Demos: Robust Orchestration for Autonomous Networking


Loading...

Date

2023

Publication Type

Conference Paper

ETH Bibliography

yes

Citations

Altmetric

Data

Abstract

Research in wireless sensor networks has resulted in a remarkable breadth of highly capable systems. However, while specialized protocols perform well in the setting they were designed for, they often lack the ability to quickly adapt once operating conditions change drastically. Of particular importance is resilience to node and link failures, as clusters of nodes that lost their leader or split apart need to re-organize and find each other again. With Demos, we present a low-power wireless protocol that ensures robust network orchestration despite such failures. Demos rapidly finds consensus on leadership with its cluster coordination mechanism even if the set of nodes fluctuates by introducing new election quorums. In addition, a novel cluster discovery scheme enables autonomous clusters to merge on the fly and maximize network coverage. Experiments with controlled mobility on a multi-hop network of 24 nodes demonstrate that Demos maintains a reliable data exchange despite severe disruptions and adapts to changes within seconds. We further find that Demos' ability to continuously coordinate and discover achieves highly robust orchestration of fully autonomous clusters.

Publication status

published

Book title

EWSN '23: Proceedings of the 2023 International Conference on Embedded Wireless Systems and Networks

Journal / series

Volume

Pages / Article No.

219 - 230

Publisher

Association for Computing Machinery

Event

International Conference on Embedded Wireless Systems and Networks (EWSN 2023)

Edition / version

Methods

Software

Geographic location

Date collected

Date created

Subject

Fault tolerance; Consensus; Network orchestration; Concurrent transmissions; Wireless sensor networks

Organisational unit

03429 - Thiele, Lothar (emeritus) / Thiele, Lothar (emeritus) check_circle

Notes

Conference lecture held on September 27, 2023.

Funding

180545 - NCCR Automation (phase I) (SNF)

Related publications and datasets

Is documented by:
Is supplemented by: 10.5905/ethz-1007-675