Dissertations / Theses on the topic 'Distributed event-based system'

The demand for parallel and distributed discrete event simulation (PDES) is rapidly growing due to the advent of middleware programs which allow multiple processes running on one or more machines to interact across networks. High Level Architecture (HLA) proposed by DoD is the standard middleware designed for distributed simulation environment. DEVS/HLA, developed in this dissertation, is a parallel and distributed modeling and simulation environment which employs a sound system theory, modeling formalism (extended DEVS) and system homomorphisms in its design. The environment includes a highly efficient message filtering scheme called quantization and is based on a risk-free PDES simulation protocol that exploits simultaneous events. In its implementation, DEVS/HLA employs hierarchical and modular object-oriented technology. To the user it presents a high level modeling paradigm and a highly reliable distributed HLA-compliant environment. This dissertation presents an analysis of quantization-based message filtering and some very promising empirical results that clarify the tradeoff between reduced message bandwidth demand and error incurred due to message reduction. The results relate bandwidth utilization and error against quantum size for federations executing on DEVS/HLA in Unix and NT networking platforms in both LAN and WAN environments. The theoretical and empirical results indicate that predictive quantization can be very scaleable due to reduced local computation demands as well as having extremely favorable communication reduction/simulation fidelity tradeoffs. How the solution extends to real-time DEVS simulation and implications for the design of real time infrastructures are topics for further research.

Globalisation has come as a double-edged blade for information technology (IT) companies; providing growth opportunities and yet posing many challenges. Software development is moving from a monolithic model to a distributed approach, where many entities and organisations are involved in the development process. Risk management an important area to deal with all the kinds of technical and social issues within companies planning and programming schedules, and this new way of working requires more attention to be paid to the temporal, socio-cultural and control aspects than before. Multinational companies like IBM have begun to consider how to address the distributed nature of its projects across the globe. With outlets across the globe, the company finds various people of different cultures, languages and ethics working on a single and bigger IT projects from different locations. Other IT companies are facing the same problems, despite there being many kinds of approaches available to handle risk management in large scale IT companies. IBM commissioned the Distributed Risk Management Process (DRiMaP) model as a suitable solution. This model focused on the collaborative and on-going control aspects, and paid attention to the need for risk managers, project managers and management to include risk management into all phases of projects and the business cycle. The authors of the DRiMaP model did not subject it to extensive testing. This research sets out to evaluate, improve and extend the model process and thereby develop a new and dynamic approach to distributed information systems development. To do this, this research compares and contrasts the model with other risk management approaches. An Evolutionary Model is developed, and this is subjected to empirical testing through a hybrid constructive research approach. A survey is used to draw out the observations of project participants, a structured interview gathered the opinions of project experts, a software tool was developed to implement the model, and SysML and Monte Carlo methods were applied to this to simulate the functioning of the model. The Evolutionary Model was found to partially address the shortcomings of the DRiMaP model, and to provide a valuable platform for the development of an enterprise risk management solution.

L’évolution technologique du web durant ces dernières années a favorisé l’arrivée d’environnements virtuels collaboratifs pour la modélisation 3D à grande échelle. Alors que la collaboration réunit dans un même espace partagé des utilisateurs distants géographiquement pour un objectif de collaboration commun, les ressources matérielles qu'ils apportent (calcul, stockage, 3D ...) avec leurs connaissances sont encore trop rarement utilisées et cela constitue un défi. Il s'agit en effet de proposer un système simple, performant et transparent pour les utilisateurs afin de permettre une collaboration efficace à la fois sur le volet computationnel mais aussi, bien entendu, sur l'aspect métier lié à la modélisation 3D sur le web. Pour rendre efficace le passage à l’échelle, de nombreux systèmes utilisent une architecture réseau dite "hybride", combinant client serveur et pair-à-pair. La réplication optimiste s'adapte bien aux propriétés de ces environnements répartis : la dynamicité des utilisateurs et leur nombre, le type de donnée traitées (3D) et leur taille. Cette thèse présente un modèle pour les systèmes d’édition collaborative en 3D sur le web. L'architecture cliente (3DEvent) permet de déporter les aspects métiers de la 3D au plus près de l’utilisateur sous la forme d’évènements. Cette architecture orientée événements repose sur le constat d’un fort besoin de traçabilité et d’historique sur les données 3D lors de l’assemblage d’un modèle. Cet aspect est porté intrinsèquement par le patron de conception event-sourcing. Ce modèle est complété par la définition d’un intergiciel en pair-à-pair. Sur ce dernier point, nous proposons d'utiliser la technologie WebRTC qui présente une API familière aux développeurs de services en infonuagique. Une évaluation portant sur deux études utilisateur concernant l’acceptance du modèle proposé a été menée dans le cadre de tâches d’assemblage de modèles 3D sur plusieurs groupes d’utilisateurs
Web technologies evolutions during last decades fostered the development of collaborative virtual environments for 3D design at large scale. Despite the fact that collaborative environments gather in a same shared space geographically distant users in a common objective, the hardware ressources of their clients (calcul, storage, graphics ...) are often underused because of the challenge it represents. It is indeed a matter of offering an easy-to-use, efficient and transparent collaborative system to the user supporting both computationnal and 3D design visualisation and business logic needs in heterogeneous web environments. To scale well, numerous systems use a network architecture called "hybrid", combining both client-server and peer-to-peer. Optimistic replication is well adapted to distributed application such as 3D collaborative envionments : the dynamicity of users and their numbers, the 3D data type used and the large amount and size of it.This document presents a model for 3D web-based collaborative editing systems. This model integrates 3DEvent, an client-based architecture allowing us to bring 3D business logic closer to the user using events. Indeed, the need of traceability and history awareness is required during 3D design especially when several experts are involved during the process. This aspect is intrinsec to event-sourcing design pattern. This architecture is completed by a peer-to-peer middleware responsible for the synchronisation and the consistency of the system. To implement it, we propose to use the recent web standard API called WebRTC, close to cloud development services know by developers. To evaluate the model, two user studies were conducted on several group of users concerning its responsiveness and the acceptance by users in the frame of cooperative assembly tasks of 3D models

The electronic components in modern automobiles build up a distributed system with so called electronic control units connected via bus systems. As more safety- and security-relevant functions are implemented in such systems, the more important fault detection becomes. A promising approach to fault detection is to build a system model from state machines and compare its predictions with properties observed in a real system. In the automobile, potential are communication characteristics between the distributed control units. Especially, the sequence of transmitted messages can be used as the basis for supervising the communication. This thesis investigates if data gathered during system tests can be used to create state-machine system models. Such an automatically created model reflects the observed normal system behavior and can potentially be used for fault detection purposes. The task can be seen as learning a state machine from a single long message sequence. Today’s automata learning algorithms are not designed for such singlemessage- sequence input data. Especially, learning without interaction between the original system and the learning algorithm is in general a NPcomplete task. Additionally, if only positive data from the normal behaving system is available, the task is further complicated. The well-known Angluin’s L∗ state-machine learning algorithm works in general independent from the type of input data. In order for this algorithm to be applicable, certain queries have to be answered. This work proposes a statistical approach to answer such queries. The implemented adapted Angluin algorithm showed the potential of automatic model building in fault detection systems and, in particular, the possibility of learning state machines from a single positive data stream.

Ever since distributed systems technology became increasingly popular in the real-time computing area about two decades ago, real-time distributed object computing technologies have been attracting more attention from researchers and engineers. While highly effective object-oriented methodologies are now widely adopted to reduce the development complexity and maintenance costs of large scale non-real-time software applications, real-time systems engineering practice has not kept pace with these system development methodologies. Indeed, real-time design techniques have not fully adopted the concepts of modular design and analysis which are the main virtues of object-oriented design technologies. As a consequence, the demand for object-oriented analysis, design, and implementation of large-scale real-time applications has been growing. To address the need for object-oriented real-time systems engineering environments we propose the Real-Time DEVS/CORBA (RTDEVS/CORBA) distributed object computing environment. In this dissertation, we show how this environment is an extension of previously developed DEVS-based modeling and simulation frameworks that have been shown to support an effective modeling and simulation methodology in various application areas. The major objective in developing Distributed Real-Time DEVS/CORBA is to establish a framework in which distributed real-time systems can be designed through DEVS-based modeling and simulation studies, and then migrated with minimal additional effort to be executed in the real-time distributed environment. This environment provides generic support for developing models of distributed embedded software systems, evaluating their performance and timing behavior through simulation and easing the transition from the simulation to actual executions. In this dissertation we describe, in some detail, the design and implementation of the RTDEVS/CORBA environment. It was implemented over Visibroker CORBA middleware along with the use of ACE/TAO real-time CORBA services, such as the real-time event service and the runtime scheduling service. Implementation aspects considered include time synchronization issues, priority-based message dispatching for timely message delivery, implementation of activity with threads, and other features required for simulating and executing real-time DEVS models. Finally, application examples are presented in the last part of the dissertation to show applicability of the environment to real systems-engineering problems.

Wireless sensor networks are application-specific networks that necessitate the development of specific network and information processing architectures that can meet the requirements of the applications involved. The most important challenge related to wireless sensor networks is the limited energy and computational resources of the battery powered sensor nodes. Although the central processing of information produces the most accurate results, it is not an energy-efficient method because it requires a continuous flow of raw sensor readings over the network. As communication operations are the most expensive in terms of energy usage, the distributed processing of information is indispensable for viable deployments of applications in wireless sensor networks. This method not only helps in reducing the total amount of packets transmitted and the total energy consumed by sensor nodes, but also produces scalable and fault-tolerant networks. Another important challenge associated with wireless sensor networks is that the possibility of sensory data being imperfect and imprecise is high. The requirement of precision necessitates employing expensive mechanisms such as redundancy or use of sophisticated equipments. Therefore, approximate computing may need to be used instead of precise computing to conserve energy. This thesis presents two schemes that distribute information processing for event-driven reactive applications, which are interested in higher-level information not in the raw sensory data of individual nodes, to appropriate nodes in sensor networks. Furthermore, based on these schemes, a fuzzy rule-based system is proposed that handles imprecision, inherently present in sensory data.

Thesis (Ph. D.)--Computing, Georgia Institute of Technology, 2008.
Committee Chair: Fujimoto, Richard; Committee Member: Ammar, Mostafa; Committee Member: Bader, David; Committee Member: Goldsman, David; Committee Member: Park, Haesun; Committee Member: Riley, George.

Les besoins croissants en puissance de calcul sont généralement satisfaits en fédérant de plus en plus d'ordinateurs (ou noeuds) pour former des infrastructures distribuées. La tendance actuelle est d'utiliser la virtualisation système dans ces infrastructures, afin de découpler les logiciels des noeuds sous-jacents en les encapsulant dans des machines virtuelles. Pour gérer efficacement ces infrastructures virtualisées, de nouveaux gestionnaires logiciels ont été mis en place. Ces gestionnaires sont pour la plupart hautement centralisés (les tâches de gestion sont effectuées par un nombre restreint de nœuds dédiés). Cela limite leur capacité à passer à l'échelle, autrement dit à gérer de manière réactive des infrastructures de grande taille, qui sont de plus en plus courantes. Au cours de cette thèse, nous nous sommes intéressés aux façons d'améliorer cet aspect ; l'une d'entre elles consiste à décentraliser le traitement des tâches de gestion, lorsque cela s'avère judicieux. Notre réflexion s'est concentrée plus particulièrement sur l'ordonnancement dynamique des machines virtuelles, pour donner naissance à la proposition DVMS (Distributed Virtual Machine Scheduler). Nous avons mis en œuvre un prototype, que nous avons validé au travers de simulations (notamment via l'outil SimGrid), et d'expériences sur le banc de test Grid'5000. Nous avons pu constater que DVMS se montrait particulièrement réactif pour gérer des infrastructures virtualisées constituées de dizaines de milliers de machines virtuelles réparties sur des milliers de nœuds. Nous nous sommes ensuite penchés sur les perspectives d'extension et d'amélioration de DVMS. L'objectif est de disposer à terme d'un gestionnaire décentralisé complet, objectif qui devrait être atteint au travers de l'initiative Discovery qui fait suite à ces travaux.

I sistemi software nati dall'esigenza di supportare agevolmente ed efficacemente il lavoro cooperativo, in particolare quelli orientati al supporto di azioni di soccorso in scenari di emergenza, appaiono tutt'ora fortemente limitati e frammentati. In molti casi vengono affrontate solamente specifiche dimensioni del problema complessivo, anche se il livello al quale è giunto lo sviluppo tecnologico e i risultati osservati in ambito di ricerca permettono di delineare soluzioni complete e significative per l'impiego in ambiti reali. Tale tipologia di sistemi è stata scelta per il grande interesse che desta sia dal punto di vista accademico, essendo costituita da molteplici sotto--sistemi spesso eterogenei che debbono necessariamente interagire e supportare l'azione umana, sia dal punto di vista industriale, interpretando la necessità crescente di iniettare nel maggior numero possibile di livelli sociali la forte dipendenza (il supporto allo stesso tempo) dalle scienze tecnologiche ed informatiche, per rafforzare e talvolta estendere le possibilità dell'essere umano in quanto tale. Dopo una prima fase in cui verrà delineato un quadro concettuale piuttosto dettagliato circa i principali elementi e problematiche che caratterizzano la classe di sistemi considerati, sarà dato spazio alla validazione di tali principi e considerazioni emerse, confrontandosi con la progettazione e sviluppo in forma prototipale di un sotto--sistema relativo ad un caso di studio reale, significativo per l'ambito di applicazione, nato dalla collaborazione con l'Università degli Studi di Bologna di un'azienda della regione Emilia--Romagna. Il sistema software realizzato vuole essere innanzi tutto la risposta alle esigenze emerse nel caso di studio trattato, in modo tale da potersi sostituire agli attuali limitati supporti alla cooperazione, ma anche un esperimento che possa essere considerato un artefatto centrale da utilizzare come base di conoscenza condivisa, in cui vengano fattorizzati i concetti e meccanismi chiave, fondamentali per sviluppi futuri.

Networked control over wireless networks is of growing importance in many application domains such as industrial control, building automation and transportation systems. Wide deployment however, requires systematic design tools to enable efficient resource usage while guaranteeing close-loop control performance. The control system may be greatly affected by the inherent imperfections and limitations of the wireless medium and malfunction of system components. In this thesis, we make five important contributions that address these issues.  In the first contribution, we consider event- and self-triggered control and investigate how to efficiently tune and execute these paradigms for appropriate control performance. Communication strategies for aperiodic control are devised, where we jointly address the selection of medium-access control and scheduling policies. Experimental results show that the best trade-off is obtained by a hybrid scheme, combining event- and self-triggered control together with contention-based and contention-free medium access control. The second contribution proposes an event-based method to select between fast and slow periodic sampling rates. The approach is based on linear quadratic control and the event condition is a quadratic function of the system state. Numerical and experimental results show that this hybrid controller is able to reduce the average sampling rate in comparison to a traditional periodic controller, while achieving the same closed-loop control performance. In the third contribution, we develop compensation methods for out-of-order communications and time-varying delays using a game-theoretic minimax control framework. We devise a linear temporal coding strategy where the sensor combines the current and previous measurements into a single packet to be transmitted. An experimental evaluation is performed in a multi-hop networked control scenario with a routing layer vulnerability exploited by a malicious application. The experimental and numerical results show the advantages of the proposed compensation schemes. The fourth contribution proposes a distributed reconfiguration method for sensor and actuator networks. We consider systems where sensors and actuators cooperate to recover from faults. Reconfiguration is performed to achieve model-matching, while minimizing the steady-state estimation error covariance and a linear quadratic control cost. The reconfiguration scheme is implemented in a room heating testbed, and experimental results demonstrate the method's ability to automatically reconfigure the faulty system in a distributed and fast manner. The final contribution is a co-simulator, which combines the control system simulator Simulink with the wireless network simulator COOJA. The co-simulator integrates physical plant dynamics with realistic wireless network models and the actual embedded software running on the networked devices. Hence, it allows for the validation of the complete wireless networked control system, including the study of the interactions between software and hardware components.

QC 20140929

Distributed Event-Based Systems (DEBSs) are middleware supporting the interaction of publisher and subscriber components via events. In DEBSs, the subscribers to be notified when an event is announced are decided at run-time without requiring publisher components to know the name or locations of the subscribers, nor the subscribers to know the name or locations of the publishers. This low coupling between components makes DEBSs suitable for applications with a large or unpredictable number of autonomous components. The development of applications in DEBSs is an ad hoc process poorly supported by current software engineering methodologies. Moreover, the behaviours exhibited by these systems and their applications are not well understood, and no suitable models exist where these behaviours can be described and analyzed. The main concern of this thesis is the development of such models. Specifically, we develop formalisms and models supporting the specification, prediction, and validation of the behaviour exhibited by the middleware and the applications executing on it. Our main contributions to the area are: new formalisms for the representation of DEBSs and their applications, and for the specification of both, system and application properties; a categorization of the features related to the definition, announcement, and notification of events in DEBSs and, in general, event-based systems; models representing the categorized DEBS features; case studies detailing models and properties for specific systems; a prototype tool for the verification of DEBSs and applications. The formalisms developed expose the location of the actions in the modelled systems and support the specification of several forms of location-awareness and adaptive behaviour.

碩士
國立交通大學
資訊科學系
89
The Internet proliferates rapidly day after day. It seems to grow up even faster after the new millennium. Although more and more data exchange systems build up with event-based distributed system (such as CORBA Event/Notification Service, Java Message Service, TIBCO TIB), but it still gets some troubles. There is a difficult technical threshold de facto. For using the event-based distributed system as an underlying communication model, these programmers should consider not only the traditional data logic but also the event-based programming model. This is not an easy job especially in today that the excellent programmers are high-priced to hire. The small enterprises do not have enough resources to build up a suitable event-based distributed system. This research proposed a system development environment, named Ghostwriter. It is an engine between client applications and underlying event-based distributed system and configures its inner settings with EventML (Event Markup Language). The existence of Ghostwriter is to cross the technical threshold of event-based distributed system, and clearly separates the duties from system analysts and application programmers. It helps to constructs systems with a simple, efficient process.

A Distributed Event-based System (DEBS) exhibits its desired behavior through its functional components collaborating with each other via event exchanging. Due to loose-coupling and flexibility, DEBS applications have become increasingly popular. Indeed, such systems are expected to appear in various application domains such as large-scale Internet applications and ubiquitous computing. Notwithstanding their popularity, current DEBS applications are still often developed in an informal process and are not modularized. On the individual event level, current DEBS developers can define what events a component can accept and publish, and, by registering event handlers, what action an event can trigger. Currently, developers lack structuring mechanisms for representing event interactions and dependencies in a modular way. While current research has made fruitful contributions to various aspects in the DEBS paradigm, such as, event delivery, event detection and composition, event visibility, its emphasis is on the individual event level. In this thesis, we advocate that by designing a new DEBS metamodel with extended behavioral interfaces and high-level structure mechanisms, we can (1) define an interface-based modular approach to model and design DEBS applications, (2) implement a prototype framework on a P2P network that provides built-in support to our proposed interface-based DEBS development, and (3) provide case studies illustrating the interface-based development process and the applicability of our proposed approach.

Τα τελευταία χρόνια οι εφαρμογές συνεχούς μετάδοσης ροών πληροφορίας στο διαδίκτυο έχουν γίνει ιδιαίτερα δημοφιλείς. Με τον συνεχώς αυξανόμενο ρυθμό εισόδου νέων αντικειμένων πληροφορίας, γίνεται ολοένα και πιο επιτακτική η ανάγκη για την ανάπτυξη πληροφορικών συστημάτων που να μπορούν να προσφέρουν στους χρήστες τους μόνο εκείνες τις πληροφορίες που τους ενδιαφέρουν, φιλτράροντας τεράστιους όγκους από άσχετες για τον κάθε χρήστη, πληροφορίες. Ένα μοντέλο διάδοσης πληροφορίας ικανό να ενσωματώσει τέτοιου είδους ιδιότητες, είναι το μοντέλο δημοσιεύσεων/συνδρομών βασισμένο στο περιεχόμενο ( content-based publish/subscribe) Βασική συνεισφορά μας στο χώρο είναι η εφαρμογή του μοντέλου δημοσιεύσεων/συνδρομών βασισμένου στο περιεχόμενο (content-based publish/subscribe) πάνω στα δίκτυα ομοτίμων ώστε να μπορέσουμε να προσφέρουμε στους χρήστες υψηλή εκφραστικότητα κατά την δήλωση των ενδιαφερόντων τους, λειτουργώντας σε ένα πλήρως κατανεμημένο και κλιμακώσιμο περιβάλλον. Ο κορμός των προτεινόμενων λύσεων σε αυτή τη διατριβή είναι: (α) η ανάπτυξη αλγορίθμων για την αποθήκευση των κλειδιών των δημοσιεύσεων σε κατάλληλους κόμβους του δικτύου με βάση τις συνθήκες στο περιεχόμενο που έχουν δηλωθεί και (β) αλγορίθμων δρομολόγησης δημοσιεύσεων στο διαδίκτυο έτσι ώστε να ((συναντούν)) αυτούς τους κόμβους οι οποίοι περιέχουν συνδρομές που ικανοποιούνται από την πληροφορία της δημοσίευσης. Οι προτεινόμενοι αλγόριθμοι υλοποιήθηκαν και εξετάσθηκαν ενδελεχώς με προσομοίωση μελετώντας την απόδοσή τους με βάση μετρικές όπως: η δίκαιη κατανομή του φόρτου στους κόμβους του δικτύου από τη διακίνηση μηνυμάτων κατά την επεξεργασία των συνδρομών/δημοσιεύσεων, ο συνολικός αριθμός μηνυμάτων που διακινούνται, ο συνολικός όγκος επιπλέον πληροφορίας που απαιτούν οι αλγόριθμοι να εισέλθει στο δίκτυο (network bandwidth), και ο χρόνος που απαιτείται για την ανεύρεση των συνδρομών που συζευγνύουν με κάθε δημοσίευση.
In the past few years the continuous data streams applications have become particularly popular. With the continuously increasing rate of entry of new information, it becomes imperative the need for developing appropriate infrastructures that will offer only the information that users are interested for, filtering out large volumes of irrelevant for each user, information. The content-based publish/subscribe model, is capable of handling large volumes of data traffic in a distributed, fully decentralized manner. Our basic contribution in this research area is the coupling of the content-based publish/subscribe model with the structured (DHT-based) peer-to-peer networks, offering high expressiveness to users on stating their interests. The proposed infrastructure operated in a distributed and scalable environment. The proposed solutions in this thesis are related to the development and testing: (a) of a number of algorithms for subscription processing in the network and (b) of a number of algorithms for processing the publication events. The proposed algorithms were developed and thoroughly tested with a detailed simulation-based experimentation. The performance metrics are: the fair distribution of load in the nodes of network from the distribution of messages while processing subscriptions and publication events, the total number of messages that are generated, the total volume of additional information that is required from the algorithms to operate, and the time that is required for matching publication events to subscriptions.