DHHT-RAID: A Distributed Heterogeneous Scalable Architecture for Dynamic Storage Environments | IEEE Conference Publication | IEEE Xplore

DHHT-RAID: A Distributed Heterogeneous Scalable Architecture for Dynamic Storage Environments


Abstract:

In this work we present storage allocation strategies for huge Storage Area Networks. The major approach, the new scalable DHHT-RAID strategy, it is based on Distributed ...Show More

Abstract:

In this work we present storage allocation strategies for huge Storage Area Networks. The major approach, the new scalable DHHT-RAID strategy, it is based on Distributed Heterogeneous Hash Tables, where the storage environment consists of a dynamic set of heterogeneous devices, serving a dynamic set of data. Additionally each device that joins or is removed from the system, generates only minimal impact on the availability of data even if any of known RAID levels is applied. Furthermore such dynamics will cause a minimal impact on the access bandwidth to preserve the specific RAID attributes and a balanced data distribution. Additionally, beneath such dynamics this strategy features all RAID levels at any time, coexistent, with embedded autonomous rebalancing, and without massive time consuming re-stripes neither administrative intensive offset tables nor manual reallocations, and it allocates the approximately maximal capacity, only be restrained by the applied RAID Levels.
Date of Conference: 21-23 May 2007
Date Added to IEEE Xplore: 04 June 2007
Electronic ISBN:978-1-5090-8717-4

ISSN Information:

Conference Location: Niagara Falls, ON, Canada

1 Introduction

Huge storage systems scalable during runtime, demanded by 24[h].7[d] applications, with low administrational costs are of tremendous importance. The sustained increasing consumption and generation of data, especially multimedia documents, emphasizes the importance of scalable, highly available, and fast storage systems. New distributed placement schemes for P2P networks, like [3], [11] are in spite of join and leave operations able to preserve a balanced workload spread with less overhead, in an environment consisting of heterogeneous nodes. Such operations are similar in Storage Area Networks aka. SANs, but they occur less frequently. There, the ongoing growth requires to add or replace devices, whether the device replacements might have different attributes than the original ones. These operations imply massive and thus expansive data movements. So it is eligible that the costs (moved data), to obtain a balanced allocation in the system, after these changes, should be minimal.

Contact IEEE to Subscribe

References

References is not available for this document.