python, for implementing prototype systems.
- Basic understanding of data encryption and security techniques for secure storage.
