Distributed Programming Paradigm Research Articles

Scalable Diversified Top-k Pattern Matching in Big Graphs

Typically, graph pattern matching is expressed in terms of subgraph isomorphism. Graph simulation and its variants were introduced to reduce the time complexity and obtain more meaningful results in big graphs. Among these models, the matching subgraphs obtained by tight simulation are more compact and topologically closer to the pattern graph than results produced by other approaches. However, the number of resulting subgraphs can be huge, overlapping each other and unequally relaxed from the pattern graph. Hence, we introduce a ranking and diversification method for tight simulation results, which allows the user to obtain the most diversified and relevant matching subgraphs. This approach exploits the weights on edges of the big graph to express the interest of the matching subgraph by tight simulation. Furthermore, we provide distributed scalable algorithms to evaluate the proposed methods based on distributed programming paradigms. The experiments on real data graphs succeed in demonstrating the effectiveness of the proposed models and the efficiency of the associated algorithms. The result diversification reached 123% within a time frame that does not exceed 40%, on average, of the duration required for tight simulation graph pattern matching.

A Novel Nodesets-Based Frequent Itemset Mining Algorithm for Big Data using MapReduce

Due to the rapid growth of data from different sources in organizations, the traditional tools and techniques that cannot handle such huge data are known as big data which is in a scalable fashion. Similarly, many existing frequent itemset mining algorithms have good performance but scalability problems as they cannot exploit parallel processing power available locally or in cloud infrastructure. Since big data and cloud ecosystem overcomes the barriers or limitations in computing resources, it is a natural choice to use distributed programming paradigms such as Map Reduce. In this paper, we propose a novel algorithm known as A Nodesets-based Fast and Scalable Frequent Itemset Mining (FSFIM) to extract frequent itemsets from Big Data. Here, Pre-Order Coding (POC) tree is used to represent data and improve speed in processing. Nodeset is the underlying data structure that is efficient in discovering frequent itemsets. FSFIM is found to be faster and more scalable in mining frequent itemsets. When compared with its predecessors such as Node-lists and N-lists, the Nodesets save half of the memory as they need only either pre-order or post-order coding. Cloudera's Distribution of Hadoop (CDH), a MapReduce framework, is used for empirical study. A prototype application is built to evaluate the performance of the FSFIM. Experimental results revealed that FSFIM outperforms existing algorithms such as Mahout PFP, Mlib PFP, and Big FIM. FSFIM is more scalable and found to be an ideal candidate for real-time applications that mine frequent itemsets from Big Data.

Using software visualization to support the teaching of distributed programming

In this paper, we introduce MARVEL, a system designed to simplify the teaching of MapReduce, a popular distributed programming paradigm, through software visualization. At its core, it allows a teacher to describe and recreate a MapReduce application by interactively requesting, through a graphical interface, the execution of a sequence of MapReduce transformations that target an input dataset. Then, the execution of each operation is illustrated on the screen by playing an appropriate graphical animation stage, highlighting aspects related to its distributed nature. The sequence of all animation stages, played back one after the other in a sequential order, results in a visualization of the whole algorithm. The content of the resulting visualization is not simulated or fictitious, but reflects the real behavior of the requested operations, thanks to the adoption of an architecture based on a real instance of a distributed system running on Apache Spark. On the teacher’s side, it is expected that by using MARVEL he/she will spend less time preparing materials and will be able to design a more interactive lesson than with electronic slides or a whiteboard. To test the effectiveness of the proposed approach on the learner side, we also conducted a small scientific experiment with a class of volunteer students who formed a control group. The results are encouraging, showing that the use of software visualization guarantees students a learning experience at least equivalent to that of conventional approaches.

An Analysis of Distributed Document Clustering Using MapReduce Based K-Means Algorithm

Clustering is considered as one of the important data mining techniques. Document clustering is among many applications of clustering. The traditional clustering algorithms are proven inefficient for clustering rapidly generating large real world datasets. As a solution, traditional clustering algorithms are modified using distributed programming paradigm. MapReduce is a popular distributed programming paradigm designed for Hadoop distributed framework. This paper demonstrates a MapReduce based modification of K-Means clustering algorithm for document datasets. The result shows that the proposed algorithm is efficient than traditional K-Means for all size of document datasets clustering. The experiments also show that the MapReduce clustering works more efficiently when the dataset size and Hadoop cluster sizes are large.

Tree Partitioning Reduction

Solving tridiagonal linear-equation systems is a fundamental computing kernel in a wide range of scientific and engineering applications, and its computation can be modeled with parallel algorithms. These parallel solvers are typically designed to compute problems whose data fit in a common shared-memory space where all the cores taking part in the computation have access. However, when the problem size is large, data cannot be entirely stored in the common shared-memory space, and a high number of high-latency communications are performed. One alternative is to partition the problem among different memory spaces. At this point, conventional parallel algorithms do not facilitate the partition of computation in independent tiles, since each reduction depends on equations that may be in different tiles. This article proposes an algorithm based on a tree reduction, called the Tree Partitioning Reduction (TPR) method, which partitions the problem into independent slices that can be partially computed in parallel within different common shared-memory spaces. The TPR method can be implemented for any parallel and distributed programming paradigm. Furthermore, in this work, TPR is efficiently implemented for CUDA GPUs to solve large size problems, providing highly competitive performance results with respect to existing packages, being, on average, 22.03× faster than CUSPARSE.

Learning distributed discrete Bayesian Network Classifiers under MapReduce with Apache Spark

The challenge of scalability has always been a focus on Machine Learning research, where improved algorithms and new techniques are proposed in a constant basis to deal with more complex problems. With the advent of Big Data, this challenge has been intensified, in which new large scale datasets overwhelm the majority of available techniques. The community has resorted to Cloud Computing and distributed programming paradigms as the most immediate solution where Apache Spark has proven to be the most promising framework. In this paper we focus on the problem of supervised classification, exploring the family of the so called Bayesian Network Classifiers by studying their adaptability to the MapReduce and Apache Spark frameworks. We will analyse a range of algorithms and propose distributed versions of them. Our approach is based on a general framework for learning this probabilistic models from large scale and high dimensional data, the latter being a problem with less support in the literature. We also present an extensive experimental evaluation of our proposal over a wide set of problems and different elastic configurations of a computing cluster to show the full extent of the scalability properties of our framework. Additional material and the software to reproduce our experiments can be found on the supplementary website http://simd.albacete.org/supplements/distributed_bncs.html.

Leveraging MapReduce to efficiently extract associations between biomedical concepts from large text data

Large biomedical text data represents an important source of information that not only enables researchers to discover in-depth knowledge about biological systems, but also helps healthcare professionals do evidence-based medicine in clinical settings. However, investigating and analyzing these data is often both data-intensive and computation-intensive. In this paper, we investigate how to use MapReduce, a parallel and distributed programming paradigm, to efficiently mine the associations between biomedical concepts extracted from a large set of biomedical articles. First, biomedical concepts were obtained by matching text to Unified Medical Language System (UMLS) Metathesaurus, a biomedical vocabulary and standard database. Then we developed a MapReduce algorithm that could be used to calculate a category of interestingness measures defined on the basis of a 2×2 contingency table. This algorithm consists of two MapReduce jobs and takes a stripes approach to reduce the number of intermediate results. Experiments were conducted using Amazon Elastic MapReduce (EMR) with an input of 33,960 articles from TREC (Text REtrieval Conference) 2006 Genomics Track. Performance test indicated that our algorithm had approximately linear scalability and was more efficient than a ‘‘pairs’’ approach in the literature. The physician in our project team evaluated a subset of the association mining results related to drug-disease treatment and found that meaningful association rules ranked high.

MEMoMR: Accelerate MapReduce via reuse of intermediate results

SummaryMapReduce has been widely regarded as a flexible, scalable, and easy‐to‐use distributed programming paradigm for big data processing such as social network data analysis on cloud computing platforms. To embrace the upcoming of big data era, many efforts have been devoted to accelerating the MapReduce performance from different aspects, especially intermediate result reusing like Dache. In this paper, we observe that existing intermediate result reusing mechanism is not efficient enough as many I/O operations are wasted. Efficient reusing of the intermediate results could potentially improve the MapReduce performance. Inspired by such fact, we propose a framework named MEMoMR (more efficient intermediate result reusing for MapReduce) by introducing a novel reusing mechanism that can substantially reduce the I/O overhead. To this end, we invent a new metadata description method and apply it in the reusing phase. We practically realize MEMoMR and evaluate its performance by implementing it in a real cluster. The experiment results show that MEMoMR can improve the system performance as high as 23.4%, comparing against Dache. Copyright © 2015 John Wiley & Sons, Ltd.

Parallel Programming Paradigms and Frameworks in Big Data Era

With Cloud Computing emerging as a promising new approach for ad-hoc parallel data processing, major companies have started to integrate frameworks for parallel data processing in their product portfolio, making it easy for customers to access these services and to deploy their programs. We have entered the Era of Big Data. The explosion and profusion of available data in a wide range of application domains rise up new challenges and opportunities in a plethora of disciplines--ranging from science and engineering to biology and business. One major challenge is how to take advantage of the unprecedented scale of data--typically of heterogeneous nature--in order to acquire further insights and knowledge for improving the quality of the offered services. To exploit this new resource, we need to scale up and scale out both our infrastructures and standard techniques. Our society is already data-rich, but the question remains whether or not we have the conceptual tools to handle it. In this paper we discuss and analyze opportunities and challenges for efficient parallel data processing. Big Data is the next frontier for innovation, competition, and productivity, and many solutions continue to appear, partly supported by the considerable enthusiasm around the MapReduce paradigm for large-scale data analysis. We review various parallel and distributed programming paradigms, analyzing how they fit into the Big Data era, and present modern emerging paradigms and frameworks. To better support practitioners interesting in this domain, we end with an analysis of on-going research challenges towards the truly fourth generation data-intensive science.

A Survey of Parallel Programming Models and Tools in the Multi and Many-Core Era

In this work, we present a survey of the different parallel programming models and tools available today with special consideration to their suitability for high-performance computing. Thus, we review the shared and distributed memory approaches, as well as the current heterogeneous parallel programming model. In addition, we analyze how the partitioned global address space (PGAS) and hybrid parallel programming models are used to combine the advantages of shared and distributed memory systems. The work is completed by considering languages with specific parallel support and the distributed programming paradigm. In all cases, we present characteristics, strengths, and weaknesses. The study shows that the availability of multi-core CPUs has given new impulse to the shared memory parallel programming approach. In addition, we find that hybrid parallel programming is the current way of harnessing the capabilities of computer clusters with multi-core nodes. On the other hand, heterogeneous programming is found to be an increasingly popular paradigm, as a consequence of the availability of multi-core CPUs+GPUs systems. The use of open industry standards like OpenMP, MPI, or OpenCL, as opposed to proprietary solutions, seems to be the way to uniformize and extend the use of parallel programming models.

Aspect-Oriented Formal Techniques of Cyber Physical Systems

Cyber-physical systems pose considerable technical challenges, ranging from the distributed programming paradigms to networking protocols with timeliness as a structuring concern, including systems theory that combines physical concerns and computational concerns. Formal specification techniques for such systems have to be able to describe all these concerns. Unfortunately, a single specification technique that is well suited for all these concerns s is yet not available. Instead one finds various specialized techniques that are very good at describing individual concerns of cyber-physical system. This observation has led to research into the combination and semantic integration of specification techniques. This paper proposes a formwork for specifying cyber physical systems based on aspect-oriented formal method, which exploits the diversity and power of existing formal specification languages. There is no requirement that different aspects of a system should be expressed in the same language. So the different aspects can be specified by one formal specification technique or different formal specification techniques. The proposed aspect–oriented formal framework is such a formwork. On the one hand, it can deal with continuous-time systems based on sets of ordinary differential equations. On the other hand, it can deal with discrete-event systems, without continuous variables or differential equations. We present a combination of the formal methods Timed-CSP, ZimOO and differential (algebraic) equations or differential logic. Each method can describe certain aspects of a cyber physical system: CSP can describe communication, concurrent and real-time requirements; ZimOO expresses complex data operations; differential (algebraic) equations model the dynamics and control (DC) parts. Two case studies illustrate the specification process of aspect-oriented formal specification for cyber physical systems.

A dynamically reconfigurable architecture for smart grids

This paper introduces a proposal for a smart management platform built on the distributed programming paradigm. Besides it is based on an information model and event architecture. The information model defines a set of clear interfaces to manage every grid node. We built an actual implementation of a low-cost embeddable control/meter device that may even be attached to individual appliances. These devices behave as conventional autonomous remote distributed objects and provide full support for the information model and also for integration with the communication middleware. The resulting event architecture provides great flexibility to manage information flows from services and applications.

Implementing Virtual Agents: a HABA-Based Approach

In recent years developments in gaming technology have focused on Massively Multi-User Persistent Worlds. This virtual communities host thousands of users interacting with each other through the Internet in real time. Such communities require new distributed programming paradigms for implementing the virtual life management. In this work we present a practical methodology to add Virtual Agents to Massively Multi-User Persistent Worlds. To demonstrate the use of our methodology we present a realworld application. It adds a set of virtual agents to a 3D chat. The agents survey the users to obtain data to evaluate the processes and components of the virtual world.

Extending software component models with the master–worker paradigm

Recent advances in computing and networking technologies—such as multi-core processors and high bandwidth wide area networks—lead parallel infrastructures to reach a higher degree of complexity. Programmers have to face with both parallel and distributed programming paradigms when designing an application. This is especially true when dealing with e-Science applications. Moreover, as parallel processing is moving to the mainstream, it does not seem appropriate to rely on low-level solutions requiring expert knowledge. This paper studies how to combine modern programming practices such as those based on software components and one of the most important parallel programming paradigms which is the well-known master–worker paradigm. The goal is to provide a simple and resource transparent model while enabling an efficient utilization of resources. The paper proposes a generic approach to embed the master–worker paradigm into software component models and describes how this generic approach can be implemented within an existing software component model. The overall approach is validated with synthetic experiments on clusters and the Grid’5000 testbed.

Implementing Virtual Agents: a HABA-Based Approach

In recent years developments in gaming technology have focused on Massively Multi-User Persistent Worlds. This virtual communities host thousands of users interacting with each other through the Internet in real time. Such communities require new distributed programming paradigms for implementing the virtual life management. In this work we present a practical methodology to add Virtual Agents to Massively Multi-User Persistent Worlds. To demonstrate the use of our methodology we present a realworld application. It adds a set of virtual agents to a 3D chat. The agents survey the users to obtain data to evaluate the processes and components of the virtual world.

Protecting mobile code in the wild

Proposed applications for mobile code systems include autonomous shopping agents, autonomic systems, distributed sensor network applications, and interstellar space networks. I argue the case for mobile code systems as the next-generation distributed programming paradigm and discuss the security problems that must be addressed before this vision is practically realizable. The security discussion will focus on protecting mobile code programs that execute in the wild from malicious actions by remote hosts.

The JavaSeal Mobile Agent Kernel

Mobile agents show promise as a new distributed programming paradigm in which locality plays a central role—programs that are able to move closer to their data can overcome limitations of connectivity, latency or bandwidth. Mobility also enables distributed systems to evolve; for instance, the deployment of a new service over a network can be programmed as part of the service itself. Of course, moving programs introduces new challenges. One of these is related to program structure: How much of a computation should be moved? Where are the boundaries between mobile and immobile entities drawn? A second challenge is to provide security guarantees: How can the actions of mobile agent be controlled? And what kinds of security properties can we realistically expect to enforce? We answer these questions within the framework of the JavaSeal mobile agent system kernel. JavaSeal provides several abstractions for constructing agent systems in Java. Our basic building block is the seal which is a nested encapsulated computation fragment with sharply delineated boundaries. Strands are sequential threads of computation bound to a seal. Capsules transfer passive seals and objects over communication channels; Traffic over channels is regulated by portals. We argue that these abstractions are sufficient to program secure mobile agent systems. An electronic commerce application built over our kernel is used as a demonstrator.

A Robust Distributed Adaptive Finite Element Program for Coupled Fluid-Structure Problems

This paper describes an implementation of an adaptive finite element program for coupled fluid-structure problems using a network of workstations. A pool of task programming paradigm suitable for a heterogeneous distributed workstation environment is presented. The issues of load balancing and fault recovery are explored. Numerical results for this distributed programming paradigm are presented and compared with sequential and parallel programming models.

A client/server control architecture for robot navigation

We address the problem of task decomposition in a mobile robot navigation system. A typical robot navigation algorithm consists of a number of concurrent modules. In order to accomplish the common goal of the navigation task, they share resources on the robot. In such a system resource access control and information sharing must be properly managed. We approach these two issues by using a client/server distributed programming paradigm. Two types of server are defined: data server and hardware server. An indoor navigation system developed using the client/server model described above is presented.

Distributed object-based simulation in WAVE

A distributed programming paradigm and language, WAVE, is described which allows flexible implementations of multi-user, dynamically programmed, distributed systems. The language acts in a spatial pattern-matching, logic flow mode, without any centralised control. One application of the WAVE technique is in the field of distributed object-based simulation. It can be used to construct multi-user simulations in which the population and distribution of models may be altered freely and dynamically, under control of strands of logic self-navigating through a processing network. Suitable choice of network topology can ensure that the total processing load scales linearly with simulation size, paving the way to simulations with no effective size limit.