Results 1  10
of
14
Perspectives on the CAP Theorem
"... Almost twelve years ago, in 2000, Eric Brewer introduced the idea that there is a fundamental tradeoff between consistency, availability, and partition tolerance. This tradeoff, which has become known as the CAP Theorem, has been widely discussed ever since. In this paper, we review the CAP Theore ..."
Abstract

Cited by 6 (0 self)
 Add to MetaCart
(Show Context)
Almost twelve years ago, in 2000, Eric Brewer introduced the idea that there is a fundamental tradeoff between consistency, availability, and partition tolerance. This tradeoff, which has become known as the CAP Theorem, has been widely discussed ever since. In this paper, we review the CAP Theorem and situate it within the broader context of distributed computing theory. We then discuss the practical implications of the CAP Theorem, and explore some general techniques for coping with the inherent tradeoffs that it implies. 1
Failure Detectors Encapsulate Fairness
, 2010
"... Failure detectors have long been viewed as abstractions for the synchronism present in distributed system models. However, investigations into the exact amount of synchronism encapsulated by a given failure detector have met with limited success. The reason for this is that traditionally, models of ..."
Abstract

Cited by 4 (1 self)
 Add to MetaCart
Failure detectors have long been viewed as abstractions for the synchronism present in distributed system models. However, investigations into the exact amount of synchronism encapsulated by a given failure detector have met with limited success. The reason for this is that traditionally, models of partial synchrony are specified with respect to real time, but failure detectors do not encapsulate real time. Instead, we argue that failure detectors encapsulate the fairness in computation and communication. Fairness is a measure of the number of steps executed by one process relative either to the number of steps taken by another process or relative to the duration for which a message is in transit. We argue that partially synchronous systems are perhaps better specified with fairness constraints (rather than realtime constraints) on computation and communication. We demonstrate the utility of this approach by specifying the weakest system models to implement failure detectors in the ChandraToueg hierarchy.
Partial synchrony based on set timeliness
, 2009
"... We introduce a new model of partial synchrony for readwrite shared memory systems. This model is based on the notion of set timeliness—a natural and straightforward generalization of the seminal concept of timeliness in the partially synchrony model of Dwork, Lynch and Stockmeyer [11]. Despite its ..."
Abstract

Cited by 4 (0 self)
 Add to MetaCart
We introduce a new model of partial synchrony for readwrite shared memory systems. This model is based on the notion of set timeliness—a natural and straightforward generalization of the seminal concept of timeliness in the partially synchrony model of Dwork, Lynch and Stockmeyer [11]. Despite its simplicity, the concept of set timeliness is powerful enough to define a family of partially synchronous systems that closely match individual instances of the tresilient kset agreement problem among n processes, henceforth denoted (t, k, n)agreement. In particular, we use it to give a partially synchronous system that is is synchronous enough for solving (t, k, n)agreement, but not enough for solving two incrementally stronger problems, namely, (t + 1, k, n)agreement, which has a slightly stronger resiliency requirement, and (t, k − 1, n)agreement, which has a slightly stronger agreement requirement. This is the first partially synchronous system that separates between these subconsensus problems. The above results show that set timeliness can be used to study and compare the partial synchrony requirements of problems that are strictly weaker than consensus.
Algorithms for Extracting Timeliness Graphs
 Lecture Notes in Computer Science
"... We consider asynchronous messagepassing systems in which some links are timely and processes may crash. Each run defines a timeliness graph among correct processes: (p, q) is an edge of the timeliness graph if the link from p to q is timely (that is, there is bound on communication delays from p t ..."
Abstract

Cited by 3 (1 self)
 Add to MetaCart
We consider asynchronous messagepassing systems in which some links are timely and processes may crash. Each run defines a timeliness graph among correct processes: (p, q) is an edge of the timeliness graph if the link from p to q is timely (that is, there is bound on communication delays from p to q). The main goal of this paper is to approximate this timeliness graph by graphs having some properties (such as being trees, rings,...). Given a family S of graphs, for runs such that the timeliness graph contains at least one graph in S then using an extraction algorithm, each correct process has to converge to the same graph in S that is, in a precise sense, an approximation of the timeliness graph of the run. For example, if the timeliness graph contains a ring, then using an extraction algorithm, all correct processes eventually converge to the same ring and in this ring all nodes will be correct processes and all links will be timely. We first present a general extraction algorithm and then a more specific extraction algorithm that is communication efficient (i.e., eventually all the messages of the extraction algorithm use only links of the extracted graph). 1
Specifying and Implementing an Eventual Leader Service for Dynamic Systems
, 2010
"... Abstract: The election of an eventual leader in an asynchronous system prone to process crashes is an important problem of faulttolerant distributed computing. This problem is known as the implementation of the failure detector Ω. Nearly all papers that propose algorithms implementing such an eventu ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
(Show Context)
Abstract: The election of an eventual leader in an asynchronous system prone to process crashes is an important problem of faulttolerant distributed computing. This problem is known as the implementation of the failure detector Ω. Nearly all papers that propose algorithms implementing such an eventual leader service consider a static system. In contrast this paper considers a dynamic system, i.e., a system in which processes can enter and leave. The paper has three contributions. It first proposes a specification of Ω suited to dynamic systems. Then, it presents and proves correct an algorithm implementing this specification. Finally, the paper discusses the notion of an eventual leader suited to dynamic systems. It introduces an additional property related to system stability. The design of an algorithm satisfying this last property remains an open challenging problem.
DOI 10.1007/s0044601001064 Adaptive progress: a gracefullydegrading liveness property
"... Abstract We introduce a simple liveness property for shared object implementations that is gracefully degrading depending on the degree of synchrony in each run. This property, called adaptive progress, provides a gradual bridge between obstructionfreedom and waitfreedom in partiallysynchronous s ..."
Abstract
 Add to MetaCart
(Show Context)
Abstract We introduce a simple liveness property for shared object implementations that is gracefully degrading depending on the degree of synchrony in each run. This property, called adaptive progress, provides a gradual bridge between obstructionfreedom and waitfreedom in partiallysynchronous systems. We show that adaptive progress can be achieved using very weak shared objects. More precisely, every object has an implementation that ensures adaptive progress and uses only abortable registers (which are weaker than safe registers). As part of this work, we present a new leader election abstraction that processes can use to dynamically compete for leadership such that if there is at least one timely process among the current candidates for leadership, then a timely leader is eventually elected among the candidates. We also show that this abstraction can be implemented using abortable registers. 1
Partial Synchrony Based on Set Timeliness [Extended Abstract]
"... We introduce a new model of partial synchrony for readwrite shared memory systems. This model is based on the notion of set timeliness—a natural and straightforward generalization of the seminal concept of timeliness in the partially synchrony model of Dwork, Lynch and Stockmeyer [8]. Despite its s ..."
Abstract
 Add to MetaCart
(Show Context)
We introduce a new model of partial synchrony for readwrite shared memory systems. This model is based on the notion of set timeliness—a natural and straightforward generalization of the seminal concept of timeliness in the partially synchrony model of Dwork, Lynch and Stockmeyer [8]. Despite its simplicity, the concept of set timeliness is powerful enough to describe the first partially synchronous system for read/write shared memory that separates consensus and set agreement: we show that this system has enough timeliness for solving set agreement but not enough for solving consensus. Set timeliness also allows us to define a family of partially synchronous systems of n processes, denoted S k n (1≤k≤n−1), which closely matches the family of kantiΩ failure detectors that were recently shown to be the weakest failure detectors for the kset agreement problem: We prove that for 1≤k≤n−1, S k n is synchronous enough to implement kantiΩ but not enough to implement (k−1)antiΩ. The results above show that set timeliness can be used to study and compare the partial synchrony requirements of problems that are strictly weaker than consensus.
unknown title
"... Abstract. We define quittable consensus, a natural variation of the consensus problem, where processes have the option to agree on “quit ” if failures occur, and we relate this problem to the wellknown problem of nonblocking atomic commit. We then determine the weakest failure detectors for these t ..."
Abstract
 Add to MetaCart
Abstract. We define quittable consensus, a natural variation of the consensus problem, where processes have the option to agree on “quit ” if failures occur, and we relate this problem to the wellknown problem of nonblocking atomic commit. We then determine the weakest failure detectors for these two problems in all environments, regardless of the number of faulty processes.
Stabilizing Leader Election in Partial Synchronous Systems with Crash Failures?
"... This article deals with stabilization and faulttolerance. We consider two types of stabilization: the self and the pseudo stabilization. Our goal is to implement the self and/or pseudo stabilizing leader election in systems with process crashes, weak reliability, and synchrony assumptions. We t ..."
Abstract
 Add to MetaCart
(Show Context)
This article deals with stabilization and faulttolerance. We consider two types of stabilization: the self and the pseudo stabilization. Our goal is to implement the self and/or pseudo stabilizing leader election in systems with process crashes, weak reliability, and synchrony assumptions. We try to propose, when it is possible, communicationefficient implementations. Our approach allows to obtain algorithms that tolerate both transient and crash failures. Note that some of our solutions are adapted from existing faulttolerant algorithms. The motivation here is not to propose new algorithms but merely to show some assumptions required to obtain stabilizing leader elections in systems with crash failures. In particular, we focus on the borderline assumptions where we go from the possibility to have selfstabilizing solutions to the possibility to only have pseudostabilizing ones. Key words: Faulttolerance, selfstabilization, pseudostabilization, leader election. 1