An information-interconnectivity-based retrieval method for network attached storage

Authors:
Iliyak Georgiev;Ivo I. Georgiev
Affiliations:
Metro State College of Denver, Denver, CO;San Francisco State University, San Francisco, CA
Venue:
Proceedings of the 1st conference on Computing frontiers
Year:
2004

Citing 10
Cited 0

Automated hoarding for mobile computers

Proceedings of the sixteenth ACM symposium on Operating systems principles
Network attached storage architecture

Communications of the ACM
Data-driven evolution of data mining algorithms

Communications of the ACM - Evolving data mining into solutions for insights
The Evolving Field of Distributed Storage

IEEE Internet Computing
Chameleon: Hierarchical Clustering Using Dynamic Modeling

Computer
Active Disks for Large-Scale Data Processing

Computer
The Case for Efficient File Access Pattern Modeling

HOTOS '99 Proceedings of the The Seventh Workshop on Hot Topics in Operating Systems
A Network-Aware Distributed Storage Cache for Data Intensive Environments

HPDC '99 Proceedings of the 8th IEEE International Symposium on High Performance Distributed Computing
Performance study of iSCSI-based storage subsystems

IEEE Communications Magazine
Object-based storage

IEEE Communications Magazine

Quantified Score

Hi-index	0.00

Visualization

Abstract

Network attached disk storage is characterized by independent network attachment and embedded intelligence. For Internet applications, it provides the key functionality of geographical replication and intelligent retrieval of data objects.The paper describes a latency reducing method based on the relative interconnectivity between data objects. We follow the locality-of-reference principle to partition interrelated data objects on close disk areas or network storage nodes. The method incorporates a clustering algorithm to support smarter placement of related objects and read-ahead group caching. Objects that are associated together are clustered in the same group and can be read from disk and cached together. The proposed clustering and cache algorithms do not use floating point, allowing direct and fast implementation on a variety of disk controllers.