Gero Mühl

Engineering Event-Based Systems with Scopes

Event notification services enable loose coupling and they are therefore becoming an essential part of distributed systems design. However, the development of event services follows the early stages of programming language evolution, disregarding the need for efficient mechanisms to structure event-based applications. In this paper, the wellknown notion of scopes is introduced to event-based systems. We show that limiting the visibility of events is a simple yet powerful mechanism that allows to identify application structure and offers a module construct for the loosely coupled components in event-based systems. We are able to customize the semantics of scoped event notification services by binding meta-objects to the application structure that reify important aspects of notification delivery, like interface mappings and transmission policies. The scoping concept facilitates design and implementation by offering encapsulation and adaption of syntax and semantics of eventbased systems.

Filter Similarities in Content-Based Publish/Subscribe Systems

Matching notifications to subscriptions and routing notifications from producers to interested consumers are the main problems in large-scale publish/subscribe systems.Most previously proposed distributed notification services either use flooding or, if filtering is performed, they assume that each event broker has global knowledge about all active subscriptions. Both approaches degrade the scalability of notification services as the former wastes network resources and the latter generates overly large routing tables.In this paper we describe content-based routing algorithms that exploit filter similarities in order to reduce the size of routing tables and the number of control messages that are exchanged among the brokers in order to keep the routing tables up-to-date. In particular, the proposed algorithms do not assume global knowledge about all active subscriptions. Furthermore, we describe how these optimizations can be supported if the underlying data and filter model is based on structured records.

Evaluating advanced routing algorithms for content-based publish/subscribe systems

We present an evaluation of advanced routing algorithms for content-based publish/subscribe systems that focuses on the inherent characteristics of routing algorithms (routing table sizes and filter forwarding overhead) instead of system-specific parameters (CPU load etc.). The evaluation is based on a working prototype instead of simulations and compares several routing algorithms to each other. Moreover, the effects of locality among the interests of the consumers are investigated. The results offer new insights into the behavior of content-based routing algorithms. Firstly, advanced routing algorithms can be considered mandatory in large-scale publish/subscribe systems. Secondly, the use of advertisements considerably improves scalability. Thirdly, advanced routing algorithms operate efficiently in more dynamic environments than was previously thought. Finally, the good behavior of the algorithms improves even if the interests of the consumers are not evenly distributed, which can be expected in practice.

QoS-Aware composition of web services: an evaluation of selection algorithms

A composition arranges available services resulting in a defined flow of executions. Before the composition is carried out, a discovery service identifies candidate services. Then, a selection process chooses the optimal candidates. This paper discusses how the selection can consider different Quality-of-Service (QoS) categories as selection criteria to select the most suitable candidates for the composition. If more than one category is used for optimisation, a multi-dimensional optimisation problem arises which results in an exponential computation effort for computing an optimal solution. We explain the problem and point out similarities to other combinatorial problems – the knapsack problem and the resource constraint project scheduling problem (RCPSP). Based on this discussion, we describe possible heuristics for these problems and evaluate their efficiency when used for web service candidate selection.

Generic Constraints for Content-Based Publish/Subscribe

Publish/subscribe interaction enables the loosely coupled exchange of asynchronous notifications. Matching notifications to subscriptions and routing notifications from producers to interested consumers are the main problems in large-scale publish/subscribe systems. Content-based selection provides great flexibility because it allows to categorize notifications with respect to multiple dimensions but it also requires complex routing strategies. Previous work has tightly coupled content-based routing and matching algorithms to the constraints that can be used to select notifications of interest. This paper introduces the idea of generic constraints for content-based selection that are separated from the underlying matching and routing algorithms. Therefore, new constraints can be added easily. Moreover, this paper introduces the idea of filter merging that can improve the scalability of a notification service. Finally, we present novel algorithms for matching, covering, and merging.

Modular event-based systems

Event-based systems are developed and used to integrate components in loosely coupled systems. Research and product development have focused so far on efficiency issues but neglected methodological support to build such systems. In this article, the modular design and implementation of an event system is presented which supports scopes and event mappings, two new and powerful structuring methods that facilitate engineering and coordination of components in event-based systems. We give a formal specification of scopes and event mappings within a trace-based formalism adapted from temporal logic. This is complemented by a comprehensive introduction to the event-based style, its benefits and requirements.

A modular approach to build structured event-based systems

Event-based systems are developed and used as a coordination model to integrate components in loosely coupled systems. Research and product development focused so far on efficiency issues but neglected methodological support to build such systems. In this paper, we present the modular design and implementation of an event system which supports scopes and event mappings, two new and powerful structuring methods that facilitate engineering and coordination of components in event-based systems. The approach is based on a trace-based specification method adapted from temporal logic.

'Publish-subscribe grows up: support for management, visibility control, and heterogeneity

Message-oriented middleware is used to decouple the operation of cooperating applications. Existing approaches have concentrated mainly on scalability issues, but dynamic business processes and the integration of a wide range of data sources and applications, require a middleware that is customizable. The Rebeca publish-subscribe service uses scoping to structure both middleware and applications. It thus offers advanced routing mechanisms to subsystems that need high scalability and it allows for heterogeneous message models that are transparently mapped onto each other.

Programming Abstractions for Content-Based Publish/Subscribe in Object-Oriented Languages

Asynchronous event-based communication facilitates loose coupling and eases the integration of autonomous, heterogeneous components into complex systems. Many middleware platforms for event-based communication follow the publish/subscribe paradigm. Despite the usefulness of such systems, their programming support is currently limited. Usually, publish/subscribe systems only exhibit low-level programming abstractions to application developers. In this paper we investigate programming abstractions for content-based publish/subscribe middleware in object-oriented languages, how they can be integrated in applications, and their implications on middleware implementation. We focus on the definition of filters and their implementation, the handling of notifications and meta-data, and programming support for composite events. We have implemented the presented approach for our content-based publish/subscribe middleware Rebeca.

Self-Management – Potentiale, Probleme, Perspektiven

ZUSAMMENFASSUNG Gordon Moores Gesetz vom exponentiellen Wachstum der Transistordichte pro Quadrat-Zoll hat seit 1965 die IT-Industrie geprägt. Mit der damit einhergehenden Explosion der Rechnerleistung wurde die Software immer leistungsfähiger, und man ist dazu übergegangen, Rechnersysteme zu vernetzen und Anwendungen zu verteilen. Eine Folge dieser Entwicklungen ist die rapide zunehmende Komplexität der modernen Informationstechnologie. 40 Jahre nach Moores Entdeckung droht eben diese Tatsache, dem bisherigen exponentiellen Wachstum natürliche Grenzen zu setzen. Moderne, vernetzte Rechnersysteme, wie sie in der Industrie weit verbreitet sind, sind schon heute zu komplex als dass sie auf manuellem Wege, d.h., durch menschliche Administratoren, in einem optimalen Betriebszustand gehalten werden können. Die Folgen sind eine unzureichende Ausnutzung vorhandener Ressourcen, wiederkehrende Fehlerzustände und Lücken in der Absicherung gegen mutwillige Angriffe auf die System-Integrität. Dies führt zu erheblichen finanziellen Mehraufwendungen bzw. Verlusten. Ein permanent überfordertes Administrationspersonal, trägt durch eigene Fehler ein Übriges bei. Schenkt man den jüngst aufkeimenden Initiativen von IT-Giganten wie IBM, Microsoft und Sun Glauben, so heißt die Lösung dieser Misere automatisiertes Management. Vernetzte Rechnersysteme sollen sich auf lange Sicht selbst verwalten. Man erhofft sich hiervon ein effektiveres Management und eine Freistellung von Personal, welches sich dann um wichtigere Aufgaben kümmern kann. In diesem Beitrag beleuchten wir den aktuellen Stand und die Perspektiven im Bereich des Self-Managements. Des Weiteren diskutieren wir offene Fragen, welche auf dem Weg zu selbstverwaltenden Systemen zu lösen sind.