No abstract

No abstract

Advancement in the field of information technology has enabled us to optimize the use of physical infrastructure with the help of virtualization, most of which is only used in specific applications like vitalizing a local area network on switch or vitalizing a compute resource to run more than one operating system on a desktop.The benefits of same has not been completely implemented and utilized in education and research environments where we require a large amount of IT resources. Cloud computing has emerged as one step further solution to deploy all virtualized IT resources as per need on self-service basis as a rental method for the users. Different cloud deployment methods have emerged according to the need of users, institutions and communities. This paper describes the framework of design and implementing a simple private cloud for education and research based on chip level virtualization with hypervisor to establish a simple model of Infrastructure as a service in cloud computing. General TermsPrivate cloud framework

Cloud computing is a collective, optimized usage of IT resources which can be accessed from anywhere and performs the task of providing IT resources and services through various platforms as a service for users. Cloud users use different services of Cloud service providers and eventually end up keeping their data (in various forms) in cloud multi tenant environment. Multi-tenancy in cloud environment makes data vulnerable though with times different forms of threats and corresponding securities are already implemented in cloud environment. Although many of these services provide key functionality such as uploading and retrieving files by a precise user, more on going to the side of advanced services it offer features such as shared folders, real-time collaboration, and minimization of data transfers or unlimited storage space. As data is placed publically it require to search ways to protect the data from unauthorized access, files are uploaded publically and need to retrieve them securely with token ensuring possession proofs. In this paper we have presented a case of multilevel security application for ensuring data integrity (prevention and detection) in cloud environment.

GPU acceleration of compute-intensive applications has emerged as a new research frontier with phenomenal success-rates. Such applications are characterized by large data-sets being processed by singular functional units (FUs) often described as SIMD (Single Instruction Multiple Data) computing.Moreover, with the proliferation of internet and its easy access on myriad devices, has resulted in huge amount of data generation. Initially, such data was considered disconnected and not related. But with the advent of semantic web, data has been found to be highly co-related and relevant. Organizing such huge amount of data and subsequently processing requires parallel processing framework that is both distributed and scalable. Graphical processing units (GPUs) are being actively probed in the domain of Big Data analysis, machine learning, and augmented reality since such applications are characterized by massive data spanned and generated over distributed network. GPUs provide a parallel programming framework using CUDA (Compute Unified Device Architecture) that can be utilized to efficiently collate and make inferences on these massive data-sets. Further, GPU multicores are available at commodity rates thus providing an option for cheap and low-power alternatives.The exponential growth of semantic web and the resultant generation of large-scale RDF (Resource Description Framework) triples pose new challenges in the domain of RDFstorage and retrieval. RDF data consist of triples which need to be efficiently indexed. Following are some of the many challenges related to efficient indexing of RDF triples: As RDF-triples extensively contain recursive redundancies, self-joins so formed are inefficient. Self-joins also lead to large scale null values.This paper presents the research initiatives of conducting literature survey of contemporary indices including those under active research, which matches the goals as outlined in above sections. Comparing the efficiency of different variety of indices that have been suggested for large data-sets (Map reduce, B+ tree Hashed Index, 3-level-cascade hash index, braided B+ tree index, etc.)