Efficient Multichannel Rendezvous Algorithms without Global Channel Enumeration

Abstract

The multichannel rendezvous problem (MRP) is a critical challenge for neighbor discovery in IoT applications, requiring two users to find each other by hopping among available channels over time. This paper addresses the MRP in scenarios where a global channel enumeration system is unavailable. To tackle this challenge, we propose a suite of low-complexity multichannel rendezvous algorithms based on locality-sensitive hashing (LSH), tailored for environments where channel labels are unique L-bit identifiers rather than globally coordinated indices. Inspired by consistent hashing techniques in distributed systems, we develop the LC-LSH and LC-LSH4 algorithms for synchronous and asynchronous settings, respectively. These algorithms significantly reduce implementation complexity while maintaining expected time-to-rendezvous (ETTR) performance comparable to state-of-the-art methods that require global channel enumeration. To ensure bounded maximum time-to-rendezvous (MTTR) in the asynchronous setting, we further introduce the ASYM-LC-LSH4 and QR-LC-LSH4 algorithms by embedding multiset-enhanced modular clock and quasi-random techniques into our framework. Extensive simulations demonstrate that the proposed algorithms achieve performance comparable to state-of-the-art LSH algorithms in both synchronous and asynchronous settings, even without a global channel enumeration system.