## On Diffusing Updates in a Byzantine Environment (1999)

### Download From

IEEE### Download Links

- [www.cs.huji.ac.il]
- [arxiv.org]
- [www.ece.cmu.edu]
- [www.cs.unc.edu]
- [users.ece.cmu.edu]
- DBLP

### Other Repositories/Bibliography

Venue: | In Proceedings of the 18th IEEE Symposium on Reliable Distributed Systems |

Citations: | 28 - 5 self |

### BibTeX

@INPROCEEDINGS{Malkhi99ondiffusing,

author = {Dahlia Malkhi and Yishay Mansour and Michael K. Reiter},

title = {On Diffusing Updates in a Byzantine Environment},

booktitle = {In Proceedings of the 18th IEEE Symposium on Reliable Distributed Systems},

year = {1999},

pages = {134--143},

publisher = {IEEE Computer Society}

}

### Years of Citing Articles

### OpenURL

### Abstract

We study how to efficiently diffuse updates to a large distributed system of data replicas, some of which may exhibit arbitrary (Byzantine) failures. We assume that strictly fewer than t replicas fail, and that each update is initially received by at least t correct replicas. The goal is to diffuse each update to all correct replicas while ensuring that correct replicas accept no updates generated spuriously by faulty replicas. To achieve reliable diffusion, each correct replica accepts an update only after receiving it from at least t others. We provide the first analysis of epidemic-style protocols for such environments. This analysis is fundamentally different from known analyses for the benign case due to our treatment of fully Byzantine failures---which, among other things, precludes the use of digital signatures for authenticating forwarded updates. We propose two epidemic-style diffusion algorithms and two measures that characterize the efficiency of diffusion algorithms in general. We characterize both of our algorithms according to these measures, and also prove lower bounds with regards to these measures that show that our algorithms are close to optimal.

### Citations

2054 | Probability theory and its applications (Vol - Feller - 1950 |

1874 | Randomized Algorithms - Motwani, Raghavan - 1995 |

1496 | Probability inequalities for sums of bounded random variables - Hoeffding - 1963 |

1244 | The byzantine generals problem - Lamport, Shostak, et al. - 1982 |

704 | Epidemic algorithms for replicated database maintenance - Demers, Greene, et al. - 1987 |

605 |
Host extensions for IP multicasting
- Deering
- 1989
(Show Context)
Citation Context ...called antientropysand apparently initially proposed in [BLNS82], was adopted in Xerox's Clearinghouse project (see [DGH+87]) and the Ensemble system [BHO+98]. Similar ideas also underly IP-Multicast =-=[Dee89]-=- and MUSE (for USENET News propagation) [LOM94]. This anti-entropy technique forms the basis for one of the algorithms (Random) that we study here. As described previously, however, the analysis provi... |

596 |
An Introduction to Computational Learning Theory
- Kearns, Vazirani
- 1994
(Show Context)
Citation Context ...= 1], and X m = P n i=1 X m i . Then, P rob[U m E[U m ] s] P rob[X m E[X m ] s] : From now on we will prove the bounds for X m and they will apply also to U m . First, using a Chernoff bound (see [KV94]) we have that, P rob X m + () 1 2 E[X m + () ] e E[X m + () ] 8 : For m + () = 2nRs;t (24=n) 1=2R;t , we have E[X m + () ] s, and hence P rob[X m + () s] es=4 = q + () : For the ... |

529 | Weighted Voting for Replicated Data - Gifford - 1979 |

403 | Byzantine quorum systems
- Malkhi, Reiter
- 1998
(Show Context)
Citation Context ...system called Phalanx [MR98b].) Fleet replicates data so that it will survive even the malicious corruption of some data replicas, and does so using adaptations of quorum systems to such environments =-=[MR98a]-=-. A characteristic of these replication techniques that is important for this discussion is that each update is sent to only a relatively small subset (quorum) of servers, but one that is guaranteed t... |

218 | Atomic Broadcast: From Simple Message Diusion to Byzantine Aggreement - Cristian, Aghili, et al. - 1985 |

192 | Bimodal multicast - Birman, Hayden, et al. - 1999 |

169 | Secure agreement protocols: Reliable and atomic group multicast in Rampart - Reiter - 1994 |

148 | Asynchronous consensus and broadcast protocols - Bracha, Toueg - 1985 |

141 | Authenticated algorithms for byzantine agreement - Dolev, Strong - 1983 |

124 |
On spreading a rumor
- Pittel
- 1987
(Show Context)
Citation Context ...lures. By comparison, in systems admitting only benign failures there are known algorithms for diffusing updates with O(log n) delay, including one on which the Random algorithm studied here is based =-=[Pit87]-=-. 8 Simulation Results Figure 2 depicts simulation results of the Random and Tree-Random algorithms. The figure portrays the delay of the two methods for varying system sizes (on a logarithmic scale),... |

117 | The SecureRing protocols for securing group communication - Kihlstrom, Moser, et al. - 1998 |

90 | Secure and scalable replication in Phalanx
- Malkhi, Reiter
- 1998
(Show Context)
Citation Context ... application of our work on message diffusion is a data replication system called Fleet. (Fleet is not yet documented, but is based on similar design principles as a predecessor system called Phalanx =-=[MR98b]-=-.) Fleet replicates data so that it will survive even the malicious corruption of some data replicas, and does so using adaptations of quorum systems to such environments [MR98a]. A characteristic of ... |

82 |
The Clearinghouse: A Decentralized Agent for Locating Named Objects in a Distributed Environment
- Oppen, Dalal
(Show Context)
Citation Context ...on is useful for driving replicated data toward a consistent state over time, and has found application for this purpose, e.g., in USENET News [LOM94], and in the Grapevine [BLNS82] and Clearinghouse =-=[OD81-=-] systems. The quality of a diffusion protocol is typically defined by the delay until the update has This paper appears in Proceedings of the 18th IEEE Symposium on Reliable Distributed Systems (SRD... |

46 | A high-throughput secure reliable multicast protocol - Malkhi, Reiter - 1997 |

41 |
Drinking from the firehose: multicast USENET news
- Lidl, Osborne, et al.
- 1994
(Show Context)
Citation Context ...d system is propagated to the rest of the system. Diffusion is useful for driving replicated data toward a consistent state over time, and has found application for this purpose, e.g., in USENET News =-=[LOM94-=-], and in the Grapevine [BLNS82] and Clearinghouse [OD81] systems. The quality of a diffusion protocol is typically defined by the delay until the update has This paper appears in Proceedings of the ... |

28 |
A survey of information authentication
- Simmons
- 1992
(Show Context)
Citation Context ...when the sender of a message is not reliably provided by the network, the sender can be authenticated using techniques that require no cryptographic assumptions (for a survey of these techniques, see =-=[Sim92]-=-). Employing digital signatures, on the other hand, would require assumptions limiting the computational power of faulty replicas. Third, pairwise authentication typically incurs a low computation ove... |

13 |
an exercise in distributed computing
- Grapevine
- 1982
(Show Context)
Citation Context ...rest of the system. Diffusion is useful for driving replicated data toward a consistent state over time, and has found application for this purpose, e.g., in USENET News [LOM94], and in the Grapevine =-=[BLNS82-=-] and Clearinghouse [OD81] systems. The quality of a diffusion protocol is typically defined by the delay until the update has This paper appears in Proceedings of the 18th IEEE Symposium on Reliable... |

11 | Total ordering algorithms for asynchronous byzantine systems - Moser, Melliar-Smith - 1995 |

8 | Probabilistic Byzantine Quorum Systems - Malkhi, Reiter, et al. - 2008 |