Visible to the public Performance Evaluation of a Fragmented Secret Share System

TitlePerformance Evaluation of a Fragmented Secret Share System
Publication TypeConference Paper
Year of Publication2017
AuthorsUkwandu, E., Buchanan, W. J., Russell, G.
Conference Name2017 International Conference On Cyber Situational Awareness, Data Analytics And Assessment (Cyber SA)
KeywordsBig Data, big data storage, cloud computing, composability, consistent data availability, cryptography, Data, Data fragmentation, data handling, data integrity, data retrievals, disaster contention, Encryption, file recovery, fragmented secret share system, Human Behavior, key, keyless mechanism, large scale data infrastructure, large-scale data leakage, Memory, metadata, Metrics, performance evaluation, pubcrawl, public storage environments, resilience, Resiliency, resilient mechanism, Robustness, secret shares, secure data storage, Secure File Sharing, storage locations, storage management, threshold sizes, thresholds scheme
AbstractThere are many risks in moving data into public storage environments, along with an increasing threat around large-scale data leakage. Secret sharing scheme has been proposed as a keyless and resilient mechanism to mitigate this, but scaling through large scale data infrastructure has remained the bane of using secret sharing scheme in big data storage and retrievals. This work applies secret sharing methods as used in cryptography to create robust and secure data storage and retrievals in conjunction with data fragmentation. It outlines two different methods of distributing data equally to storage locations as well as recovering them in such a manner that ensures consistent data availability irrespective of file size and type. Our experiments consist of two different methods - data and key shares. Using our experimental results, we were able to validate previous works on the effects of threshold on file recovery. Results obtained also revealed the varying effects of share writing to and retrieval from storage locations other than computer memory. The implication is that increase in fragment size at varying file and threshold sizes rather than add overheads to file recovery, do so on creation instead, underscoring the importance of choosing a varying fragment size as file size increases.
Citation Keyukwandu_performance_2017