Early Work on Efficient Patching for Coordinating Edge Applications

TL;DR

This paper investigates efficient methods for updating multiple edge applications with prioritized updates to optimize performance and bandwidth, proposing a scalable system called DSOC for orchestrating these updates.

Contribution

It introduces a setup and benchmarks for understanding update impacts and proposes the DSOC system for reliable, distributed edge application orchestration with prioritized updates.

Findings

Prioritized updates improve overall system performance.

Benchmark results demonstrate reduced bandwidth usage.

DSOC enables scalable and reliable edge application management.

Abstract

Multiple applications running on Edge computers can be orchestrated to achieve the desired goal. Orchestration of applications is prominent when working with Internet of Things based applications, Autonomous driving and Autonomous Aerial vehicles. As the applications receive modified classifiers/code, there will be multiple applications that need to be updated. If all the classifiers are synchronously updated there would be increased throughput and bandwidth degradation. On the other hand, delaying updates of applications which need immediate update hinders performance and delays progress towards end goal. The updates of applications should be prioritized and updates should happen according to this priority. This paper explores the setup and benchmarks to understand the impact of updates when multiple applications working to achieve same objective are orchestrated with prioritized updates. We discuss methods to build a distributed, reliable and scalable system called "DSOC"(Docker Swarm Orchestration Component).