Design simulation and performance analysis of information dispersal algorithm based cloud storage system

Abstract

In this work, a system named IDA-based Multi-Cloud Storage System (I-MCSS) is proposed that attempts to maintain the tradeoff between QoS parameters and the cost paid by cloud user with a secure dispersion of information at various locations. The dispersal cost in I-MCSS is optimized while fulfilling the cloud users preferred set of QoS requirements. For the same, the notion of knapsack as well as information dispersal algorithm is deployed collectively. newlineI-MCSS enables the optimal selection of CSPs as per QoS requirement of the cloud user and stores the data over the multi-cloud environment to provide implicit security. I-MCSS enables cloud brokers to: newline capture user QoS requirements and determine the score of the CSPs newline select optimal subset of CSPs for storage newline split and disperse the data over multiple CSPs so that the data is stored as per QoS requirements and implicit security for the sensitive data is maintained newlineAHP method has been used to assign relative importance to QoS parameters from which relative weights are computed for each QoS parameter. Weights generated from the AHP process are captured and processed by a MCDM method called TOPSIS. TOPSIS generates the scores for different CSPs registered with the cloud broker which are further used for optimal selection of CSPs. Knapsack algorithm has been used to define best five security level plans available under the defined budget. For different settings of experiments, it was observed that with the increase in security level, more CSPs were selected and hence higher implicit security can be provided for storing data fragments. Cloud users may choose the desired security level and accordingly the cloud broker fragment and store the data using Rabin s IDA (Information Dispersal Algorithm) approach. IDA helps in dispersing data parts to subset of CSPs which are generated by Knapsack algorithm that maintains a balance between cloud users desired QoS requirements and cost.