Distributed Algorithms for a Replication Problem of Popular Network Data

The replication of popular data objects can effectively reduce the access time and bandwidth requirements of network services. We study the replication problem in the model of distributed replication groups and propose two distributed algorithms: an approximation optimal replication algorithm, which...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Journal of network and systems management Ročník 24; číslo 1; s. 34 - 56
Hlavní autoři: Jiang, Xiaofeng, Li, Jun, Xi, Hongsheng
Médium: Journal Article
Jazyk:angličtina
Vydáno: New York Springer US 01.01.2016
Springer Nature B.V
Témata:
Algorithms
Approximation
Asynchronous
Bandwidths
Communications Engineering
Complexity
Computer Communication Networks
Computer networks
Computer Science
Computer Systems Organization and Communication Networks
Data replication
Experiments
Information Systems and Communication Service
Isolationism
Mathematical analysis
Networks
Operations Research/Decision Theory
Optimization
Replication
Reproducibility
Requirements
Servers
Simulation
Storage
Storage capacity
Replication of popular data
Access cost
Asynchronous distributed algorithm
Distributed replication group
ISSN:1064-7570, 1573-7705
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:The replication of popular data objects can effectively reduce the access time and bandwidth requirements of network services. We study the replication problem in the model of distributed replication groups and propose two distributed algorithms: an approximation optimal replication algorithm, which is an asynchronous distributed algorithm as it takes more time to be completed. However its performance approaches the optimal algorithm, and a fast replication algorithm that is very suitable as the initial algorithm of the approximation optimal algorithm. We give a proof of the complexity of the algorithms, and show that the time and communication complexities of the algorithms are polynomial with respect to the number of objects and the maximum storage capacities of the servers. Finally, simulation experiments are performed to investigate the performance of the algorithms, and the results show that the two algorithms can effectively solve the replication problem.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
content type line 23
ISSN:1064-7570
1573-7705
DOI:10.1007/s10922-014-9338-0