Units of Computation in Fault-Tolerant Distributed Systems

Ahuja, M.; Mishra, Shivakant

doi:10.1006/jpdc.1996.1277

Cited by 2 publications

(2 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…2 P as the set of processes in which t he e v ent occurs. For any e 2 E, 2 It is possible to assign equal timestamps to A and B that s a tisfy all four assumptions: For example, any a bstract event gets timestamp 481. However, this timestamp does not characterize weak precedence correctly for abstract events other than A and B.…”

Section: Basic De Nitions and Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Vector time and causality among abstract events in distributed computations

Basten

Kunz

Black

et al. 1997

Distributed Computing

View full text Add to dashboard Cite

An important problem in analyzing distributed computations is the amount of information. In event-based models, even for simple applications, the n umber of events is large and t he c a usal structure is complex. Event a bstraction can be used to r e d uce the a p parent complexity o f a distributed computation.This paper discusses one important aspect of event a bstraction: causality among a bstract events. Following Lamport 24 , two c a usality relations are de ned on abstract events, called weak and strong precedence. A general theoretical framework based on logical vector time is developed in which s e v eral meaningful timestamps for abstract events are derived. These timestamps can be used to e ciently determine c a usal relationships between arbitrary abstract events. The class of convex abstract events i s i d enti ed as a subclass of abstract events t hat i s general enough to be widely applicable and restricted enough to simplify timestamping s c hemes used for characterizing w eak precedence. We explain why s u ch a simpli cation seems not possible for strong precedence.

show abstract

Section: Basic De Nitions and Resultsmentioning

confidence: 99%

“…However, under the a bove assumptions, it is not di cult t o see that a n y reasonable timestamp for A is always smaller than the same timestamp for B. 2 Hence, one n eeds at least two t imestamps to c haracterize weak precedence among a bstract events.…”

Section: Timestamps and Precedence Tests For Abstract Eventsmentioning

confidence: 99%

Vector time and causality among abstract events in distributed computations

Basten

Kunz

Black

et al. 1997

Distributed Computing

View full text Add to dashboard Cite

show abstract

Load balancing with multiple token policy

Kulkarni

Sengupta²

Proceedings Seventh International Conference on Parallel and Distributed Systems (Cat. No.PR00568)

View full text Add to dashboard Cite

In distributed systems uneven arrivals of the tasks may overload few hosts where as some of the hosts are lightly loaded. This load imbalance prevents distributed systems from delivering its performance to its capacity. Load balancing has been advocated as a means of improving performance and reliability of distributed systems. A new load balancing approach [14] is proposed to deal with this problem. In this paper we extend this model with multiple tokens. With some parameters in the algorithm set to intelligent values, the algorithm promises better load balancing results.

show abstract

Units of Computation in Fault-Tolerant Distributed Systems

Cited by 2 publications

References 31 publications

Vector time and causality among abstract events in distributed computations

Vector time and causality among abstract events in distributed computations

Load balancing with multiple token policy

Contact Info

Product

Resources

About