52 research outputs found
Simplifying Fault-Tolerance: Providing the Abstraction of Crash Failures
No full textion of Crash Failures Rida A. Bazzi y Gil Neiger GIT--CC--93/12 February 16, 1993 Abstract The difficulty of designing fault-tolerant distributed algorithms increases with the severity of failures that an algorithm must tolerate. This paper considers methods that automatically translate algorithms tolerant of simple crash failures into ones tolerant of more severe failures. These translations simplify the design task by allowing algorithm designers to assume that processors fail only by stopping. Such translations can be quantified by two measures: fault-tolerance, which is a measure of how many processors must remain nonfaulty for the translation to be correct, and round-complexity, which is a measure of how the translation increases the running time of an algorithm. Understanding these translations and their limitations with respect to these measures can provide insight into the relative impact of different models of faulty behavior on the ability to provide faulttolerant appl..
Osservazioni sull'onomastica in alcune opere teatrali di Frisch, Fassbinder, Wiesel e Mitterer
No full textSeveral Jewish figures that appear in four contemporary theatre plays are interesting for diverse reasons and under different aspects. This onomastic investigation focuses on the works of the German Rainer Werner Fassbinder’s Der Müll, die Stadt und der Tod, the Swiss Max Frisch’s Andorra, the American Elie Wiesel’s Le procès de Shamgorod and the Austrian Felix Mitterer’s Kein schöner Land. The names of the – de facto or supposed – Jewish protagonists will be taken under examination, as well as other relevant anthroponyms and toponyms. Among some bizarre given and family names present in his play, Fassbinder doesn’t give a name to his Jewish character: his choice led to a huge controversy. Frisch’s Andri is not Jewish by birth, but those around him are convinced of that, and this aspect is reflected in his first name; not only Andorra, also the toponyms in Wiesel’s play deserve an examination. Finally, among the functions fulfilled by various names in Kein schöner Land, one choice the author makes draws our attention
To My Parents Acknowledgments
No full textI would like to begin by expressing my gratitude to my advisor, Mustaque Ahamad. It was his guidance and encouragement which has made this thesis possible. I would also like to thank the members of my dissertation commit-tee, Gil Neiger, Umakishore Ramachandran, Karsten Schwan and Divyakant Agrawal, for their time and helpful suggestions which have vastly improved this thesis. Several people associated with the Clouds project have contributed with their assistance, ideas and coding: In particular, Sathis Menon who helped me find my bearings in the Clouds project and was always around when I needed help; R. Ananthanarayanan and Ajay Mohindra for the discussions on DSM related issues which to a great extent simplified my implementations. I would like to thank my cubicle-mates Vibby Gottemukkala, M. Chelliah and Srinivas Doddapaneni for their companionship during the last five years. Also, Rida Bazzi for the late night tea sessions, while we were both writing our theses, where we argued about everything from the format of the thesis to lif
Detection of Global State Predicates
No full textthose of the authors and should not be construed as an official Department of Defense position, policy, or decision. Author's address: Cornell Universit
Distributed Consensus Revisited
No full textDistributed Consensus is a classical problem in distributed computing. It requires the correct processors in a distributed system to agree on a common value despite the failure of other processors. This problem is closely related to other problems, such as Byzantine Generals, Approximate Agreement, and k-Set Agreement. This paper examines a variant of Distributed Consensus that considers agreement on a value that is more than a single bit and requires that the agreed upon value be one of the correct processors' input values. It shows that, for this problem to be solved in a system with arbitrary failures, it is necessary that more processors remain correct than for solutions to Distributed Consensus and for cases where agreement is only a single bit. Specifically, the number of processors that must be correct is a function of the size of the domain of values used. Two existing consensus algorithms are modified to solve this stronger variant. College of Computing Georgia Institute of Te..
The Complexity of Almost-Optimal Coordination
No full textThe problem of fault-tolerant coordination is fundamental in distributed computing. In the past, researchers have considered the complexity of achieving optimal simultaneous coordination under various failure assumptions. This paper studies the complexity of achieving simultaneous coordination in synchronous systems in the presence of send/receive omission failures. It had been shown earlier that achieving optimal simultaneous coordination in these systems requires NP-hard local computation. In this paper, we study almost-optimal coordination, which requires processors to coordinate within a constant additive or multiplicative number of rounds of the coordination time of an optimal protocol. We show that achieving almost-optimal coordination also requires NP-hard computation. College of Computing Georgia Institute of Technology Atlanta, Georgia 30332-0280 This work was supported in part by the National Science Foundation under grants CCR-9106627 and CCR-9301454. y This author was s..
- …
