DeAI Research
Source · IEEE International Conference on Edge Computing and Communications (EDGE)

CargoSync: Efficient Containerd Image Updates in Low-Bandwidth Edge Environments with Rsync-based Delta Compression

2024Ioannis Protogeros; Michalis Karamousadakis; Michail Rizakis; Antonios Porichis

cid ad624a75241f8df8c30e758b38f37298630f6378dc5334744ecedb5b7ea962f9 · original · open access · via openalex

Title: CargoSync: Efficient Containerd Image Updates in Low-Bandwidth Edge Environments with Rsync-based Delta Compression
Authors: Ioannis Protogeros; Michalis Karamousadakis; Michail Rizakis; Antonios Porichis
Venue: IEEE International Conference on Edge Computing and Communications (EDGE)
Year: 2024
Canonical: doi:10.1109/edge62653.2024.00013
URL: https://doi.org/10.1109/edge62653.2024.00013
Collector: openalex
OA: true

ABSTRACT
Containers have emerged as the standard for software deployment, offering simplicity, low overhead, and portability. In the context of updating containers in Edge environments, the constrained bandwidth poses the critical challenge of minimizing data transfer and optimizing the update process. This paper introduces CargoSync, a Containerd-based tool designed for efficient image updates. It leverages the rsync tool for delta compression, reducing the size of updates towards saving network resources. The proposed method is evaluated against traditional pull-based updates showcasing notable improvements in data transfer efficiency as well as update times even for large container images when bandwidth is severely limited. Our tool competes favourably with Moby-based counterpart Balena-engine’s rsync delta updates in our performance evaluations, making it the first Containerd-based tool to effectively adopt this approach. Additionally, CargoSync addresses the challenge of distributing updates to multiple machines concurrently, demonstrating scalability benefits. The results indicate the potential of CargoSync for optimizing container image updates in resource-constrained environments.