Distributed Environment Research Articles

Optimized Feature Learning for Anti-Inflammatory Peptide Prediction Using Parallel Distributed Computing

With recent advancements in computational biology, high throughput Next-Generation Sequencing (NGS) has become a de facto standard technology for gene expression studies, including DNAs, RNAs, and proteins; however, it generates several millions of sequences in a single run. Moreover, the raw sequencing datasets are increasing exponentially, doubling in size every 18 months, leading to a big data issue in computational biology. Moreover, inflammatory illnesses and boosting immune function have recently attracted a lot of attention, yet accurate recognition of Anti-Inflammatory Peptides (AIPs) through a biological process is time-consuming as therapeutic agents for inflammatory-related diseases. Similarly, precise classification of these AIPs is challenging for traditional technology and conventional machine learning algorithms. Parallel and distributed computing models and deep neural networks have become major computing platforms for big data analytics now required in computational biology. This study proposes an efficient high-throughput anti-inflammatory peptide predictor based on a parallel deep neural network model. The model performance is extensively evaluated regarding performance measurement parameters such as accuracy, efficiency, scalability, and speedup in sequential and distributed environments. The encoding sequence data were balanced using the SMOTETomek approach, resulting in a high-accuracy performance. The parallel deep neural network demonstrated high speed up and scalability compared to other traditional classification algorithms study’s outcome could promote a parallel-based model for predicting anti-Inflammatory Peptides.

Multi-Dimensional Resource Allocation in Distributed Data Centers Using Deep Reinforcement Learning

With the development of edge-cloud computing technologies, distributed data centers (DCs) have been extensively deployed across the global Internet. Since different users/applications have heterogeneous requirements on specific types of ICT resources in distributed DCs, how to optimize such heterogeneous resources under dynamic and even uncertain environments becomes a challenging issue. Traditional approaches are not able to provide effective solutions for multi-dimensional resource allocation that involves the balanced utilization across different resource types in distributed DC environments. This paper presents a reinforcement learning based approach for multi-dimensional resource allocation (termed as NESRL-MRM) that is able to achieve balanced utilization and availability of resources in dynamic environments. To train NESRL-MRM’s agent with sufficiently quick wall-clock time but without the loss of exploration diversity in the search space, a natural evolution strategy (NES) is employed to approximate the gradient of the reward function. To realistically evaluate the performance of NESRL-MRM, our simulation evaluations are based on real-world workload traces from Amazon EC2 and Google datacenters. Our results show that NESRL-MRM is able to achieve significant improvement over the existing approaches in balancing the utilization of multi-dimensional DC resources, which leads to substantially reduced blocking probability of future incoming workload demands.

Configuration and Placement of Serverless Applications Using Statistical Learning

In the last decade, serverless computing emerged as a new compelling paradigm for the deployment of cloud applications and services. It represents an evolution of cloud computing with a simplified programming model, that aims to abstract away most operational concerns. Running serverless applications requires users to configure multiple parameters, such as memory, CPU, cloud provider, etc. While relatively simpler, configuring such parameters correctly while minimizing cost and meeting delay constraints is not trivial. In this paper, we present, a framework that uses Bayesian Optimization to find the optimal resource configuration and placement for functions in a serverless application. uses statistical learning techniques to intelligently collect samples and predict the cost and execution time of a serverless function across unseen configuration values. Our framework uses the predicted cost and execution time on available locations to select the “best” configuration parameters and placement for running a serverless application while satisfying customer objectives. We evaluate on AWS Lambda with real-world applications consisting of multiple functions (both linear chains and service graphs), where we successfully found optimal/near-optimal configurations. We also evaluate over a wide range of simulated distributed cloud environments that confirm the efficacy of our approach.

Influence of a Distributed Learning Environment on Learners’ Performance

Online high-quality teaching and learning behaviors can take place anytime and anywhere, which comprehensively breaks through isolation of physical space. Individual learning in distributed learning environment is an activity that is transformed from process of individual self-cognition to social cognition of the group. It attaches great importance to interactive behaviors in learning. By strengthening cooperative learning, learning motivation can be further stimulated and learning performance can be improved. A questionnaire about the influence of distributed learning environment on learners’ learning performance was designed. Influence of four components of distributed learning environment interaction on learners’ learning performance was measured, and the mediating effect of collaborative learning on the interaction between learners’ learning performance and distributed learning environment was analyzed. Results show systematic management support of self-knowledge, support of learners’ expression and support of self-reflection can significantly improve learning performance of learners. Collaborative learning plays a full mediating effect in the interaction between distributed learning environment and learning performance of learners. Grade is significant at 0.01 level for learning performance. Conclusions are of great significance to promote occurrence of effective learning behaviors in interactive process of collaborative learning from the perspective of distributed learning environment.

Becoming an Online Wine Taster: An Ethnographic Study on the Digital Mediation of Taste

There has been a growing interest in taste research in the HCI and CSCW communities. However, the focus is more on stimulating the senses, while the socio-cultural aspects have received less attention. However, individual taste perception is mediated through social interaction and collective negotiation and is not only dependent on physical stimulation. Therefore, we study the digital mediation of taste by drawing on ethnographic research of four online wine tastings and one self-organized event. Hence, we investigated the materials, associated meanings, competences, procedures, and engagements that shaped the performative character of tasting practices. We illustrate how the tastings are built around the taste-making process and how online contexts differ in providing a more diverse and distributed environment. We then explore the implications of our findings for the further mediation of taste as a social and democratized phenomenon through online interaction.

GA Based Task Allocation In Agile Software Development In Distributed Environment

GA Based Task Allocation In Agile Software Development In Distributed Environment

Real-Time Streaming Analytics using Big Data Paradigm and Predictive Modelling based on Deep Learning

With the evolution of distributed streaming platforms analysing humongous time series data, which is streamed continuously from IoT devices become lot easier. In most of the IoT networks the data are in motion or in data centre/cloud. It is possible to process this data in real time similar to edge devices using the big data framework. In data intensive applications predictive analytics require more resources to perform complex computations. Apache Flink framework is capable of performing real time streaming of schema less data and scales very high in distributed environment with low latency, it is used to collect and store the data in the cloud. This work suggests a suitable environment to collect, transport, preprocess and aggregate the data stream to perform predictive analytics using deep learning models. Deep learning automatically extracts features and builds models after training, it has the potential to solve problems that can't be solved by conventional machine learning models. Therefore, the use of algorithms based on deep learning is recommended for forecasting temporal data. Also, we discuss a number of different deep learning forecasting models and analyse the performance of different deep learning forecasting models in order to determine which one is the effective model for single step, multi step and multi variant methods based on error functions with respect to streamed sensor data.

B-SAHIC: A blockchain based secured and automated health insurance claim processing system

In recent years, blockchain technology has been successfully used in many distributed environments where different stakeholders, who do not have any trust between them, interact with each other through a secured and transparent platform. The use of blockchain in healthcare insurance industry has not been studied methodically. In this study, we develop a blockchain based Secured and Automated Health Insurance Claim (B-SAHIC) processing system. First, we design an Entity-Relationship (ER) diagram identifying all actors with their respective data and relations between them. We also develop the business model and algorithms for all necessary steps. We implement the system in Hyperledger Fabric and deploy a smart contract to implement these algorithms. We introduce CouchDB to store the OffChain data where we store World-State Database. B-SAHIC provides a web-based portal for all actors who interact with the blockchain. Privacy of clients’ claim-related data is ensured by encrypting treatment-related data with a new key that is derived uniquely for each new submission done of clients’ personal documents. We have also deployed Hyperledger Explorer, a user-friendly web application tool for monitoring the health state of each node participating on the blockchain network. We have studied the performance of B-SAHIC for two to six peer nodes. Moreover, our performance study shows that B-SAHIC is fast and scalable. In our study, the average query latency is decreased from 0.07 second (for two peer nodes) to 0.02 second (for six peer nodes) in case of 1300 queries per second while the average transaction latency remains unchanged (around 3.5 seconds) for 300 transactions per second. Moreover, B-SAHIC consumes minimum resources, around 350MB only for two peer nodes. We believe that the development process of this blockchain based platform can be applicable for the automation of other insurance industry too.

JQPro:Join Query Processing in a Distributed System for Big RDF Data Using the Hash-Merge Join Technique

In the last decade, the volume of semantic data has increased exponentially, with the number of Resource Description Framework (RDF) datasets exceeding trillions of triples in RDF repositories. Hence, the size of RDF datasets continues to grow. However, with the increasing number of RDF triples, complex multiple RDF queries are becoming a significant demand. Sometimes, such complex queries produce many common sub-expressions in a single query or over multiple queries running as a batch. In addition, it is also difficult to minimize the number of RDF queries and processing time for a large amount of related data in a typical distributed environment encounter. To address this complication, we introduce a join query processing model for big RDF data, called JQPro. By adopting a MapReduce framework in JQPro, we developed three new algorithms, which are hash-join, sort-merge, and enhanced MapReduce-join for join query processing of RDF data. Based on an experiment conducted, the result showed that the JQPro model outperformed the two popular algorithms, gStore and RDF-3X, with respect to the average execution time. Furthermore, the JQPro model was also tested against RDF-3X, RDFox, and PARJs using the LUBM benchmark. The result showed that the JQPro model had better performance in comparison with the other models. In conclusion, the findings showed that JQPro achieved improved performance with 87.77% in terms of execution time. Hence, in comparison with the selected models, JQPro performs better.

Efficient and Trusted Data Sharing in a Sharding-Enabled Vehicular Blockchain

Data sharing is essential for future autonomous vehicles in vehicular networks. Building trust in the shared data can be challenging among mutually unknown vehicles in an open and distributed environment. Blockchain is the key to establish trusted data sharing in a decentralized vehicular environment. However, the low efficiency (e.g., transaction per second, TPS) of blockchain is the bottleneck for the efficient data sharing. Moreover, the highly dynamic topology, the imperfect wireless links, and the limited radio coverage may further slow down the blockchain consensus process. To improve the efficiency and trustworthiness of data sharing in a vehicular network, we propose a sharding-enabled vehicular blockchain system. Specifically, the blockchain-enabled vehicles are dynamically allocated to different shards based on their geographical locations so that transactions are processed in parallel. We design an efficient and trusted sharding-enabled blockchain consensus mechanism by integrating a reputation-assisted intrashard consensus and a vehicle-assisted inter-shard consensus. The reputation assistance in intra-shard consensus is designed to defend against the collusive false-block attack that may happen in a small-scale shard environment. The vehicle-assisted intershard consensus is designed to tackle the issues caused by defective roadside units (RSUs). Simulation results demonstrate the improved performance on efficiency and trustworthiness of the proposed system. In addition, trusted data sharing can be established even under the conditions of malicious vehicles or attacked RSUs.

Software Training in High Energy Physics

Among the upgrades in current high energy physics (HEP) experiments and the new facilities coming online, solving software challenges has become integral to the success of the collaborations. The demand for human resources skilled in both HEP and software domains is increasing. With a highly distributed environment in human resources, the sustainability of the HEP ecosystem requires a continuous effort in the equipment of physicists with the required abilities in software development. In this paper, the collective software training program in HEP and its activities led by the HEP Software Foundation (HSF) and the Institute for Research and Innovation in Software in HEP (IRIS-HEP) are presented. Experiment-agnostic, open, and accessible modules for training have been developed, focusing on common software material with ranges from core software skills needed by everyone to advanced training required to produce high-quality sustainable software. A basic software curriculum was built, and an introductory software training event has been prepared to serve HEP entrants. This program serves individuals with transferable skills that are becoming increasingly important to careers in the realm of software and computing, whether inside or outside HEP.

Blockchain Technology: Security Issues, Healthcare Applications, Challenges and Future Trends

Blockchain technology provides a data structure with inherent security properties that include cryptography, decentralization, and consensus, which ensure trust in transactions. It covers widely applicable usages, such as in intelligent manufacturing, finance, the Internet of things (IoT), medicine and health, and many different areas, especially in medical health data security and privacy protection areas. Its natural attributes, such as contracts and consensus mechanisms, have leading-edge advantages in protecting data confidentiality, integrity, and availability. The security issues are gradually revealed with in-depth research and vigorous development. Unlike traditional paper storage methods, modern medical records are stored electronically. Blockchain technology provided a decentralized solution to the trust-less issues between distrusting parties without third-party guarantees, but the “trust-less” security through technology was easily misunderstood and hindered the security differences between public and private blockchains appropriately. The mentioned advantages and disadvantages motivated us to provide an advancement and comprehensive study regarding the applicability of blockchain technology. This paper focuses on the healthcare security issues in blockchain and sorts out the security risks in six layers of blockchain technology by comparing and analyzing existing security measures. It also explores and defines the different security attacks and challenges when applying blockchain technology, which promotes theoretical research and robust security protocol development in the current and future distributed work environment.

A Framework for User-Focused Electronic Health Record System Leveraging Hyperledger Fabric

This research study aims to examine the possibilities of Hyperledger Fabric (HLF) in the healthcare sector. The study addresses the gap in the knowledge base through developing customization techniques to enable the simplicity and efficacy of Electronic Medical Records (EMR) adoption for healthcare industry applications. The focus of this research explores methods of using blockchain technology that prioritise users. The investigation of several concepts used in developing web applications has been determined. The study identified that an open-source project, known as Hyperledger Fabric, can be utilised to construct a novel method of storing EMRs. The framework provides a test network that can be customised to satisfy the need of several projects, including storing medical records. This research additionally outlines the difficulties encountered and problems that need to be resolved before Hyperledger Fabric can be successfully implemented in healthcare systems. Considering all types of blockchains available, the needs are met by Hyperledger Fabric, which offers a distributed and secure environment for healthcare systems. Blockchain has the potential to transform healthcare by putting the patient at the centre of the system and enhancing health data protection and interoperability. Also, by using grant and revoke access mechanisms, patients have complete control over their medical information as well as authorized doctors who are allowed to view records. This functionality is made possible by the chaincode defined in the blockchain platform. The research study has both practitioner and research implications for the development of secure blockchain-based EMRs.

Control Methodologies to Mitigate and Regulate Second-Order Ripples in DC–AC Conversions and Microgrids: A Brief Review

Second-order ripples occur in the voltage and current during any DC–AC power conversion. These conversions occur in the voltage source inverters (VSIs), current source inverters (CSIs), and various single-stage inverters (SSIs) topologies. The second-order ripples lead to oscillating source node currents and DC bus voltages when there is an interconnection between the AC and DC microgrids or when an AC load is connected to the DC bus of the microgrid. Second-order ripples have various detrimental effects on the sources and the battery storage. In the storage battery, they lead to the depletion of electrodes. They also lead to stress in the converter or inverter components. This may lead to the failure of a component and hence affect the reliability of the system. Furthermore, the second-order ripple currents (SRCs) lead to ripple torque in wind turbines and lead to mechanical stress. SRCs cause a rise in the temperature of photovoltaic panels. An increase in the temperature of PV panels leads to a reduction in the power generated. Furthermore, the second-order voltage and current oscillations lead to a varying maximum power point in PV panels. Hence, the maximum power may not be extracted from it. To mitigate SRCs, oversizing of the components is needed. To improve the lifespan of the sources, storage, and converter components, the SRCs must be mitigated or kept within the desired limits. In the literature, different methodologies have been proposed to mitigate and regulate these second-order ripple components. This manuscript presents a comprehensive review of different effects of second-order ripples on different sources and the methodologies adopted to mitigate the ripples. Different active power decoupling methodologies, virtual impedance-based methodologies, pulse width modulation-based signal injection methodologies, and control methods adopted in distributed power generation methods for DC microgrids have been presented. The application of ripple control methods spans from single converters such as SSIs and VSIs to a network of interconnected converters. Furthermore, different challenges in the field of virtual impedance control and ripple mitigation in distributed power generation environments are discussed. This paper brings a review regarding control methodologies to mitigate and regulate second-order ripples in DC–AC conversions and microgrids.

Efficient, Traceable and Privacy-Aware Data Access Control in Distributed Cloud-Based IoD Systems

The emerging combination of Internet of Things (IoT) and aerospace integration aided by satellite and 6G communication techniques has stimulated the Internet of Unmanned Aerial Vehicles (UAVs), i.e., Internet of Drones (IoD). To accommodate and share the enormous real-time UAV data, cloud-based IoD is an inevitable choice to lower the heavy burden of mobile UAVs. Nevertheless, how to protect highly sensitive UAV data in such a honest-but-curious, open and distributed environment with resource-limited UAVs is a significant challenge. Although our previous work (PATLDAC) in SPNCE’21 devises a cloud-based UAV data access control scheme with policy privacy protection, limited access time and user traceability, it incurs inflexible and centralized cloud data storage and access as well as untrustworthy metadata in untrusted cloud environment for data access and user tracing. To this end, we further propose a blockchain-based privacy-aware data access control (BPADAC) scheme for distributed and secure UAV data sharing in cloud-based IoD. Based on fine-grained, traceable and privacy-preserving UAV data access characteristic of our previous work, we extend it by leveraging blockchain and Distributed Hash Table (DHT) for distributed and trustful UAV data access and storage, together with reliable and limited access mechanism to guarantee cloud UAV data sharing service provision. We also design public and undeniable user tracing mechanism to prevent user key abuse with traitor denial. Finally, we present formal security analysis and prototype the system leveraging the smart contracts of Ethereum blockchain for performance evaluation to show the feasibility of BPADAC.

GeFL: Gradient Encryption-Aided Privacy Preserved Federated Learning for Autonomous Vehicles

Autonomous vehicles (AVs) are getting popular because of their usage in a wide range of applications like delivery systems, self-driving taxis, and ambulances. AVs utilize the power of machine learning (ML) and deep learning (DL) algorithms to improve their self-driving learning experiences. The sudden surge in the number of AVs raises the need for distributed learning ecosystem to optimize their self-driving experiences at a rapid pace. Toward this goal, federated learning (FL) benefits, which can create a distributed learning environment for AVs. But, the traditional FL transfers the raw input data directly to a server, which leads to privacy concerns among the end-users. The concept of blockchain helps us to protect privacy, but it requires additional computational infrastructure. The extra infrastructure increases the operational cost for the company handling and maintaining the AVs. Motivated by this, in this paper, the authors introduced the concept of gradient encryption in FL, which preserves the users’ privacy without the additional computation requirements. The computational power present in the edge devices helps to fine-tune the local model and encrypt the input data to preserve privacy without any drop in performance. For performance evaluation, the authors have built a German traffic sign recognition system using a convolutional neural network (CNN) algorithm-based classification system and GeFL. The simulation process is carried out over a wide range of input parameters to analyze the performance at scale. Simulation results of GeFL outperform the conventional FL-based algorithms in terms of accuracy, i.e., 2% higher. Also, the amount of data transferred among the devices in the network is nearly three times less in GeFL compared to the traditional FL.

ПРОБЛЕМА АВТОМАТИЗИРОВАННОЙ ДЕТЕКЦИИ ВИКТИМНОГО ПОВЕДЕНИЯ ИНДИВИДА С ТОЧКИ ЗРЕНИЯ КАЧЕСТВА СЕРВИСА

This article discusses the issue of automatic detection of potentially victimized behavior of an individual in a digital environment in terms of the quality of the service provided. An analysis of precedents for detecting the victim behavior demonstrates a connection between the presence of multiple false accounts in an individual and a high potential for the victim behavior. Such an inference makes it possible to detect individuals prone to victimization through the detection of their false accounts. Also, the analysis of recent works in the subject area demonstrates that one of the most effective methods for detecting false aliases in the network is stylometric analysis, which, however, is the most time-consuming and requires the analysis of the largest amount of data. In this regard, the issue of compliance with the quality of services to the declared level in terms of the execution time of operations becomes relevant, which is a non-trivial task when searching and processing information in highly distributed digital environments. The article presents approaches and methods for reducing the execution time of search operations and detection of individuals with the victim behavior by adapting system parameters to search conditions, as well as by parallelizing and placing data processing close to their sources. Experimental calculations are also presented, on the basis of which conclusions are drawn about the appropriateness of using the adaptive properties of systems and elements of the fog computing concept.

TRANSFORMING STRATEGY FROM INDUSTRIAL AUTOMATION TO ADVANCED LEAN AUTOMATION

Industrial automation refers to the control of machines and processes in order to manage variables. Automated manufacturing systems have evolved as advanced technologies have been incorporated, and theoretical approaches have evolved from mass production to intelligent manufacturing, each step introducing superior manufacturing concepts and models, which allowed increasing processes, organization and work. One of the newest approaches refers to distributive automation, in which the system allows reconfiguration and self-organization of autonomous subsystems in a distribute environment. By incorporating integrative Industry 4.0 technologies, an automated industrial system becomes an intelligent, fully digitized and reconfigurable manufacturing applications platform. At the same time, through digitization and virtualization, the automated manufacturing system can employ the Advanced Lean Manufacturing conceptual support. This paper addresses the transforming challenges of an existing automated manufacturing system to the Advanced Lean adaptive automated manufacturing level, taking into account the limits due to attributes, properties and capabilities of physical manufacturing assets. The proposed solution describes a versatile and dynamic architecture, which allows the organization / reorganization of manufacturing flows, scalability and connection with the external environment, according to advanced cyber manufacturing requests.

Large-scale digital forensic investigation for Windows registry on Apache Spark.

In this study, we investigate large-scale digital forensic investigation on Apache Spark using a Windows registry. Because the Windows registry depends on the system on which it operates, the existing forensic methods on the Windows registry have been targeted on the Windows registry in a single system. However, it is a critical issue to analyze large-scale registry data collected from several Windows systems because it allows us to detect suspiciously changed data by comparing the Windows registry in multiple systems. To this end, we devise distributed algorithms to analyze large-scale registry data collected from multiple Windows systems on the Apache Spark framework. First, we define three main scenarios in which we classify the existing registry forensic studies into them. Second, we propose an algorithm to load the Windows registry into the Hadoop distributed file system (HDFS) for subsequent forensics. Third, we propose a distributed algorithm for each defined forensic scenario using Apache Spark operations. Through extensive experiments using eight nodes in an actual distributed environment, we demonstrate that the proposed method can perform forensics efficiently on large-scale registry data. Specifically, we perform forensics on 1.52 GB of Windows registry data collected from four computers and show that the proposed algorithms can reduce the processing time by up to approximately 3.31 times, as we increase the number of CPUs from 1 to 8 and the number of worker nodes from 2 to 8. Because the distributed algorithms on Apache Spark require the inherent network and MapReduce overheads, this improvement of the processing performance verifies the efficiency and scalability of the proposed algorithms.

An efficient load balancing framework for deploying resource schedulingin cloud based communication in healthcare

In distributed environments, cloud computing is widely used to manage user requests for resources and services. Resource scheduling is used to handle user requests for resources based on priorities within a given time frame. In today's environment, healthcare systems and management rely on smart devices connected to the internet. These devices deal with the massive amounts of data processed and detected by smart medical sensors without sacrificing performance factors like throughput and latency. This has prompted the requirement for load balancing among the smart operational devices to prevent any insensitivity. Load balancing is used to manage large amounts of data in both a centralised and distributed manner. In this paper, a load balancing framework for deploying resource scheduling for cloud-based healthcare environments is presented. We use reinforcement learning algorithms such as GA, SARSA, and Q-learning for resource scheduling. These algorithms are used to predict the optimal solution to manage load in cloud-based healthcare environments. The proposed mechanism is energy efficient, has a low make span, and consumes less latency time. The proposed mechanism is implemented using MATLAB. The performance of the proposed mechanism is analysed using latency, make-span, and throughput performance metrics. In the proposed mechanism, throughput is high, whereas make span is less than in the existing technique.