Faster Information Gathering in Ad-Hoc Radio Tree Networks

TL;DR

This paper presents improved deterministic algorithms for efficiently gathering rumors in ad-hoc tree networks without prior topology knowledge, achieving near-optimal time bounds with and without acknowledgment mechanisms.

Contribution

It introduces two deterministic algorithms that reduce the gathering time to O(n log log n) and O(n), improving upon previous bounds in ad-hoc radio tree networks.

Findings

O(n log log n) time algorithm without acknowledgements

O(n) time algorithm with acknowledgements

Enhanced efficiency in rumor gathering in ad-hoc tree networks

Abstract

We study information gathering in ad-hoc radio networks. Initially, each node of the network has a piece of information called a rumor, and the overall objective is to gather all these rumors in the designated target node. The ad-hoc property refers to the fact that the topology of the network is unknown when the computation starts. Aggregation of rumors is not allowed, which means that each node may transmit at most one rumor in one step. We focus on networks with tree topologies, that is we assume that the network is a tree with all edges directed towards the root, but, being ad-hoc, its actual topology is not known. We provide two deterministic algorithms for this problem. For the model that does not assume any collision detection nor acknowledgement mechanisms, we give an $O(n\log\log n)$-time algorithm, improving the previous upper bound of $O(n\log n)$. We also show that this running time can be further reduced to $O(n)$ if the model allows for acknowledgements of successful transmissions.