Publication: FINCoS++: Marco de análisis de rendimiento de motores para procesamiento de eventos complejos
Loading...
Official URL
Full text at PDC
Publication Date
2020-09
Authors
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citations
Exportar
Abstract
El procesamiento de eventos complejos (CEP) se ha convertido en uno de los campos que han emergido más rápidamente como disciplina de investigación y tendencia industrial.
La habilidad de recopilar, analizar, y reaccionar a los eventos en tiempo real se ha vuelto un componente clave de los sistemas de información. Si embargo, aún existe una carencia de información sobre el rendimiento de los motores de procesamiento de eventos complejos.
FINCoS es una herramienta que cuenta con un enfoque flexible y neutro a través del cual los usuarios pueden ejecutar rápidamente pruebas de rendimiento en una plataforma de procesamiento de eventos sin tener que configurar las rutinas de generación de carga y la conversión de eventos. A pesar de ello esta herramienta se ha quedado desactualizada con la última versión publicada en 2013, lo que hace que no se pueda utilizar con los motores CEP actuales en un área tan dinámica como es la del procesamiento de eventos complejos. Como solución a este problema en este Trabajo de Fin de Máster se ha desarrollado FINCoS++.
Las principales diferencias entre FINCoS y FINCoS++ son las siguientes: se ha actualizado y extendido el adaptador con motores CEP y se ha integrado un motor para procesamiento de eventos nuevo, nuevas funciones automatizadas para la generación de esquemas y patrones, la posibilidad de generar pruebas de rendimiento de distintos motores CEP en una o diferentes máquinas simultáneamente, y el acceso a la visualización de las estadísticas recogidas dentro de la misma aplicación.
El presente trabajo explica las definiciones esenciales para entender el procesamiento de eventos complejos, analiza los principales trabajos relacionados con comparación de rendimiento de estos motores y el progreso que ha tenido este campo, describe el desarrollo de FINCoS++ y su arquitectura, las mejoras incluidas en esta nueva versión, así como los pasos realizados para llegar a dichas mejoras. También se muestran las pruebas y resultados realizados en FINCoS++ para ilustrar su funcionamiento, concluyendo con algunas posibles ampliaciones al sistema.
Complex event processing (CEP) has become one of the fastest-growing emerging fields as a research area and industrial trend. The ability to collect, analyze, and react to events in real time has become a key component of information systems. However, there is still a lack of information on the performance of complex event processing engines. FINCoS is a framework that has flexible and neutral approach through which users can quickly run realistic performance tests on an event processing platform without having to configure the load generation routines and event conversion. Despite this, the FINCoS framework has not been updated since 2013, and therefore it cannot be used with current CEP engines in such a dynamic research area as complex event processing. As a solution to this problem, FINCoS++ has been developed in this Master’s Thesis. The main differences between FINCoS and FINCoS++ are the following: The adapter to CEP engines of FINCoS has been updated and extended for enabling additional CEP engines, and a new event processing engine has been integrated; new automated functions for schema and pattern generation have been added to this framework; it has been enabled the generation of performance tests over different CEP engines on one or more servers simultaneously; and the visualization of the collected statistics has been enabled from the main application. This thesis explains the essential definitions to understand of complex events processing, analyzes the main works related to the performance comparison of these engines and the progress has been made in this field, describes the development and architecture of FINCoS++, the improvements included in this new version, and the steps taken to reach these improvements. Tests and results carried out in FINCoS++ have been included to illustrate its operation, concluding with some possible extensions to this system.
Complex event processing (CEP) has become one of the fastest-growing emerging fields as a research area and industrial trend. The ability to collect, analyze, and react to events in real time has become a key component of information systems. However, there is still a lack of information on the performance of complex event processing engines. FINCoS is a framework that has flexible and neutral approach through which users can quickly run realistic performance tests on an event processing platform without having to configure the load generation routines and event conversion. Despite this, the FINCoS framework has not been updated since 2013, and therefore it cannot be used with current CEP engines in such a dynamic research area as complex event processing. As a solution to this problem, FINCoS++ has been developed in this Master’s Thesis. The main differences between FINCoS and FINCoS++ are the following: The adapter to CEP engines of FINCoS has been updated and extended for enabling additional CEP engines, and a new event processing engine has been integrated; new automated functions for schema and pattern generation have been added to this framework; it has been enabled the generation of performance tests over different CEP engines on one or more servers simultaneously; and the visualization of the collected statistics has been enabled from the main application. This thesis explains the essential definitions to understand of complex events processing, analyzes the main works related to the performance comparison of these engines and the progress has been made in this field, describes the development and architecture of FINCoS++, the improvements included in this new version, and the steps taken to reach these improvements. Tests and results carried out in FINCoS++ have been included to illustrate its operation, concluding with some possible extensions to this system.
Description
Trabajo de Fin de Máster en Máster en Ingeniería Informática, Facultad de Informática UCM, Departamento de, Curso 2019/2020