On the complexity of asynchronous gossip

In this paper, we study the complexity of gossip in an asynchronous, message-passing fault-prone distributed system. In short, we show that an adaptive adversary can significantly hamper the spreading of a rumor, while an oblivious adversary cannot. This latter fact implies that there exist message-efficient asynchronous (randomized) consensus protocols, in the context of an oblivious adversary. In more detail, we summarize our results as follows. If the adversary is adaptive, we show that a randomized asynchronous gossip algorithm cannot terminate in fewer than O(f(d + delta)) time steps unless Omega(n+f2) messages are exchanged, where n is the total number of processes, f is the number of tolerated crash failures, d is the maximum communication delay for the specific execution in question, and delta is the bound on relative process speeds in the specific execution. The lower bound result is to be contrasted with deterministic synchronous gossip algorithms that, even against an adaptive adversary, require only O(polylog n) time steps and O(n polylog n) messages. In the case of an oblivious adversary, we present three different randomized, asynchronous algorithms that provide different trade-offs between time complexity and message complexity. The first algorithm is based on the epidemic paradigm, and completes in O(n / (n-f) log2 n (d + δ)) time steps using O(n log3 n (d + δ)) messages, with high probability. The second algorithm relies on more rapid dissemination of the rumors, yielding a constant-time (w.r.t. n) gossip protocol: for every constant epsilon < 1, and for f ≤ n/2, there is a variant with time complexity O((1 / ε)(d+δ)) and message complexity O((1/ε)n1+εlog n (d+δ)). The third algorithm solves a weaker version of the gossip problem in which each process receives at least a majority of the rumors. This algorithm achieves constant O(d+δ) time complexity and message complexity O(n7/4 log2 n). As an application of these message-efficient gossip protocols, we present three randomized consensus protocols. Our consensus algorithms derive from combining each of our gossip protocols with the Canetti-Rabin framework, resulting in message-efficient consensus algorithms. The resulting protocols have time and message-complexity asymptotically equal to our gossip protocols. We particularly highlight the third consensus protocol, a result that is interesting in its own right: the first asynchronous randomized consensus algorithm with strictly subquadradic message-complexity, i.e., O(n7/4 log2 n).

[1]  Wei Tsang Ooi,et al.  Controlling Gossip Protocol Infection Pattern Using Adaptive Fanout , 2005, 25th IEEE International Conference on Distributed Computing Systems (ICDCS'05).

[2]  Richard M. Karp,et al.  Randomized rumor spreading , 2000, Proceedings 41st Annual Symposium on Foundations of Computer Science.

[3]  Dariusz R. Kowalski,et al.  Time and Communication Efficient Consensus for Crash Failures , 2006, DISC.

[4]  Andrzej Pelc,et al.  Dissemination of Information in Communication Networks - Broadcasting, Gossiping, Leader Election, and Fault-Tolerance , 2005, Texts in Theoretical Computer Science. An EATCS Series.

[5]  Hagit Attiya,et al.  Distributed Computing: Fundamentals, Simulations and Advanced Topics , 1998 .

[6]  Ran Canetti,et al.  Fast asynchronous Byzantine agreement with optimal resilience , 1993, STOC.

[7]  Anne-Marie Kermarrec,et al.  Efficient epidemic-style protocols for reliable and scalable multicast , 2002, 21st IEEE Symposium on Reliable Distributed Systems, 2002. Proceedings..

[8]  Jon M. Kleinberg,et al.  Spatial gossip and resource location protocols , 2001, JACM.

[9]  Doug Terry,et al.  Epidemic algorithms for replicated database maintenance , 1988, OPSR.

[10]  Andrzej Pelc Fault-tolerant broadcasting and gossiping in communication networks , 1996, Networks.

[11]  Chryssis Georgiou,et al.  Efficient Gossip and Robust Distributed Computation , 2003, DISC.

[12]  Stephen P. Boyd,et al.  Randomized gossip algorithms , 2006, IEEE Transactions on Information Theory.

[13]  Marcin Paprzycki,et al.  Distributed Computing: Fundamentals, Simulations and Advanced Topics , 2001, Scalable Comput. Pract. Exp..

[14]  Anne-Marie Kermarrec,et al.  Lightweight probabilistic broadcast , 2003, TOCS.

[15]  Nancy A. Lynch,et al.  Consensus in the presence of partial synchrony , 1988, JACM.

[16]  Michael Ben-Or,et al.  Another advantage of free choice (Extended Abstract): Completely asynchronous agreement protocols , 1983, PODC '83.

[17]  Anne-Marie Kermarrec,et al.  Probabilistic Reliable Dissemination in Large-Scale Systems , 2003, IEEE Trans. Parallel Distributed Syst..

[18]  James Aspnes,et al.  Randomized protocols for asynchronous consensus , 2002, Distributed Computing.

[19]  Nancy A. Lynch,et al.  Impossibility of distributed consensus with one faulty process , 1985, JACM.

[20]  M. Mitzenmacher,et al.  Probability and Computing: Chernoff Bounds , 2005 .

[21]  Kenneth P. Birman,et al.  Bimodal multicast , 1999, TOCS.

[22]  Nancy A. Lynch,et al.  Impossibility of distributed consensus with one faulty process , 1983, PODS '83.

[23]  Robbert van Renesse,et al.  A Gossip-Style Failure Detection Service , 2009 .

[24]  Michael Ben-Or,et al.  Another advantage of free choice (Extended Abstract): Completely asynchronous agreement protocols , 1983, PODC '83.

[25]  Dariusz R. Kowalski,et al.  Bounding Work and Communication in Robust Cooperative Computation , 2002, DISC.

[26]  Cynthia Dwork,et al.  Randomization in Byzantine Agreement , 1989, Adv. Comput. Res..

[27]  Patrick Th. Eugster,et al.  Route driven gossip: probabilistic reliable multicast in ad hoc networks , 2003, IEEE INFOCOM 2003. Twenty-second Annual Joint Conference of the IEEE Computer and Communications Societies (IEEE Cat. No.03CH37428).

[28]  Dariusz R. Kowalski,et al.  Robust gossiping with an application to consensus , 2006, J. Comput. Syst. Sci..

[29]  Patrick Th. Eugster,et al.  Probabilistic reliable multicast in ad hoc networks , 2004, Ad Hoc Networks.