Big Data management is an interesting research challenge for all storage vendors. Since data can be structured or unstructured, hence variety of storage systems has been designed to meet storage... Show More

Big Data management is an interesting research challenge for all storage vendors. Since data can be structured or unstructured, hence variety of storage systems has been designed to meet storage requirement as per organization's demands. The article focuses on different kinds of storage systems, their architecture and implementations. The first portion of the article describes different examples of structured (PostgreSQL) and unstructured databases (MongoDB, OrientDB and Neo4j) along with data models and comparative performance analysis between them. The second portion of the paper focuses on cloud storage systems. As an example of cloud storage, Google Cloud Storage and mainly its implementation details have been discussed. The aim of the article is not to eulogize any particular storage system, but to clearly point out that every storage has a role to play in the industry. It depends on the enterprise to identify the requirements and deploy the storage systems.

Erroneous or incomplete data generated from various sources can have direct impact in business analysis. Extracted data from sources need to load into data warehouse after required transformation to... Show More

Erroneous or incomplete data generated from various sources can have direct impact in business analysis. Extracted data from sources need to load into data warehouse after required transformation to reduce error and minimize data loss. This process is also known as Extraction-Transformation-Loading (ETL). High-level view of the system activities can be visualized by conceptual modeling of ETL process. It provides the advantage of pre-identification of system error, cost minimization, scope and risk assessment etc. A new modeling approach is proposed for conceptualization ETL process by using a standard Systems Modeling Language (SysML). For handling increasing complexity of any system model, it is preferable to go through verification and validation process in early stage of system development. In this article, the authors' previous work is extended by presenting a MBSE based approach to automate the SysML model's validation by using No Magic simulator. Here, the main objective is to overcome the gap between modeling and simulation and to examine the performance of the proposed SysML model. The usefulness of the authors' approach is exhibited by using a use case scenario.

The cloud computing presents a type of assignments and systems which occupy distributed resources to execute a role in a distributed way. Cloud computing make use of the online systems on the web to... Show More

The cloud computing presents a type of assignments and systems which occupy distributed resources to execute a role in a distributed way. Cloud computing make use of the online systems on the web to assist the implementation of complicated assignments; that need huge-scale computation. It was said with the intention of in our living world; we can find it challenging to balance workloads of cloud computing among assignments (jobs or tasks) and systems (machines or nodes), so the majority of the time we have to promote a condition to unbalanced assignment problems (unequal task allocations). The present article submits a new technique to solve the unequal task allocation problems. The technique is offered in an algorithmic model and put into practice on the several groups of input to investigate the presentation and usefulness of the works. An evaluation is prepared with the presented approach. It makes sure that the proposed approach provides a better outcome by comparing with some other existing algorithms.

This research article attempts to analytically determine the factors, significant for safety, in connection with driving of automotives as well as to develop a conceptual model of the driving... Show More

This research article attempts to analytically determine the factors, significant for safety, in connection with driving of automotives as well as to develop a conceptual model of the driving assistance system, using the knowledge about such factors. Millions of casualties due to road accidents, happen worldwide every year and the annual average of lives lost in India alone is about hundred and fifty thousand. The causes of such accidents are attributed to road characteristic and condition, driving faults, driving conditions or traffic environmental factors and defects or functional failure in vehicle mechanism. Studies have focused primarily on these factors without associating the ‘weather' which has been reported as in a work but as an isolated factor without including the above three. This work includes all the four stated factors in modelling the driver assistance system for automatic speed control with warning system module. Further, to predict accident rates in a particular region a model using adaptive neuro fuzzy inference system (ANFIS) is proposed in this work, which may be used by the vehicle manufactures to select the right product variant to minimise accidents.

Distributed storage allocation problems are an important optimization problem in reliable distributed storage, which aims to minimize storage cost while maximizing error recovery probability by... Show More

Distributed storage allocation problems are an important optimization problem in reliable distributed storage, which aims to minimize storage cost while maximizing error recovery probability by optimal storage of data in distributed storage nodes. A key characteristic of distributed storage is that data is stored in remote servers across a network. Thus, network resources especially communication links are an expensive and non-trivial resource which should be optimized as well. In this article, the authors present a simulation-based study of the network characteristics of a distributed storage network in the light of several allocation patterns. By varying the allocation patterns, the authors have demonstrated the interdependence between network bandwidth, defined in terms of link capacity and allocation pattern using network throughput as a metric. Motivated by observing the importance of network resource as an important cost metric, the authors have formalized an optimization problem that jointly minimizes both the storage cost and the cost of network resources. A hybrid meta heuristic algorithm is employed that solves this optimization problem by allocating data in a distributed storage system. Experimental results validate the efficacy of the algorithm.

The wireless medical sensor networks WMSN play a crucial role in healthcare monitoring remotely. In remote healthcare monitoring, the sensor nodes are deployed in patient's body for collecting... Show More

The wireless medical sensor networks WMSN play a crucial role in healthcare monitoring remotely. In remote healthcare monitoring, the sensor nodes are deployed in patient's body for collecting physiological data and transmit these data over an insecure channel. The patient's health information is highly sensitive and important. Any malicious modification in physiological data will make wrong diagnoses and harm the patient health. Therefore, privacy, data security, and user authentication are extremely important for accessing patient's real-time heath information over an insecure channel. In this regard, this article proposes a secure and robust two-factor based remote user authentication protocol for healthcare monitoring. The authentication proof has done with the help of BAN logic, which ensures that the proposed scheme provides mutual authentication and session key agreement securely. The informal security verification proves that the developed protocol is secure from various security attacks. The simulation of the proposed scheme has been done using AVISPA tool, whose simulation results confirm that the proposed scheme is secure from active and passive attacks. Performance evaluation shows that the proposed protocol is efficient in terms of security features, computation cost, communication cost, and execution time.