Search

With the increasing digitalization of the economy, it is becoming increasingly important for organizations to embrace digital innovation. The emergence of blockchain technology exemplifies the disruptive impact of digital innovation and, simultaneously, poses challenges for organizations. Hence, organizations are in transition with regard to fostering its organizational adoption. This kind of adoption is reflected in far-reaching measures in terms of technology, organization, and people. The measures are governed by project management that aligns internal measures with the outside environment. Based on this framework, our research builds on an empirical study to explore the organizational adoption of blockchain technology. We conducted in-depth interviews in 11 cases with experts leading the adoption of blockchain. The objective of our study is to better understand organizational adoption and to explore the factors influencing adoption. Our results uncover a multitude of factors and inherent tensions. The paper's contributions are twofold. First, our findings can be used as guidelines for organizations planning to adopt digital innovations such as blockchain and to effectively cope with the related phenomena. Second, the factors advance the literature on blockchain by conceptualizing its organizational adoption and aligning the factors to the organizational adoption framework. [ABSTRACT FROM AUTHOR]

Technological advancements, such as machine learning or the concept of blockchain, have the potential to reshape entire branches of industry as a disruptive technology, but also offer the possibility to optimize existing processes. In this paper, we present a concept for vendor managed inventory relationships based on blockchain technology. Hence, we initially define necessary processes and transactions that highlight a special need for transparency and trust between the participants. We designed and developed a proof of concept that addresses the weaknesses of current VMI solutions by performing certain process steps in the blockchain. Thus, we implement a method based on smart contracts, which can interact with existing systems by using different data interfaces and therefore represents an application of Blockchain 3.0. The present work is intended to explain the concept to the reader and to support researchers for the future development of further industrial applications of the blockchain. [ABSTRACT FROM AUTHOR]

The growing content delivery traffic becomes a challenge for efficient usage of network resources. Optimal utilization of scarce radio resources may be achieved by deployment of cloud capabilities in the radio access network i.e. by deployment of Multi-access Edge Computing (MEC). In this paper, we propose RESTful application programming interfaces (API) that enable third party MEC applications to send messages to end user devices in a specific geographical area. Using the API, an authorized mobile edge application may send a broadcast message to the network; monitor the delivery status of a sent broadcast message; and receive notifications about the message delivery status. The proposed APIs are described by use cases illustrating their functionality, data model representing the resources, and broadcast delivery status models as seen by mobile edge application and MEC server. Both models are formally described and verified. [ABSTRACT FROM AUTHOR]

This paper proposes a novel virtual shared memory framework, Soft Memory Box (SMB), which directly shares the memory of remote nodes among distributed processes to improve communication performance/speed via deep learning parameter sharing. [ABSTRACT FROM AUTHOR]

The CMS experiment is working to integrate an increasing number of High Performance Computing (HPC) resources into its distributed computing infrastructure. The case of the Barcelona Supercomputing Center (BSC) is particularly challenging as severe network restrictions prevent the use of CMS standard computing solutions. The CIEMAT CMS group has performed significant work in order to overcome these constraints and make BSC resources available to CMS. The developments include adapting the workload management tools, replicating the CMS software repository to BSC storage, providing an alternative access to detector conditions data, and setting up a service to transfer produced output data to a nearby storage facility. In this work, we discuss the current status of this integration activity and present recent developments, such as a front-end service to improve slot usage efficiency and an enhanced transfer service that supports the staging of input data for workflows at BSC. Moreover, significant efforts have been devoted to improving the scalability of the deployed solution, automating its operation, and simplifying the matchmaking of CMS workflows that are suitable for execution at BSC. [ABSTRACT FROM AUTHOR]

Predicting the performance of various infrastructure design options in complex federated infrastructures with computing sites distributed over a wide area network that support a plethora of users and workflows, such as the Worldwide LHC Computing Grid (WLCG), is not trivial. Due to the complexity and size of these infrastructures, it is not feasible to deploy experimental test-beds at large scales merely for the purpose of comparing and evaluating alternate designs. An alternative is to study the behaviours of these systems using simulation. This approach has been used successfully in the past to identify efficient and practical infrastructure designs for High Energy Physics (HEP). A prominent example is the Monarc simulation framework, which was used to study the initial structure of the WLCG. New simulation capabilities are needed to simulate large-scale heterogeneous computing systems with complex networks, data access and caching patterns. A modern tool to simulate HEP workloads that execute on distributed computing infrastructures based on the SimGrid and WRENCH simulation frameworks is outlined. Studies of its accuracy and scalability are presented using HEP as a case-study. Hypothetical adjustments to prevailing computing architectures in HEP are studied providing insights into the dynamics of a part of the WLCG and candidates for improvements. [ABSTRACT FROM AUTHOR]

Operational analytics is the direction of research related to the analysis of the current state of computing processes and the prediction of future states in order to anticipate imbalances and take timely measures to stabilize a complex system. There are two relevant areas in ATLAS Distributed Computing that are currently the focus of studies: user physics analysis including the forecast of popularity of data samples among users, and evaluating WLCG centers for their readiness to process user analysis payloads. Studying these areas is challenging due to the complexity involved, as it requires a comprehensive understanding of numerous boundary conditions typically found in large-scale distributed computing infrastructures. Forecasts of data popularity are problematic without the categorization of user tasks by their types (data transformation or physics analysis), which do not always appear on the surface but may induce noise, which introduces significant distortions for predictive analysis. Evaluating the WLCG resources by their analysis workloads is also a challenging task as it is necessary to find a balance between the workload of the resource, its performance, the waiting time for jobs on it, as well as the volume of jobs that it processes. This is especially difficult in a heterogeneous computing environment, where legacy resources are used along with modern high-performance machines. We will look at these areas of research in detail and discuss what tools and methods are used in our work, demonstrating results already obtained. [ABSTRACT FROM AUTHOR]

Document clustering is the process of grouping related documents with each other. It facilitates organizing search results and document management. K-means algorithm and its variant bisecting k-means have been applied for document clustering and approved good clustering results. The increased number of available documents requires utilizing of distributed computing and huge number of computer resources which are available through cloud computing. This paper introduces Var-secting k-means algorithm. In addition to generating binary tree as in Bisecting k-means algorithm, it can generate hierarchy tree with variable number of nodes per tree level. The experimental results show that Var -secting k-means algorithm utilizes distributed computing nodes better than Bisecting k-means, especially when using MapReduce programming model. [ABSTRACT FROM AUTHOR]

In this work, we introduce slot selection and co-allocation algorithms for parallel jobs in distributed computing with non-dedicated resources. A single slot is a time span that can be assigned to a task, which is a part of a job. The job launch requires a co-allocation of a specified number of slots starting synchronously. The challenge is that slots associated with different CPU nodes of distributed computational environments may have arbitrary start and finish points that do not match. Some existing algorithms assign a job to the first set of slots matching the resource request without any optimization (the first fit type), while other algorithms are based on an exhaustive search. In this paper, algorithms for effective slot selection of linear complexity on an available slots number are studied and compared with known approaches. The novelty of the proposed approach consists of allocating alternative sets of slots. It provides possibilities to optimize job scheduling. [ABSTRACT FROM AUTHOR]

PESOS has established itself as a forum that brings together software engineering researchers and practitioners working in the areas of service-oriented systems to discuss research challenges, new developments and applications, as well as methods, techniques, experiences, and tools to support engineering, evolution and adaptation of service-oriented systems. The technical advances and growing adoption of Cloud computing is creating new challenges for the PESOS the software services community to explore the approaches to better engineer software systems that are designed, developed, operated and governed in the context of the Cloud. We again attracted high-quality submissions on a diverse set of relevant topics such as better approaches to engineering service-based collaborative systems, Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS) models of cloud computing and associated software quality attributes. PESOS 2015 will continue to be the key forum for collecting case studies and artifacts for educators and researchers in this area. [ABSTRACT FROM AUTHOR]

Genetic Algorithm (GA), one of the artificial intelligence algorithms, performs much better than the other algorithms for the document clustering. However, it has problem known as the premature convergence occurrence. So, Fuzzy Logic based GA (FLGA) was proposed to solve it. Nevertheless, it has still weakness such as the parameter dependence problem. In order to overcome this problem, the Multi-Objective Genetic Algorithms (MOGAs), NSGA-II and SPEA2, have been proposed. The MOGAs showed high performance compared to other algorithms including the general GA, but their computational times have increased. In order to reduce these computational times, the distributed computing method, MATLAB Distributed Computing (MDC) with 10 processors, is applied to the MOGAs in this paper. The performances of MOGAs on MDC show about 12% higher than others. [ABSTRACT FROM PUBLISHER]

This paper first briefly introduces the Mobile Agent technology, then analyzes some other Mobile Agent communication mechanisms by comparing them with each other, based on which we generate Assistant Agent by Mobile Agent with combination of Shortest Path First (SPF) algorithm to receive the message as a new Mobile Agent communication system and give the experimental results to demonstrate. [ABSTRACT FROM PUBLISHER]

Information systems based on the cloud computing model and peer-to-peer (P2P) model are now getting popular. In the cloud computing model, a cloud of servers support thin clients with various types of service like Web pages and databases. On the other hand, every computer is peer and there is no centralized coordinator in the P2P model. It is getting more significant to discuss how to reduce the total electric power consumption of computers in information systems to realize eco-society. In this paper, we consider a Web type of application on P2P overlay networks. A model shows how much server peer consumes electric power to performWeb requests from client peers. Here we are creating VM migration to consolidate the server utilization. An algorithm is used for a client peer to select a server peer in a collection of server peers so that the total power consumption can be reduced. Lastly, we evaluate the algorithms in terms of the total power consumption and throughput compared with traditional round robin algorithm. [ABSTRACT FROM PUBLISHER]

This paper presents the impact analysis of the hyper visor layer over database applications. The conclusions were reached by performing experiments with hybrid environments, composed by bare-metal servers and two types of hyper visors. The main objectives were to compare the behaviour of database applications and to determine the maximum service capacity of a virtualized database server on different environments. With the results, it was possible to assess the scalability of a database environment, according to the type of hyper visor and the number of simultaneous connections, and decide how and when to expand or reduce the virtualized infrastructure. [ABSTRACT FROM PUBLISHER]

Asynchronous task allocation is a fundamental problem in distributed computing in which p asynchronous processes must execute a set of m tasks. Also known as write-all or do-all, this problem been studied extensively, both independently and as a key building block for various distributed algorithms. In this paper, we break new ground on this classic problem: we introduce the To-Do Tree concurrent data structure, which improves on the best known randomized and deterministic upper bounds. In the presence of an adaptive adversary, the randomized To-Do Tree algorithm has O(m+plogplog2m) work complexity. We then show that there exists a deterministic variant of the To-Do Tree algorithm with work complexity O(m+p log5 m log2 max(m, p)). For all values of m and p, our algorithms are within log factors of the O(m + p log p) lower bound for this problem. The key technical ingredient in our results is a new approach for analyzing concurrent executions against a strong adaptive scheduler. This technique allows us to handle the complex dependencies between the processes' coin flips and their scheduling, and to tightly bound the work needed to perform subsets of the tasks. [ABSTRACT FROM PUBLISHER]

This paper examines the integration of the NSF/TCPP Core Curriculum Recommendations in a liberal arts undergraduate setting. We examine how parallel and distributed computing concepts can be incorporated across the breadth of the undergraduate curriculum. As a model of such an integration, changes are proposed to Data Structures and Design and Analysis of Algorithms. These changes were implemented in Design and Analysis of Algorithms and the results were compared to previous iterations of that course taught by the same instructor. The student feedback received shows that the introduction of these topics made the course more engaging and conveyed an adequate introduction to this material. [ABSTRACT FROM PUBLISHER]

Cloud is a promising source for computing resources when solving scientific problems. Several large companies have entered the cloud computing market in recent times and are providing cloud computing services like Infrastructure as a Service (IaaS) that provide virtual computer machines on demand. These virtual machines can easily be used for the high performance computing needs by applying distributed computing methods and frameworks. The SciCloud project at the University of Tartu has studied adapting scientific computing algorithms to different distributed computing frameworks like MapReduce and its optimizations. The study has concluded that the current solutions do not fully support complex iterative algorithms, which are commonly used in scientific simulations. This paper presents a design for a new distributed computing framework, Stratus, which fully supports scientific computing algorithms and takes advantage of the characteristics that have made cloud such a convenient and popular source for computing resources. The paper also describes the motivation for creating a brand new solution, outlines its architecture and design, and gives an overview on how algorithms can be adapted to this framework. [ABSTRACT FROM PUBLISHER]

In this paper we present a Platform-as-a-Service (PaaS) approach for rapid development of wireless sensor network (WSN) applications based on the dinam-mite concept, i.e. an embedded web-based development environment and run-time platform for WSN systems integrated in a single information appliance. The PaaS is hosted by a cloud of dinam-mite nodes which facilitates the on-demand development, deployment and integration of WSN applications. We introduce the dinam Cloud architecture and focus, in this paper, on the PaaS layer established by the dinam-mite nodes. In addition to the description of this so-called dinam PaaS, a performance analysis of the dinam-mite node towards its applicability to forming a dinam PaaS layer is demonstrated. We then present the MASON mobile vehicular network as an example of such a WSN which delivers spatially and temporally fine-grained environmental measurements within the city of Beijing, and illustrate how to utilize the dinam PaaS for integrating the data from the MASON network into its back-end business system. Finally, we discuss the five essential properties of the Cloud Computing stack, according to the NIST definition, with respect to the dinam PaaS and illustrate the benefits of the dinam PaaS for system integration as well as WSN application development. [ABSTRACT FROM PUBLISHER]

Using the high fidelity model in the multidisciplinary design optimization (MDO) application is becoming more and more popular, but it brings complex coupling relationship and huge computing intensity, which need efforts to be well organized to reduce the computation time. In this paper, a parallel computing environment distributed in local area network (LAN) was established based on the collaborative optimization (CO) method. The distributed computing system communicated with Transfer Control Protocol (TCP) and used XML database to store and manage the data. The proposed system was applied in a remote sensing satellite system parameters design considering four subsystems, which includes structure, attitude control (AC), power, and thermal control (TC). The result shows that the efficiency can be improved by about 60% relative to the sequential running plan. [ABSTRACT FROM PUBLISHER]

The computational humanity is flattering extremely bulky and multifaceted. Cloud computing is becoming one of the most expanding methodologies in the computing industry. It is a novel approach for the deliverance of IT services on the World Wide Web. This model provides computing resources in the puddle for consumers, all the way through Internet. In cloud computing, resource allocation and scheduling of numerous aggregate web services is an imperative and demanding quandary. This paper estimates the various network resource allocation strategies and their applications in Cloud Computing Environment. A brief description for network resource allocation in Cloud Computing, based on differentially adapted dynamic proportions, has also been done. This paper addresses and categorizes the foremost challenges normal to the resource allocation progress of cloud computing in terms of diverse types of resource allocation techniques. [ABSTRACT FROM PUBLISHER]

In distributed computing field, there are many challenges, and one of them is leader election in distributed system. It is really tough task to find suitable and efficient algorithms for leader election. The main role of an elected leader is that it performs a centralized coordination after being selected and manages the use of a shared resource in an optimal manner. Whenever a failure occurs the new leader is elected by nodes using various algorithms so that nodes can continue working. In this paper, the proposal is a new approach, the improved heap tree mechanism for electing the coordinator. The higher efficiency and better performance in the presented algorithms with respect to the existing algorithms is validated through results. [ABSTRACT FROM PUBLISHER]

In recent years, MapReduce parallel computing model has gained lots of attentions from industry and academia. In Google, Yahoo, Facebook, etc., it has played a very good effect, which greatly simplifies the design of large-scale data-intensive applications. MapReduce-based systems were originally used to manage massive unstructured and semi-structured data, for example: to generate the inverted index, to calculate web page rank, log analysis, etc. Therefore, current MapReduce systems don't give more considerations to the optimization of structured data, for example: it uses the brute-force scanning mode to process the whole datasets, which confront the common workflow of structured data processing, range query and analysis. To address the problem, this paper propose to build a global B-tree like index on top of hadoop distributed file system for structured data, and use the global index to eliminate unnecessary map tasks during range queries, thereby reducing the overhead of data I/O and tasks scheduling, which not only reduces query response time, but also greatly optimizes system resource utilization. [ABSTRACT FROM PUBLISHER]

In recent years, many service providers have started migrating their service offerings to cloud infrastructure. Sometimes, parts of the service workflow can however not be moved to cloud environments. This can occur due to client policies, or because some services are linked to physical client-site devices. The result of the migration is then a hybrid cloud environment, where part of the services are executed within the client network, while most of the processing is moved to the cloud. Migration to the cloud enables a more flexible deployment of services, but also increases the strain on underlying networks as most tasks are partially handled in a remote cloud, and no longer just in the local network. An important question that providers must answer before new service workflows are deployed is whether they can provide the workflow with sufficient quality of service, and whether the deployment will impact existing service workflows. In this paper we discuss strategies based on multi-commodity flow problems, a subset of graph flow problems that can be used to determine whether new service workflows can be sufficiently provisioned, and whether the addition of new workflows can negatively impact the performance of existing flows. We evaluate the proposed solution by comparing the performance of three approaches with respect to the number of successful workflows and with respect to their execution speed. [ABSTRACT FROM PUBLISHER]

With network developed quickly and used widely, how to store mass data has become a problem we have to face. Traditional methods usually add more servers into to increase computing speed and storage ability, while the hardware cost is very expensive and the storage efficiency is low. Through deep study of MapReduce programming model and Hadoop framework, this paper presented a mass data storage model based on Hadoop framework, then we designed and implemented a mass data storage system for campus network based on cloud computing. This system can effectively solved issues in mass data storage management in teaching and study. This system is cheap, easily operated, fast dealing and stable. [ABSTRACT FROM PUBLISHER]

Nowadays, evacuation is mostly based on individual decisions, which is unreliable. This paper introduces a practical solution towards emergency evacuation by guiding people out of building efficiently. Built on wireless sensor network (WSN), our system is capable of learning and dynamically generating evacuation strategy. Distributed mechanism and backup mechanism feature the scalability and invulnerability. Simulations prove the efficiency and reliability of the system. [ABSTRACT FROM PUBLISHER]

In this paper, we propose three lower bounds to the constrained task allocation problem (CTAP). These three lower bounds are: a heuristic lower bound (HLB), a linear programming relaxation (LPR) lower bound and a lagrangean relaxation with subgradient optimization (LRSO) based lower bound. We empirically compare the performance of these three bounds under different task executing times. Our experiments with real-world CTAPs indicate that the LPR provides efficient and tight lower bounds for the CTAP. [ABSTRACT FROM PUBLISHER]

This paper presents an approach to perform bit rate reduction Transco ding by video segmentation. The paper shows how a high performance distributed video transcoder can be built using multiple processing units and a Message Passing Interface based parallel programming model. The computation parallelization and data distribution among computing units is discussed. For data distribution coarse grain approach is used in which significant gain in terms of execution speedup is obtained. The segmentation of video stream with (1) equal size having unequal number of intra frames and (2) unequal size having equal number of intra frames is performed to achieve high performance. The results show that the proposed distributed video transcoder provides very short startup times. [ABSTRACT FROM PUBLISHER]

This paper presents an ontology-based approach for data quality inference on streaming observation data originating from large-scale sensor networks. We evaluate this approach in the context of an existing river basin monitoring program called the Intelligent River®. Our current methods for data quality evaluation are compared with the ontology-based inference methods described in this paper. We present an architecture that incorporates semantic inference into a publish/subscribe messaging middleware, allowing data quality inference to occur on real-time data streams. Our preliminary benchmark results indicate delays of 100ms for basic data quality checks based on an existing semantic web software framework. We demonstrate how these results can be maintained under increasing sensor data traffic rates by allowing inference software agents to work in parallel. These results indicate that data quality inference using the semantic sensor network paradigm is viable solution for data intensive, large-scale sensor networks. [ABSTRACT FROM PUBLISHER]

D-Clean is a functional coordination language for distributed computation. The language was designed for the need of high-level process description and communication coordination of functional programs distributed over a cluster. The pure functional computational nodes required language primitives to control the dataflow in a distributed process-network. Therefore, in order to achieve parallel features, we created an extension for the lazy functional programming language Clean using new language elements. D-Clean is compiled to an intermediate level language called D-Box, which is designed for the description of the computational nodes. Every D-Clean construct generates a D-Box expression. The D-Box expressions hide the low level implementation details and enable direct control over the process-network. The asynchronous communication is based on language-independent middleware services. Earlier we have presented the syntax and the semantics of both coordination languages. Practical experiences showed the difficulties of distributed program development, especially in testing and debugging. This paper aims to provide software comprehension application for a better understanding and usage of the D-Clean distributed system. We model the elements and the behaviour of the D-Clean system using C++ templates. The strong type system of C++ templates guarantees the correctness of the model. Using templates we can avoid run-time overhead achieving impressive efficiency. [ABSTRACT FROM AUTHOR]

Template engines are widely used for separation of business and presentation logic. They are commonly used in web applications for clean rendering of HTML pages. Another area of usage is message formatting in distributed applications where they transform objects to appropriate representations. This paper explores the possibility of using templates for a reverse process - for creating objects starting from their representations. We present the prototype of engine that we have developed, and describe benefits and drawbacks of this approach. [ABSTRACT FROM AUTHOR]

Cloud computing is rapidly becoming an important platform for research in Software Engineering. Despite the vibe and huge literature on commercial cloud environments, there is, however, very little research on how to capture, model, design and implement new software applications that can make intelligent use of the cloud. In this paper we propose a new abstraction that explores a fundamental aspect of cloud systems - data elasticity. The Container Database (CDB) abstraction provides a cloud-based solution for scenarios where device local storage is not sufficient for manipulating data. To demonstrate the viability of the proposed approach we present an implementation of the CDB abstraction as an API designed to work with a Visual Studio 2010 plug-in, using Windows Azure Cloud services. [ABSTRACT FROM AUTHOR]

Dynamic specification mining involves discovering software behavior from traces for the purpose of program comprehension and bug detection. However, mining program behavior from execution traces is difficult for concurrent/distributed programs. Specifically, the inherent partial order relationships among events occurring across processes pose a big challenge to specification mining. In this paper, we propose a framework for mining partial orders so as to understand concurrent program behavior. Our miner takes in a set of concurrent program traces, and produces a message sequence graph (MSG) to represent the concurrent program behavior. An MSG represents a graph where the nodes of the graph are partial orders, represented as Message Sequence Charts. Mining an MSG allows us to understand concurrent program behaviors since the nodes of the MSG depict important "phases" or "interaction snippets" involving several concurrently executing processes. To demonstrate the power of this technique, we conducted experiments on mining behaviors of several fairly complex distributed systems. We show that our miner can produce the corresponding MSGs with both high precision and recall. [ABSTRACT FROM AUTHOR]

Web services providing weather information are not new. The existing web services working on this kind of fields can provide more précised information if the concerned data is collected in a distributed fashion using a sensor network. Longer the lifetime of the sensor network, longer is the service provided without interruption. In this paper we propose a web service for providing weather information with a sensor network as the backbone. We have used a Fermat point based forwarding technique to minimize the energy consumption of the sensor network which eventually helps the web service work in an uninterrupted fashion for a longer duration, as the life time of the network has prolonged. [ABSTRACT FROM AUTHOR]

Among various scheduling algorithms, duplication based scheduling algorithms provide superior performance than others. General scheduling algorithms consist of task prioritizing phase and processor selection phase. In this paper, we propose a noveltask prioritizing algorithm, called WPD, which is specialized for duplication based scheduling algorithms. A primary goal of WPD is to minimize the schedule length of applications. The performance of WPD has been observedby applying it to processor selection phase of conventional scheduling algorithm. The comparison results show that WPD provides better performance than algorithms with conventional task prioritizing phase. [ABSTRACT FROM AUTHOR]

Optimally mapping diverse independent tasks onto machines in a distributed system, in general, is an NP hard problem. Further, distributed systems augmented with GPUs are quite different from those without GPUs. In order to better utilize machines with both CPUs and GPUs, dynamic load balancing has been introduced and it can significantly affect the scheduling performance of such distributed systems. In this paper, we proposed four heuristics, adapted them into a set of common assumptions, added dynamic load balancing, and compared their testing results based on the simulation experiments. The performance of these heuristics, as well as the effect of task redistribution, was carefully studied. MCT with task redistribution out-performs other algorithms. Task redistribution can significantly improve performance, especially when the speed of CPU and GPU is close to each other. [ABSTRACT FROM AUTHOR]

Computational Grid enables the aggregation and sharing of geographically distributed computational resources for solving various large-scale applications. Due to the widespread use of resources, systems are highly prone to errors and failures. Hence fault tolerance plays a key role in grid to avoid the problem of unreliability. The two main techniques for implementing fault tolerance in grid environment are check pointing and replication. In this paper, a survey of the different replication techniques is presented. The goal of replication is to ensure at least one replica is always able to continue and complete the computation in the event of failure of other resources. This paper describes the different job replication techniques that have been proposed to improve the fault tolerance in the grid environment. This study will help the researchers who are undertaking their research in improving the fault tolerance in grid, to analyze and compare different job replication techniques with their own work. [ABSTRACT FROM AUTHOR]

Due to the variations of the network in actual distributed system, failure detectors without adaptive mechanism cannot meet the requirements of QOS of applications. Adaptive failure detectors should dynamically adjust the detecting quality according to the real-time state of the network. Assuming that delay of the message and loss of the messages is a random probability, the failure detection model based on the predicted message delay is proposed in this paper. A PAC-AFD adaptive failure detection algorithm is realized based on the above model based on the prediction from historical message delay and it contains checking idea. Experimental results show that the algorithm can relieve the effect of delay of the message and loss of the message on the failure detection while ensuring the accuracy and completeness of detection. [ABSTRACT FROM AUTHOR]

Cloud Computing is an emerging technology in which the research community and industries have recently embarked. However, the infrastructures of most cloud computing systems today are invisible to the research community, or are not explicitly designed to the researchers interested in cloud computing systems. In this paper, we depict the Cloud computing paradigm from a variety of aspects, such as definitions, features, and technologies. Moreover, some Cloud computing Systems are illustrated as well as their architectural trade-offs. It would contribute to better comprehension of the Cloud computing technology so as to enable some communities to design more efficient clouds, and the adoption of this novel computing approach is facilitated in scientific areas. [ABSTRACT FROM AUTHOR]

Hadoop, based on the popular MapReduce framework, is an open-source distributed computing framework that has been gaining much popularity and usage. It aims to allow programmers to focus on building applications that deals with processing large amount of data, without having to handle other issues when performing parallel computations. However, tuning the performance of Hadoop applications is not an easy task due to the level of abstraction the framework works on. In this paper, we present some of the challenges and issues that are to be considered in performance tuning when running applications in Hadoop. [ABSTRACT FROM AUTHOR]

M&S is used in the German Armed Forces for supporting processes in the following application areas: • analysis and planning, • procurement, • missions, and • training and exercises. These four areas of application are closely connected to each other, especially in the context of • supporting the method of Concept Development and Experimentation (CD&E), • support of networking and linking national and international simulations and information systems, • the entailing support for the realization of Network Centric Warfare (NCW). Therefore, the further development of M&S towards interconnected systems by applying a Distributed and Integrated Test Bed1 is indispensable to achieve the required capabilities and objectives. Such a test bed can be used as an armament test bed for a fast and effective evaluation of technical and technological solutions in a realistic and operational test environment to support all phases of the development and procurement cycle of material and equipment as well as the other application areas of M&S. And such a test bed has to satisfy the demands concerning the acceptance of interconnected simulations as well as to meet the requirements resulting from the NCW idea of joint and combined missions by adapting to national and international standards. In this paper, we summarize the results of a large combined study of the German government, industry and research institutes that addresses the architectural aspects for such a distributed integrated test bed and encompasses issues like ensuring fair fight conditions as well as the reproducibility and reliability of simulation results. [ABSTRACT FROM AUTHOR]

Growth in the adoption of distributed software development business models continues to outpace the development of robust governance schemes for them. In the absence of specific welltested governance schemes, distributed software teams continue to employ the normative process frameworks, and software engineering practices that were primarily developed in the colocated context. In this position paper, we present a possible research direction for developing and testing governance schemes specifically suited for distributed software development projects. Our proposal draws examples from our recent research initiatives and presents an actionable road map for researchers and practitioners who are working towards establishing governance schemes for distributed software teams. [ABSTRACT FROM AUTHOR]

A retrial queue is characterized by the feature that an incoming call that cannot receive service immediately upon arrival joins a virtual waiting room (orbit) and retries for service in a random time. Most of models in the literature assume that the servers only serve incoming calls. Recently, Artalejo and Phung-Duc [1] study two-way communication queueing models where the server makes outgoing calls in its idle time. It is assumed in [1] that outgoing calls are independent of incoming calls in the orbit. In this paper, we relax this assumption by introducing the interaction between the servers and the orbit. In particular, we consider the case where outgoing calls may be those in the orbit. For this extended model, we establish the necessary and sufficient stability condition. It turns out that the stability condition depends on the retrial rate which is a new finding in the queueing literature. [ABSTRACT FROM AUTHOR]

We focus on the numerical analysis of a coupled queueing system with Poisson arrivals and exponentially distributed service times. Such a system consists of multiple queues served by a single server. Service is synchronised meaning that there is a departure from every queue upon service completion and there is no service whenever one of the queues is empty. It was shown before that the terms in the Maclaurin series expansion of the steady-state distribution of this queueing system when the service rate is sent to 0 (overload) can be calculated efficiently. In the present paper we extend this approach to lower loads. We focus on a sequence of Taylor series expansions of the stationary distribution around increasing service rates. For each series expansion, we use Jacobi iteration to calculate the terms in the series expansion where the initial solution is the approximation found by the preceding series expansion. As the generator matrix of the queueing system at hand is sparse, the numerical complexity of a single Jacobi iteration is O(NMK), where N is the order of the series expansion, K is the number of queues and M is the size of the state space. Having a good initial solution reduces the number of Jacobi iterations considerably, meaning that we can find a sequence of good approximations of the steady state probabilities fast. [ABSTRACT FROM AUTHOR]

With the continuous development of machine learning, industry and academia have released a lot of machine learning frameworks based on distributed computing platform, and have been widely used. However, the existing framework of machine learning is limited by the limitations of machine learning algorithm itself, such as the choice of parameters and the interference of noises, the high using threshold and so on. This paper introduces the research background of machine learning framework, and combined with the commonly used random forest algorithm in machine learning classification algorithm, puts forward the research objectives and content, proposes an improved adaptive random forest algorithm (referred to as ARF), and on the basis of ARF, designs and implements the machine learning framework. [ABSTRACT FROM AUTHOR]

Nowadays, technology grows fast and cloud computing and big data become important parts of the businesses. Cloud computing brings many advantages, such as flexibility, easy recovery, automatic software updates, availability -- which allows users to work from everywhere. One of the most important characteristic is that cloud computing provides resources on demand, all we need is an internet connection. So, cloud computing is adopted by more and more companies due its benefits. An important aspect of the society is education. As the technology is expanding continuously, also the education needs to move in parallel. As a consequence, we can follow different courses online, on the eLearning platforms. There are many eLearning platforms, with different subjects of learning, some courses being provided by top universities, or top companies with certificate of completion after a final exam. Due to cloud computing benefits, eLearning platforms could be developed as services. Still, a sensitive subject regarding to cloud computing is its security. The cloud computing security is threatened from many points of view, because of its complexity. The security issues could occur at the communication level (because of shared infrastructure, virtualization, and configuration), architectural level (security challenges for data storage, web applications and APIs, access control), or contractual and legal level. The researchers are working continuously in order to provide strong security. For example, at the architectural level, could be used attribute based encryption or identity based encryption techniques. In this paper we propose a new framework for an eLearning environment, provided as a service, such that, it could be accessed from everywhere and on every type of device. We secure our framework using attribute based encryption scheme applied in cloud computing. [ABSTRACT FROM AUTHOR]

Community detection is an advanced graph operation that is used to reveal tightly-knit groups of vertices (aka. communities) in real-world networks. Given the intractability of the problem, efficient heuristics are used in practice. Yet, even the best of these state-of-the-art heuristics can become computationally demanding over large inputs and can generate workloads that exhibit inherent irregularity in data movement on manycore platforms. In this paper, we posit that effective acceleration of the graph community detection operation can be achieved by reducing the cost of data movement through a combined innovation at both software and hardware levels. More specifically, we first propose an efficient software-level parallelization of community detection that uses approximate updates to cleverly exploit a diminishing returns property of the algorithm. Secondly, as a way to augment this innovation at the software layer, we design an efficient Wireless Network on Chip (WiNoC) architecture that is suited to handle the irregular on-chip data movements exhibited by the community detection algorithm under both unicast- and broadcast-heavy cache coherence protocols. Experimental results show that our resulting WiNoCenabled manycore platform achieves on average 52% savings in execution time, without compromising on the quality of the outputs, when compared to a traditional manycore platform designed with a wireline mesh NoC and running community detection without employing approximate updates. [ABSTRACT FROM AUTHOR]

Data intensive science focuses on extracting, exploring and mining knowledge from a huge ever growing data that is obtained from various instruments and simulations. Scientific applications earlier were using High Performance Computers, then focus was on cluster computing and then towards grid computing. As the technologies in the networking domain improved, distributed networking and cloud computing have become prevalent. Scientific applications using cloud computing paradigm, faces several challenges such as data transportation cost, computing costs, data privacy and data security. A small dedicated cluster, with a cloud like interface, designed from low cost commodity machines, is often affordable by a group or an individual researcher, and is reasonably cost-effective. In this paper, the authors demonstrate how a low cost inexpensive cluster setup can be designed and customized as a cloud based data analytical services framework, to provide data storage and analytical capabilities, for exploratory needs of a small or medium scale scientific application. [ABSTRACT FROM AUTHOR]

The aim of the paper is to create a polymorphic informational model of a ZVS Buck quasi-resonant DC-DC converter for the modeling purposes of the object. For the creation of the model is applied flexible open standards for setting, storing, publishing and exchange of data in distributed information environment. The created model is useful for creation of many and different by type variants with different configuration of the composing elements and different inner model of the examined object. [ABSTRACT FROM AUTHOR]

The performance and scalability of cellular automata, when executed on parallel/distributed machines, are limited by the necessity of synchronizing all the nodes at each time step, i.e., a node can execute only after the execution of the previous step at all the other nodes. However, these synchronization requirements can be relaxed: a node can execute one step after synchronizing only with the adjacent nodes. In this fashion, different nodes can execute different time steps. This can be a notable advantageous in many novel and increasingly popular applications of cellular automata, such as smart city applications, simulation of natural phenomena, etc., in which the execution times can be different and variable, due to the heterogeneity of machines and/or data and/or executed functions. Indeed, a longer execution time at a node does not slow down the execution at all the other nodes but only at the neighboring nodes. This is particularly advantageous when the nodes that act as bottlenecks vary during the application execution. The goal of the paper is to analyze the benefits that can be achieved with the described asynchronous implementation of cellular automata, when compared to the classical all-to-all synchronization pattern. The performance and scalability have been evaluated through a Petri net model, as this model is very useful to represent the synchronization barrier among nodes. We examined the usual case in which the territory is partitioned into a number of regions, and the computation associated with a region is assigned to a computing node. We considered both the cases of mono-dimensional and two-dimensional partitioning. The results show that the advantage obtained through the asynchronous execution, when compared to the all-to-all synchronous approach is notable, and it can be as large as 90% in terms of speedup. [ABSTRACT FROM AUTHOR]

Infrastructures supporting distributed scientific collaborations must address competing goals in both providing high performance access to resources while simultaneously securing the infrastructure against security threats. The NetBASILISK project is attempting to improve the security of such infrastructures while not adversely impacting their performance. This paper will present our work to create a benchmark and monitoring infrastructure that allows us to test for any degradation in transferring data into a NetBASILISK protected site. [ABSTRACT FROM AUTHOR]