Reliable Multi-cloud Storage Architecture Based on Erasure Code to Improve Storage Performance and Failure Recovery

Abstract

Cloud computing is a way to increase the capacity dynamically without investing in new infrastructure, training new users as well as licensing new software. The increasing popularity of cloud storage services has lead companies that handle critical content to think about using these services for their daily storage needs. Power systems records, huge Medical datasets and financial content are partial critical content that could be moved to the cloud. To this, cloud storage providers reveal challenges in achieving content availability, reliability, storage overhead and content integrity that require high efficiency on single cloud storage architecture. Cloud storage disperses content blocks over single cloud storage nodes. This study proposes Reliable Multi-Cloud Storage Architecture (RMCSA) based on Erasure Codes, that content can be dispersed to different cloud storage providers to reduce challenges which lead to Lock-In. A Multi-Cloud Control Node will manage other Control Nodes evolved in the cloud as well as service migration across sub-clouds. Currently, replication is the most applied technique which disperses content replica but standing against content loss becomes critical, hence ineffective. To assure stability in storage costs with best practice to assure content failure or loss recovery, we applied a Maximum Distance Separable such as Reed-Solomon to our RMCSA. Analysis shows that RMCSA resolves a number of storage issues and recommends Zfec library as erasure code emulator due to its accuracy.