Abstract Wireless vehicular communication is used to enhance traffic safety and to minimize congestion, thereby leading to increased driving efficiency. A malicious node can transmit an inaccurate message to trigger inevitable situations by pretending to be multiple (other) vehicles. Therefore, it is critical to identify malicious nodes as well as fake messages generated by such nodes, and discard such messages quickly. In a Sybil attack, an attacker participates in the network with multiple forged identities in order to disrupt the fundamental operations of VANET. Sybil attacks are particularly easy to launch in VANETs due to the open and broadcast nature of communication medium. In this paper, we present the implementation of simulated Sybil attack scenario in VANET and its consequences on the performance of the network. We also propose a lightweight, scalable and distributed detection approach based on the difference in movement patterns of Sybil nodes and legitimate nodes. In our approach, each Road Side Unit (RSU) computes, stores and verifies various parameter values including RSS, distance, angle of passing-by vehicles through passive overhearing process to detect Sybil attackers. The combination of different parameters makes our detection approach highly accurate. We validate our results on realistic traces obtained from a multi-agent microscopic traffic simulator (MMTS). Simulation results show the effectiveness of the proposed approach to locate Sybil nodes with a different number of network parameters.

Abstract OWL (Web Ontology Language) is the standard language for Semantic Web and is used in defining ontologies for Web. Temporal event data are ubiquitous in nature. Temporal data can be represented qualitatively using temporal relations in OWL, enabling temporal ordering of events which plays a vital role in task planners. The basic Allen’s temporal interval relations can be used to describe relations in OWL. Allen’s interval algebra is a well known formalism used to represent and reason the temporal knowledge. In this work, Allen’s interval algebra is extended by Reference Event based Temporal (REseT) relations to reduce the ambiguity in the before relation. The extended formalism is used in the representation of relations between time intervals and the viability of ordering of events in ontology is elucidated. This paper proposes a temporal knowledge representation and reasoning based event ordering system which helps in the temporal ordering of events. The advantage of this method is that it does not introduce any additional constructs in OWL and hence the existing reasoning tools and DL based query languages are capable of generating the linear order of events. The system is investigated experimentally using the COW (Correlates of War) dataset and has been evaluated using the Percent_ Similarity measure.

Abstract Performance and power consumption analysis and characterization for computational benchmarks is important for processor designers and benchmark developers. In this paper, we characterize and analyze different High Performance Computing workloads. We analyze benchmarks characteristics and behavior on various processors and propose a performance estimation analytical model to predict performance for different processor microarchitecture parameters. Performance model is verified to predict performance within <5% error margin between estimated and measured data for different processors. We also propose a power estimation analytical model to estimate power consumption with low error deviation.

Abstract Composition of High Level Petri Nets in terms of joining relevant places and/or transitions is considered in the paper. In case of place composition type safe and combined types composition is contemplated. The process of composition is proposed and analysed in three separate cases with respect to the general approaches with minimal composition interface in each case but an analogous extension of the interface follows immediately from the approach introduced. Properties of the composition mechanism are analysed, namely preserving of boundedness, liveness and deadlock freedom. Conditions for preserving of the desired properties are introduced. Usability of the compositional mechanism is analysed in the process of de/compositional analysis.

Abstract The general simulation of n-body systems often requires the simulation of pairwise interaction events between the objects. The naive method of simulating these events is an algorithm that polls each pair of objects for an interaction every time step. This algorithm has O(n 2)operations per time step in the number of objects. However, this method scales very well to multiple cores. In this paper, we propose a novel method of pairwise simulation that saves a significant amount of computation time by predicting possible future outcomes rather than reacting to them. We demonstrate the implementation of this method, as well as demonstrate that it has amortized O(n) complexity per time step. We also demonstrate an implementation to allow this algorithm to scale comparably well to multiple cores of a shared memory machine.

Abstract GPUSVM (Graphic Processing Unit Support Vector Machine) is a Computing Unified Device Architecture (CUDA) based Support Vector Machine (SVM) package. It is designed to offer an end-user a fully functional and user friendly SVM tool which utilizes the power of GPUs. The core package includes an efficient cross validation tool, a fast training tool and a predicting tool. In this article, we first introduce the background theory of how we build our parallel SVM solver using CUDA programming model. Then we compare our GPUSVM package with the popular state of the art Libsvm package on several well known datasets. The preliminary results have shown one to two orders of magnitude speed improvement in both training and predicting phases compared to Libsvm using our Tesla server.

Abstract The goal of this article is to provide a rigorous conceptual framework for defining the concepts of dependability and resilience. Since the seventies, the terms dependability and resilience have been used in nearly all the scientific and technological fields related to Information and Communication systems. The introduction and use of these concepts in all these fields makes it difficult to have a common and precise definition. Having such a definition is nevertheless mandatory for the software and systems engineering research community that create development processes, languages and tools to support the engineering of products that would be required to be dependable or resilient. For this, we introduce an abstract and generic terminology defined mathematically to be used when speaking about dependability and resiliency. We also provide some abstract semantic descriptions to these terminological elements. This formal framework is defined from a software engineering perspective, which means that we define its components such that they are useful for the development or improvement of analysis, architectural design, detailed design, implementation, verification and maintenance phases. To this aim, we provide the necessary elements in accordance with a model driven engineering perspective that enable the definition of a new modelling language for dependable and resilient systems.

Abstract With the advent of multi-core computer systems, real-time (RT) scheduling on multiprocessors (MPs) is gaining more and more practical relevance. Basic algorithms developed in the 1970s or earlier are strongly influential on state-of-the-art preemptive hard real-time scheduling algorithms. These relationships will be revealed and visualized by four genealogies. Here, the two classification dimensions prioritization dynamics and migration dynamics serve as a framework for the systematization. Finally, such genealogical reconstructions could even lead to the discovery of new algorithms. An extra look will be devoted to the group of fully dynamic scheduling algorithms with full migration in which algorithms can reach full utilization, neglecting scheduling and context switching overheads. The important question for a set of RT scheduling algorithms serving as a basis for more sophisticated ones will be dealt with.

Abstract Network analytic method designed for the analysis of static networks promise to identify significant relational patterns that correlate with important structures in the complex system the network is derived from. In this mini review, three groups of network analytic methods are discussed: centrality indices, network motifs, and clustering algorithms. We show that so far these methods have mainly been used in a descriptive way, but that they show promising possibilities to be used for prediction and classification. We thus conclude the article with a discussion of how benchmark sets and evaluation criteria could look like to realize this promise.