17
9

Communication cost of consensus for nodes with limited memory

Abstract

Motivated by applications in blockchains and sensor networks, we consider a model of nn nodes trying to reach consensus on their majority bit. Each node ii is assigned a bit at time zero, and is a finite automaton with mm bits of memory (i.e., 2m2^m states) and a Poisson clock. When the clock of ii rings, ii can choose to communicate, and is then matched to a uniformly chosen node jj. The nodes jj and ii may update their states based on the state of the other node. Previous work has focused on minimizing the time to consensus and the probability of error, while our goal is minimizing the number of communications. We show that when m>3logloglog(n)m>3 \log\log\log(n), consensus can be reached at linear communication cost, but this is impossible if m<logloglog(n)m<\log\log\log(n). We also study a synchronous variant of the model, where our upper and lower bounds on mm for achieving linear communication cost are 2logloglog(n)2\log\log\log(n) and logloglog(n)\log\log\log(n), respectively. A key step is to distinguish when nodes can become aware of knowing the majority bit and stop communicating. We show that this is impossible if their memory is too low.

View on arXiv
Comments on this paper