Improved Self Fused Check pointing Replication for Handling Multiple Faults in Cloud Computing
Journal Title: International Journal on Computer Science and Engineering - Year 2012, Vol 4, Issue 6
Abstract
The performance of checkpointing replication fault tolerance technique is severely bottlenecks due to handling of number of replicas generated for a large number of nodes to tolerate multiple faults such as multiple failure of nodes, processes etc. In fusion based approach, these checkpointing replicas stored at large number of computing nodes is aggregated into some data structure to handle efficiently through fused data structure. These impose higher overheads of fusing a large numbers of checkpointing replicas. In this paper, a self fused checkpointing replication (SFCR) for cloud computing is proposed. All checkpointing replicas assigned to store at a particular node are stored in a self-fused-shared-checkpointing-replicas file already created and located at every node rather than storing as a separate checkpointing element and than fusing. Thus, it eliminates the need of further fusing of the checkpointing replicas stored at different checkpointing replicas storage nodes, as checkpointing replicas assign to store a particular node stored in an already created fuse file at every checkpointing replicas storage node. It improves the performance without affecting the specified fault tolerant capabilities as failure of any node will result in loss of all replicas irrespective of separate checkpointing file or shared checkpointing fused file. Costs to maintain these set of self fused shared files are obviously less than the number of separate replicated files in terms of time and efforts that it takes to create, and update a file. Thus, proposed approach is enhancement of performance without compromising the specified fault tolerance capability. At the same time when system seems to be prone to many number of faults, some specific self fused shared files consist of important and critical data that can be further replicated to enhance the fault tolerant capability of a group of important and critical nodes or processes at run time. Thus, it also provides adaptive dynamic fault tolerance capability in a simple manner. Effectiveness of proposed technique is also presented
Authors and Affiliations
Sanjay Bansal , Sanjeev Sharma , Ishita Trivedi , Mrinalika Ghosh
Keywords
Related Articles
- EP ID EP156018
- DOI -
- Views 101
- Downloads 0
Speaker Identification using Row Mean of DCT and Walsh Hadamard Transform
In this paper we propose a unique approach to text dependent speaker identification using transformation techniques such as DCT (Discrete Cosine Transform) and WHT (Walsh and Hadamard Transform). The feature vectors for...
Analysis of Selective Routing Strategies for Fault Tolerance In Wireless Sensor Networks
Wireless sensor networks are intended to have large number of sensor nodes which are widely deployed in a distributed environment. Steep increase in number of nodes, which is a deviant to WSN eventually, affects the comm...
Human Talent Prediction in HRM using C4.5 Classification Algorithm
In HRM, among the challenges for HR professionals is to manage an organization’s talents, especially to ensure the right person for the right job at the right time. Human talent prediction is an alternative to handle thi...
Human Age Prediction and Classification Using Facial Image
Human face is one of the most important sources of the information, which can be utilized for personal verification and identification. This paper discusses the method of finding the human age using the facial image of a...
A Survey on Early Determination of Zero Quantized Coefficients in Video Coding
In video encoding, there are a large number of discrete cosine transform (DCT) coefficients of the prediction residue which are quantized to zeros. Therefore, it is desired to design a method which can early detect zero...