A novel FPGA-based evolvable hardware system based on multiple processing arrays

In this paper, an architecture based on a scalable and flexible set of Evolvable Processing arrays is presented. FPGA-native Dynamic Partial Reconfiguration (DPR) is used for evolution, which is done intrinsically, letting the system to adapt autonomously to variable run-time conditions, including the presence of transient and permanent faults. The architecture supports different modes of operation, namely: independent, parallel, cascaded or bypass mode. These modes of operation can be used during evolution time or during normal operation. The evolvability of the architecture is combined with fault-tolerance techniques, to enhance the platform with self-healing features, making it suitable for applications which require both high adaptability and reliability. Experimental results show that such a system may benefit from accelerated evolution times, increased performance and improved dependability, mainly by increasing fault tolerance for transient and permanent faults, as well as providing some fault identification possibilities. The evolvable HW array shown is tailored for window-based image processing applications.

Marco computacional para la definición de un modelo bio-inspirado en la señalización celular

Las Redes de Procesadores Evolutivos-NEP propuestas en [Mitrana et al., 2001], son un modelo computacional bio-inspirado a partir de la evolución de poblaciones de células, definiendo a nivel sintáctico algunas propiedades biológicas. En este modelo, las células están representadas por medio de palabras que describen secuencias de ADN. Informalmente, en algún instante de tiempo, el sistema evolutivo está representado por una colección de palabras cada una de las cuales representa una célula. El espacio genotipo de las especies, es un conjunto que recoge aquellas palabras que son aceptadas como sobrevivientes (es decir, como \correctas"). Desde el punto de vista de la evolución, las células pertenecen a especies y su comunidad evoluciona de acuerdo a procesos biológicos como la mutación y la división celular. éstos procesos representan el proceso natural de evolución y ponen de manifiesto una característica intrínseca de la naturaleza: el paralelismo. En este modelo, estos procesos son vistos como operaciones sobre palabras. Formalmente, el modelo de las NEP constituyen una arquitectura paralela y distribuida de procesamiento simbólico inspirada en la Máquina de conexión [Hillis, 1981], en el Paradigma de Flujo Lógico [Errico and Jesshope, 1994] y en las Redes de Procesadores Paralelos de Lenguajes (RPPL) [Csuhaj-Varju and Salomaa, 1997]. Al modelo NEP se han ido agregando nuevas y novedosas extensiones hasta el punto que actualmente podemos hablar de una familia de Redes de Procesadores Bio-inspirados (NBP) [Mitrana et al., 2012b]. Un considerable número de trabajos a lo largo de los últimos años han demostrado la potencia computacional de la familia NBP. En general, éstos modelos son computacionalmente completos, universales y eficientes [Manea et al., 2007], [Manea et al., 2010b], [Mitrana and Martín-Vide, 2005]. De acuerdo a lo anterior, se puede afirmar que el modelo NEP ha adquirido hasta el momento un nivel de madurez considerable. Sin embargo, aunque el modelo es de inspiración biológica, sus metas siguen estando motivadas en la Teoría de Lenguajes Formales y las Ciencias de la Computación. En este sentido, los aspectos biológicos han sido abordados desde una perspectiva cualitativa y el acercamiento a la realidad biológica es de forma meramente sintáctica. Para considerar estos aspectos y lograr dicho acercamiento es necesario que el modelo NEP tenga una perspectiva más amplia que incorpore la interacción de aspectos tanto cualitativos como cuantitativos. La contribución de esta Tesis puede considerarse como un paso hacia adelante en una nueva etapa de los NEPs, donde el carácter cuantitativo del modelo es de primordial interés y donde existen posibilidades de un cambio visible en el enfoque de interés del dominio de los problemas a considerar: de las ciencias de la computación hacia la simulación/modelado biológico y viceversa, entre otros. El marco computacional que proponemos en esta Tesis extiende el modelo de las Redes de Procesadores Evolutivos (NEP) y define arquitectura inspirada en la definición de bloques funcionales del proceso de señalización celular para la solución de problemas computacionales complejos y el modelado de fenómenos celulares desde una perspectiva discreta. En particular, se proponen dos extensiones: (1) los Transductores basados en Redes de Procesadores Evolutivos (NEPT), y (2) las Redes Parametrizadas de Procesadores Evolutivos Polarizados (PNPEP). La conservación de las propiedades y el poder computacional tanto de NEPT como de PNPEP se demuestra formalmente. Varias simulaciones de procesos relacionados con la señalización celular son abordadas sintáctica y computacionalmente, con el _n de mostrar la aplicabilidad e idoneidad de estas dos extensiones. ABSTRACT Network of Evolutionary Processors -NEP was proposed in [Mitrana et al., 2001], as a computational model inspired by the evolution of cell populations, which might model some properties of evolving cell communities at the syntactical level. In this model, cells are represented by words which encode their DNA sequences. Informally, at any moment of time, the evolutionary system is described by a collection of words, where each word represents one cell. Cells belong to species and their community evolves according to mutations and division which are defined by operations on words. Only those cells accepted as survivors (correct) are represented by a word in a given set of words, called the genotype space of the species. This feature is analogous with the natural process of evolution. Formally, NEP is based on an architecture for parallel and distributed processing inspired from the Connection Machine [Hillis, 1981], the Flow Logic Paradigm [Errico and Jesshope, 1994] and the Networks of Parallel Language Processors (RPPL) [Csuhaj-Varju and Salomaa, 1997]. Since the date when NEP was proposed, several extensions and variants have appeared engendering a new set of models named Networks of Bio-inspired Processors (NBP) [Mitrana et al., 2012b]. During this time, several works have proved the computational power of NBP. Specifically, their efficiency, universality, and computational completeness have been thoroughly investigated [Manea et al., 2007, Manea et al., 2010b, Mitrana and Martín-Vide, 2005]. Therefore, we can say that the NEP model has reached its maturity. Nevertheless, although the NEP model is biologically inspired, this model is mainly motivated by mathematical and computer science goals. In this context, the biological aspects are only considered from a qualitative and syntactical perspective. In view of this lack, it is important to try to keep the NEP theory as close as possible to the biological reality, extending their perspective incorporating the interplay of qualitative and quantitative aspects. The contribution of this Thesis, can be considered as a starting point in a new era of the NEP model. Then, the quantitative character of the NEP model is mandatory and it can address completely new different types of problems with respect to the classical computational domain (e.g. from the computer science to system biology). Therefore, the computational framework that we propose extends the NEP model and defines an architecture inspired by the functional blocks from cellular signaling in order to solve complex computational problems and cellular phenomena modeled from a discrete perspective. Particularly, we propose two extensions, namely: (1) Transducers based on Network of Evolutionary Processors (NEPT), and (2) Parametrized Network of Polarized Evolutionary Processors (PNPEP). Additionally, we have formally proved that the properties and computational power of NEP is kept in both extensions. Several simulations about processes related with cellular signaling both syntactical and computationally have been considered to show the model suitability.

Desarrollo de una herramienta de minitorización y control de aplicaciones distribuidas en un entorno de integración continua

Debido a la creciente relevancia de la computación en la nube y de los sistemas distribuidos, cobran también creciente interés las herramientas que ayudan a los desarrolladores y administradores a desempeñar sus funciones con la mayor eficacia posible. Por ello el objetivo principal de este trabajo es el desarrollo de una herramienta capaz de crear y controlar un entorno de almacenamiento de claves distribuidas desde una máquina local e independiente, aumentando la productividad mediante la automatización de todas las tareas. La herramienta desarrollada tiene la capacidad necesaria para integrarse tanto en proyectos que se encuentren en marcha como para proyectos que aún no hayan comenzado y proveer una solución sencilla, eficaz, y, sobre todo, útil. A lo largo del trabajo se ha realizado una gran tarea de análisis para determinar cuáles serán, de entre las posibilidades existentes, las más apropiadas para su implementación, teniendo en cuenta las tecnologías líderes disponibles en el estado del arte. Ello ha requerido también la obtención de una mejor comprensión de su funcionamiento interno. Se han realizado diferentes diseños que se han analizado y discutido en detalle para encontrar la solución que mejor se adaptaba a los objetivos propuestos. Y finalmente se ha desarrollado una herramienta ligera y sencilla, pero con un gran potencial para la administración. ---ABSTRACT---Due to the growing relevance of cloud computing and distributed systems it seems interesting to take into account the importance of the administration tools that help developers and administrators fulfill their duties in the most efficient ways. Because of this motivation, the main objective of this project is the development of a tool capable of creating and controlling a distributed key storing environment from a local and independent machine, improving the productivity thanks to the automation of all the involved tasks. The developed tool is able to integrate itself into already running projects as well as in not-yet-started ones, providing a simple, efficient and overall useful solution. During this project big tasks of research and analysis have taken place in order to determine, from the existent possibilities, the most suitable for its implementation, taking into account the leading technologies in the sector, which are described in the state of the art section. This has required the acquisition of a better insight of their inner workings. Some different designs have been made and have been discussed in detail with the intention of finding the solution that best suits the proposed objectives. And finally a lightweight and simple tool has been developed, which presents a very big potential for administration tasks.

Modelo de arquitectura para gestión cooperativa de sistemas y servicios distribuidos basado en agentes autónomos

La creciente complejidad, heterogeneidad y dinamismo inherente a las redes de telecomunicaciones, los sistemas distribuidos y los servicios avanzados de información y comunicación emergentes, así como el incremento de su criticidad e importancia estratégica, requieren la adopción de tecnologías cada vez más sofisticadas para su gestión, su coordinación y su integración por parte de los operadores de red, los proveedores de servicio y las empresas, como usuarios finales de los mismos, con el fin de garantizar niveles adecuados de funcionalidad, rendimiento y fiabilidad. Las estrategias de gestión adoptadas tradicionalmente adolecen de seguir modelos excesivamente estáticos y centralizados, con un elevado componente de supervisión y difícilmente escalables. La acuciante necesidad por flexibilizar esta gestión y hacerla a la vez más escalable y robusta, ha provocado en los últimos años un considerable interés por desarrollar nuevos paradigmas basados en modelos jerárquicos y distribuidos, como evolución natural de los primeros modelos jerárquicos débilmente distribuidos que sucedieron al paradigma centralizado. Se crean así nuevos modelos como son los basados en Gestión por Delegación, en el paradigma de código móvil, en las tecnologías de objetos distribuidos y en los servicios web. Estas alternativas se han mostrado enormemente robustas, flexibles y escalables frente a las estrategias tradicionales de gestión, pero continúan sin resolver aún muchos problemas. Las líneas actuales de investigación parten del hecho de que muchos problemas de robustez, escalabilidad y flexibilidad continúan sin ser resueltos por el paradigma jerárquico-distribuido, y abogan por la migración hacia un paradigma cooperativo fuertemente distribuido. Estas líneas tienen su germen en la Inteligencia Artificial Distribuida (DAI) y, más concretamente, en el paradigma de agentes autónomos y en los Sistemas Multi-agente (MAS). Todas ellas se perfilan en torno a un conjunto de objetivos que pueden resumirse en alcanzar un mayor grado de autonomía en la funcionalidad de la gestión y una mayor capacidad de autoconfiguración que resuelva los problemas de escalabilidad y la necesidad de supervisión presentes en los sistemas actuales, evolucionar hacia técnicas de control fuertemente distribuido y cooperativo guiado por la meta y dotar de una mayor riqueza semántica a los modelos de información. Cada vez más investigadores están empezando a utilizar agentes para la gestión de redes y sistemas distribuidos. Sin embargo, los límites establecidos en sus trabajos entre agentes móviles (que siguen el paradigma de código móvil) y agentes autónomos (que realmente siguen el paradigma cooperativo) resultan difusos. Muchos de estos trabajos se centran en la utilización de agentes móviles, lo cual, al igual que ocurría con las técnicas de código móvil comentadas anteriormente, les permite dotar de un mayor componente dinámico al concepto tradicional de Gestión por Delegación. Con ello se consigue flexibilizar la gestión, distribuir la lógica de gestión cerca de los datos y distribuir el control. Sin embargo se permanece en el paradigma jerárquico distribuido. Si bien continúa sin definirse aún una arquitectura de gestión fiel al paradigma cooperativo fuertemente distribuido, estas líneas de investigación han puesto de manifiesto serios problemas de adecuación en los modelos de información, comunicación y organizativo de las arquitecturas de gestión existentes. En este contexto, la tesis presenta un modelo de arquitectura para gestión holónica de sistemas y servicios distribuidos mediante sociedades de agentes autónomos, cuyos objetivos fundamentales son el incremento del grado de automatización asociado a las tareas de gestión, el aumento de la escalabilidad de las soluciones de gestión, soporte para delegación tanto por dominios como por macro-tareas, y un alto grado de interoperabilidad en entornos abiertos. A partir de estos objetivos se ha desarrollado un modelo de información formal de tipo semántico, basado en lógica descriptiva que permite un mayor grado de automatización en la gestión en base a la utilización de agentes autónomos racionales, capaces de razonar, inferir e integrar de forma dinámica conocimiento y servicios conceptualizados mediante el modelo CIM y formalizados a nivel semántico mediante lógica descriptiva. El modelo de información incluye además un “mapping” a nivel de meta-modelo de CIM al lenguaje de especificación de ontologías OWL, que supone un significativo avance en el área de la representación y el intercambio basado en XML de modelos y meta-información. A nivel de interacción, el modelo aporta un lenguaje de especificación formal de conversaciones entre agentes basado en la teoría de actos ilocucionales y aporta una semántica operacional para dicho lenguaje que facilita la labor de verificación de propiedades formales asociadas al protocolo de interacción. Se ha desarrollado también un modelo de organización holónico y orientado a roles cuyas principales características están alineadas con las demandadas por los servicios distribuidos emergentes e incluyen la ausencia de control central, capacidades de reestructuración dinámica, capacidades de cooperación, y facilidades de adaptación a diferentes culturas organizativas. El modelo incluye un submodelo normativo adecuado al carácter autónomo de los holones de gestión y basado en las lógicas modales deontológica y de acción.---ABSTRACT---The growing complexity, heterogeneity and dynamism inherent in telecommunications networks, distributed systems and the emerging advanced information and communication services, as well as their increased criticality and strategic importance, calls for the adoption of increasingly more sophisticated technologies for their management, coordination and integration by network operators, service providers and end-user companies to assure adequate levels of functionality, performance and reliability. The management strategies adopted traditionally follow models that are too static and centralised, have a high supervision component and are difficult to scale. The pressing need to flexibilise management and, at the same time, make it more scalable and robust recently led to a lot of interest in developing new paradigms based on hierarchical and distributed models, as a natural evolution from the first weakly distributed hierarchical models that succeeded the centralised paradigm. Thus new models based on management by delegation, the mobile code paradigm, distributed objects and web services came into being. These alternatives have turned out to be enormously robust, flexible and scalable as compared with the traditional management strategies. However, many problems still remain to be solved. Current research lines assume that the distributed hierarchical paradigm has as yet failed to solve many of the problems related to robustness, scalability and flexibility and advocate migration towards a strongly distributed cooperative paradigm. These lines of research were spawned by Distributed Artificial Intelligence (DAI) and, specifically, the autonomous agent paradigm and Multi-Agent Systems (MAS). They all revolve around a series of objectives, which can be summarised as achieving greater management functionality autonomy and a greater self-configuration capability, which solves the problems of scalability and the need for supervision that plague current systems, evolving towards strongly distributed and goal-driven cooperative control techniques and semantically enhancing information models. More and more researchers are starting to use agents for network and distributed systems management. However, the boundaries established in their work between mobile agents (that follow the mobile code paradigm) and autonomous agents (that really follow the cooperative paradigm) are fuzzy. Many of these approximations focus on the use of mobile agents, which, as was the case with the above-mentioned mobile code techniques, means that they can inject more dynamism into the traditional concept of management by delegation. Accordingly, they are able to flexibilise management, distribute management logic about data and distribute control. However, they remain within the distributed hierarchical paradigm. While a management architecture faithful to the strongly distributed cooperative paradigm has yet to be defined, these lines of research have revealed that the information, communication and organisation models of existing management architectures are far from adequate. In this context, this dissertation presents an architectural model for the holonic management of distributed systems and services through autonomous agent societies. The main objectives of this model are to raise the level of management task automation, increase the scalability of management solutions, provide support for delegation by both domains and macro-tasks and achieve a high level of interoperability in open environments. Bearing in mind these objectives, a descriptive logic-based formal semantic information model has been developed, which increases management automation by using rational autonomous agents capable of reasoning, inferring and dynamically integrating knowledge and services conceptualised by means of the CIM model and formalised at the semantic level by means of descriptive logic. The information model also includes a mapping, at the CIM metamodel level, to the OWL ontology specification language, which amounts to a significant advance in the field of XML-based model and metainformation representation and exchange. At the interaction level, the model introduces a formal specification language (ACSL) of conversations between agents based on speech act theory and contributes an operational semantics for this language that eases the task of verifying formal properties associated with the interaction protocol. A role-oriented holonic organisational model has also been developed, whose main features meet the requirements demanded by emerging distributed services, including no centralised control, dynamic restructuring capabilities, cooperative skills and facilities for adaptation to different organisational cultures. The model includes a normative submodel adapted to management holon autonomy and based on the deontic and action modal logics.

GPGPU implementation of growing neural gas: application to 3D scene reconstruction

Self-organising neural models have the ability to provide a good representation of the input space. In particular the Growing Neural Gas (GNG) is a suitable model because of its flexibility, rapid adaptation and excellent quality of representation. However, this type of learning is time-consuming, especially for high-dimensional input data. Since real applications often work under time constraints, it is necessary to adapt the learning process in order to complete it in a predefined time. This paper proposes a Graphics Processing Unit (GPU) parallel implementation of the GNG with Compute Unified Device Architecture (CUDA). In contrast to existing algorithms, the proposed GPU implementation allows the acceleration of the learning process keeping a good quality of representation. Comparative experiments using iterative, parallel and hybrid implementations are carried out to demonstrate the effectiveness of CUDA implementation. The results show that GNG learning with the proposed implementation achieves a speed-up of 6× compared with the single-threaded CPU implementation. GPU implementation has also been applied to a real application with time constraints: acceleration of 3D scene reconstruction for egomotion, in order to validate the proposal.

Large scale data processing using MapReduce

As massive data sets become increasingly available, people are facing the problem of how to effectively process and understand these data. Traditional sequential computing models are giving way to parallel and distributed computing models, such as MapReduce, both due to the large size of the data sets and their high dimensionality. This dissertation, as in the same direction of other researches that are based on MapReduce, tries to develop effective techniques and applications using MapReduce that can help people solve large-scale problems. Three different problems are tackled in the dissertation. The first one deals with processing terabytes of raster data in a spatial data management system. Aerial imagery files are broken into tiles to enable data parallel computation. The second and third problems deal with dimension reduction techniques that can be used to handle data sets of high dimensionality. Three variants of the nonnegative matrix factorization technique are scaled up to factorize matrices of dimensions in the order of millions in MapReduce based on different matrix multiplication implementations. Two algorithms, which compute CANDECOMP/PARAFAC and Tucker tensor decompositions respectively, are parallelized in MapReduce based on carefully partitioning the data and arranging the computation to maximize data locality and parallelism.

Multi-Tenant Virtual GPUs for Optimising Performance of a Financial Risk Application

Graphics Processing Units (GPUs) are becoming popular accelerators in modern High-Performance Computing (HPC) clusters. Installing GPUs on each node of the cluster is not efficient resulting in high costs and power consumption as well as underutilisation of the accelerator. The research reported in this paper is motivated towards the use of few physical GPUs by providing cluster nodes access to remote GPUs on-demand for a financial risk application. We hypothesise that sharing GPUs between several nodes, referred to as multi-tenancy, reduces the execution time and energy consumed by an application. Two data transfer modes between the CPU and the GPUs, namely concurrent and sequential, are explored. The key result from the experiments is that multi-tenancy with few physical GPUs using sequential data transfers lowers the execution time and the energy consumed, thereby improving the overall performance of the application.

Computer aided parallelisation of unstructured mesh codes

Abstract not available

Task partitioning and priority assignment for distributed hard real-time systems

In this paper, we propose the Distributed using Optimal Priority Assignment (DOPA) heuristic that finds a feasible partitioning and priority assignment for distributed applications based on the linear transactional model. DOPA partitions the tasks and messages in the distributed system, and makes use of the Optimal Priority Assignment (OPA) algorithm known as Audsley’s algorithm, to find the priorities for that partition. The experimental results show how the use of the OPA algorithm increases in average the number of schedulable tasks and messages in a distributed system when compared to the use of Deadline Monotonic (DM) usually favoured in other works. Afterwards, we extend these results to the assignment of Parallel/Distributed applications and present a second heuristic named Parallel-DOPA (P-DOPA). In that case, we show how the partitioning process can be simplified by using the Distributed Stretch Transformation (DST), a parallel transaction transformation algorithm introduced in [1].

Parallel real-time support for distributed adaptive embedded applications

Real-time embedded applications require to process large amounts of data within small time windows. Parallelize and distribute workloads adaptively is suitable solution for computational demanding applications. The purpose of the Parallel Real-Time Framework for distributed adaptive embedded systems is to guarantee local and distributed processing of real-time applications. This work identifies some promising research directions for parallel/distributed real-time embedded applications.

Towards a flexible and dynamic replication control for distributed real-time embedded systems with QoS interdependencies

Replication is a proven concept for increasing the availability of distributed systems. However, actively replicating every software component in distributed embedded systems may not be a feasible approach. Not only the available resources are often limited, but also the imposed overhead could significantly degrade the system's performance. The paper proposes heuristics to dynamically determine which components to replicate based on their significance to the system as a whole, its consequent number of passive replicas, and where to place those replicas in the network. The results show that the proposed heuristics achieve a reasonably higher system's availability than static offline decisions when lower replication ratios are imposed due to resource or cost limitations. The paper introduces a novel approach to coordinate the activation of passive replicas in interdependent distributed environments. The proposed distributed coordination model reduces the complexity of the needed interactions among nodes and is faster to converge to a globally acceptable solution than a traditional centralised approach.

Flexible and dynamic replication control for interdependent distributed real-time embedded systems

Replication is a proven concept for increasing the availability of distributed systems. However, actively replicating every software component in distributed embedded systems may not be a feasible approach. Not only the available resources are often limited, but also the imposed overhead could significantly degrade the system’s performance. This paper proposes heuristics to dynamically determine which components to replicate based on their significance to the system as a whole, its consequent number of passive replicas, and where to place those replicas in the network. The activation of passive replicas is coordinated through a fast convergence protocol that reduces the complexity of the needed interactions among nodes until a new collective global service solution is determined.

Distributed Linear Precoding and User Selection in Coordinated Multicell Systems

In this manuscript we tackle the problem of semidistributed user selection with distributed linear precoding for sum rate maximization in multiuser multicell systems. A set of adjacent base stations (BS) form a cluster in order to perform coordinated transmission to cell-edge users, and coordination is carried out through a central processing unit (CU). However, the message exchange between BSs and the CU is limited to scheduling control signaling and no user data or channel state information (CSI) exchange is allowed. In the considered multicell coordinated approach, each BS has its own set of cell-edge users and transmits only to one intended user while interference to non-intended users at other BSs is suppressed by signal steering (precoding). We use two distributed linear precoding schemes, Distributed Zero Forcing (DZF) and Distributed Virtual Signalto-Interference-plus-Noise Ratio (DVSINR). Considering multiple users per cell and the backhaul limitations, the BSs rely on local CSI to solve the user selection problem. First we investigate how the signal-to-noise-ratio (SNR) regime and the number of antennas at the BSs impact the effective channel gain (the magnitude of the channels after precoding) and its relationship with multiuser diversity. Considering that user selection must be based on the type of implemented precoding, we develop metrics of compatibility (estimations of the effective channel gains) that can be computed from local CSI at each BS and reported to the CU for scheduling decisions. Based on such metrics, we design user selection algorithms that can find a set of users that potentially maximizes the sum rate. Numerical results show the effectiveness of the proposed metrics and algorithms for different configurations of users and antennas at the base stations.

Allocation of Parallel Real-Time Tasks in Distributed Multi-core Architectures supported by an FTT-SE Network

Distributed real-time systems such as automotive applications are becoming larger and more complex, thus, requiring the use of more powerful hardware and software architectures. Furthermore, those distributed applications commonly have stringent real-time constraints. This implies that such applications would gain in flexibility if they were parallelized and distributed over the system. In this paper, we consider the problem of allocating fixed-priority fork-join Parallel/Distributed real-time tasks onto distributed multi-core nodes connected through a Flexible Time Triggered Switched Ethernet network. We analyze the system requirements and present a set of formulations based on a constraint programming approach. Constraint programming allows us to express the relations between variables in the form of constraints. Our approach is guaranteed to find a feasible solution, if one exists, in contrast to other approaches based on heuristics. Furthermore, approaches based on constraint programming have shown to obtain solutions for these type of formulations in reasonable time.

Distributed document migration between ERP systems by means of messaging and event sourcing

This bachelor’s thesis, written for Lappeenranta University of Technology and implemented in a medium-sized enterprise (SME), examines a distributed document migration system. The system was created to migrate a large number of electronic documents, along with their metadata, from one document management system to another, so as to enable a rapid switchover of an enterprise resource planning systems inside the company. The paper examines, through theoretical analysis, messaging as a possible enabler of distributing applications and how it naturally fits an event based model, whereby system transitions and states are expressed through recorded behaviours. This is put into practice by analysing the implemented migration systems and how the core components, MassTransit, RabbitMQ and MongoDB, were orchestrated together to realize such a system. As a result, the paper presents an architecture for a scalable and distributed system that could migrate hundreds of thousands of documents over weekend, serving its goals in enabling a rapid system switchover.