A Distributed Hash Table-based Approach for Providing a File System Neutral, Robust and Resilient Storage Infrastructure

Parissis, George; Apostolopoulos, Theodore K.

doi:10.1109/ngi.2009.5175751

Cited by 4 publications

(5 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

Section: Introductionmentioning

confidence: 99%

“…Cluster file systems such as Lustre, Gluster, relies on the powerful, dedicated metadata servers for conducting all file metadata operations, leading to a hierarchical structure that scales well only in combination with high-speed, low-latency and symmetric network links [3]. One distributed network file system like Wofs that splits a file into many small objects, stores these objects in remote file servers, and uses a special B+ tree to manage the metadata of these objects.…”

Section: Introductionmentioning

confidence: 99%

“…Besides, the system can promise the complexity due to frequent insertion and deletion not being higher than original B+ tree. [1], [3], [4] III. BACKGROUND THEORY…”

Section: Introductionmentioning

confidence: 99%

“…Recovering the system state depends on the weight of the underlying nodes. The higher the applied storage tree level reaches, the longer the time to be completed [1].Cluster file systems such as Lustre, Gluster, relies on the powerful, dedicated metadata servers for conducting all file metadata operations, leading to a hierarchical structure that scales well only in combination with high-speed, low-latency and symmetric network links [3].One distributed network file system like Wofs that splits a file into many small objects, stores these objects in remote file servers, and uses a special B+ tree to manage the metadata of these objects. Besides, it uses the object-range locking policy to avoid data incoherence and improve performance [4].…”

mentioning

confidence: 99%

See 3 more Smart Citations

A Framework for Power of 2 Based Scalable Data Storage in Object-Based File System

Aung¹,

Thein²

2013

IJFCC

View full text Add to dashboard Cite

Abstract-As the amount of data in the today's storage systems has been growing over times, the ideas to expand the new resources are being considered to meet the system's requirement. Adding or removing resources will make throughput of the current state changes. To lower cost or make efficiency in maintaining the system's balance, one trial by another that serve nothing before is being replaced. B+ tree-based indexing algorithm is widely being popular for managing data dynamically in today's storage systems. Fast data insertion, deletion and searching are also concerned with the system's performance. Those criteria are heavily dependent on the order size and height of the tree used because it determines how large a B+ Tree will become and how much the system can hold data and throughput. The proposed system modifies the traditional B+ Tree in the form power of 2-based for data expansion and it is designed on object-based file system. Index Terms-B+ tree, power of 2-based clustering I. INTRODUCTIONIn modern storage systems, changes have occurred over time as new requirements are added on demand to supply needed capacity or bandwidth. Those file systems have been making storage management simplified and try to hide most of the complexities behind in order to get the systems performance high. More importantly, distributed file system workloads are inherently dynamic, with significant variation in data and metadata access as active applications and data sets change over time. Consequently, the performance of the storage system depends on the way objects included in the file system's attributes which can be managed by using appropriate method [1].But, the higher the desired level to get, the more storage overheads will be cost and the longer the time to maintain data consistency. Even RAID-based storage systems get longer time to rebuild as increases in disk capacity outpace increases in bandwidth. These may be challenging to ensure high reliability and scalability for large-scale storage systems [2].The B+ tree is easy to implement for being balanced structure in which all path from root to the nodes are same. Besides, the algorithm "B+ Tree" mostly used by the file system usually has been grown by splitting nodes (internal or leaf) in the tree and such expansion causes the tree level to be high. This results in frequent I/O operations to be taken much longer time throughout the whole process. Being raised more levels, system scales logarithmically and then storage's Manuscript received February 1, 2013; revised March 12, 2013. The authors are with the University of Computer Studies, Yangon, Republic of Union of Myanmar (e-mail: ohnmaraung2008@gmail.com, nilarthein@gmail.com) useful capacity grows more slowly as it gets larger. Such natures are now being facing in those systems deploying tree-based approach for data storage, which becomes a major problem to be solved. This has resulted from increasing B+ Tree size associated with its order. Until now, there is no system based on B+ tree indexing structure with limi...

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

“…Besides, the system can promise the complexity due to frequent insertion and deletion not being higher than original B+ tree. [1], [3], [4] III. BACKGROUND THEORY…”

Section: Introductionmentioning

confidence: 99%

mentioning

confidence: 99%

See 2 more Smart Citations

A Framework for Power of 2 Based Scalable Data Storage in Object-Based File System

Aung¹,

Thein²

2013

IJFCC

View full text Add to dashboard Cite

show abstract

“…• A thorough analysis of the design requirements that a distributed storage system should satisfy so that it can provide an efficient solution when deployed in high performance clusters and large scale networks [61]. • The design and the open source implementation of such a system [62], [63].…”

Section: • Cooperative Networkmentioning

confidence: 99%

Block-level storage in structured overlay networks

Parisis¹,

Παρίσης²

View full text Add to dashboard Cite

BLAST: Off-the-Shelf Hardware for Building an Efficient Hash-Based Cluster Storage System

Parissis

Xylomenos

Gritzalis

2009

2009 Sixth IFIP International Conference on Network and Parallel Computing

View full text Add to dashboard Cite

A Distributed Hash Table-based Approach for Providing a File System Neutral, Robust and Resilient Storage Infrastructure

Cited by 4 publications

References 19 publications

A Framework for Power of 2 Based Scalable Data Storage in Object-Based File System

A Framework for Power of 2 Based Scalable Data Storage in Object-Based File System

Block-level storage in structured overlay networks

BLAST: Off-the-Shelf Hardware for Building an Efficient Hash-Based Cluster Storage System

Contact Info

Product

Resources

About