Marc Schaaf

Asynchronous messaging for OSGi

OSGi is a popular Java-based platform, which has its roots in the area of embedded systems. However, nowadays it is used more and more in enterprise systems. To fit this new application area, OSGi has recently been extended with the Remote Services specification. This specification enables distribution, which OSGi was previously lacking. However, the specification provides means for synchronous communication only and leaves out asynchronous communication. As an attempt to fill a gap in this field, we propose, implement and evaluate an approach for the integration of asynchronous messaging into OSGi.

Cloud-Based Event-Processing Architecture for Opinion Mining

The viability of cloud computing for information-intensive tasks arising in real-time opinion mining and sentiment analysis of large online text streams is described. We show how a smart distributed architecture enables an efficient and scalable design for opinion mining on internet-based content that answers key challenges, such as integrating heterogeneous data sources and adapting to events through dynamic system configuration. In particular, we present a novel approach of semantic complex event processing in a cloud environment capturing different levels of information, such as event data (e.g. content from various heterogeneous, distributed sources) as well as associations identified during the opinion mining and sentiment analysis process (e.g. dynamic co-reference resolution).

Context-aware website personalization

Today there is a need in on-line shops to support the visitors of a web page appropriately by analyzing their current situation. In this paper we introduce a model which supports the identification of the content the user is interested in and the shopping strategy of the current session. We use context information extracted from enterprise data, content data, the current and the historical behaviour of the user. This allows us to learn more about the interests and needs of the user. We monitor and analyze this content information at runtime and use it for the adaption of the web site during the users navigation.

Security of a Public Cloud

This paper is about the current state of security in cloud computing. It shows some of the key benefits and the major drawbacks that come around with swapping out services and infrastructure to a public cloud. Based on these benefits and drawbacks, K.O. (knock-out) criteria will be identified, which can be seen as the minimum basis for secure cloud environment. Then a final conclusion about security in cloud computing will be given.

Event processing with dynamically changing focus: Doctoral consortium paper

Event Processing technologies are likely to play an increasingly important role in future IT systems due to the increasing demand for on-line analytical systems as well as big data processing applications. Event Processing is particularly suitable for those applications due to its active processing approach combined with scalability. Todays approaches for achieving scalability are however focused on rather static event stream partitioning mechanisms to allow parallelization. Such approaches are well suited as long as a feasible partitioning for all processing tasks can be found. However when such a scaling mechanism is faced with processing tasks with dynamically changing focus areas, no effective pre-determined partitioning can be found which massively hampers with the required scalability. This paper presents the first steps towards a focused processing framework to overcome these limitations.

Towards a Timely Root Cause Analysis for Complex Situations in Large Scale Telecommunications Networks

Abstract Large scale telecommunications networks need to be continuously monitored to detect problems and react accordingly to ensure the networks stability. Current monitoring systems are well capable to monitor such large scale installations for simple situations like failing routers, links or abnormal link utilization. However, current systems fail to provide near real-time in-depth root cause analysis for nation scale networks. This is particularly important in cases where the monitoring dashboard is flooded with notifications caused by side-effects of one bigger failure. In our project we develop a processing system capable of providing this root cause analysis in a timely fashion even for large telecommunications networks. To cope with these challenges we defined a specialized processing model for the detection and analysis of complex error situations which is largely based on event stream processing mechanisms. In this paper we discuss the first stage of this process and its initial implementation.

Ereignisverarbeitung zur Flexiblen Dynamischen Informationsverarbeitung in Smart Cities

ZusammenfassungEin Pfeiler des Smart City Konzeptes ist die Verfügbarkeit von großen Datenmengen, sowie die Fähigkeit diese flexibel und intelligent in Nahe-Echtzeit analysieren und verarbeiten zu können. Als technologische Grundlage für diese oft unterschätze Datenanalyse und -verarbeitungsfähigkeit werden vielfach Technologien aus dem Bereich der Ereignisverarbeitung genannt. In diesem Kapitel werden zentrale Konzepte der Ereignisverarbeitung kurz vorgestellt und deren Nutzen, Grenzen und zukünftige Entwicklungen anhand eines Szenarios aus dem Bereich ad-hoc Car-Sharings beschrieben.AbstractA pillar of the Smart City concept is the availability of big data, together with the ability to process and analyse it flexibly and intelligently in near real time. Event processing technologies are often referred to as a technical basis of the—often underestimated—big data processing and analysis capability. The chapter briefly introduces its central concepts and ideas. We outline the advantages, limits, and future developments of the technology using the example of an ad-hoc car sharing scenario in a Smart City.

Enhancing OSGi with Asynchronous Messaging

OSGi is a Java-based framework intended for embedded systems. In this paper, we design, implement, evaluate and demonstrate an approach to enhancing OSGi with asynchronous messaging. Our approach is non-invasive. It guarantees reliable delivery of messages. Furthermore, our approach enables a remote asynchronous communication not only between OSGi containers but also between an OSGi container and a foreign system.

Integrating Structured Peer-to-Peer Networks Into OM4SPACE Project

In this paper, we provide an overview of the OM4SPACE project and different types of peer-to-peer (P2P) networks (both structured and unstructured), identify requirements for integrating P2P networks into the project and discuss which of the P2P networks meets these requirements best. We also propose an architecture for such integration.

Automatic Scaling of Complex-Event Processing Applications in Eucalyptus

Cloud computing is particularly well suited for applications with a variable workload such as, for example, complex-event processing (CEP) applications. Since these applications experience variability in utilization of resources, they are calling for an infrastructure that can dynamically scale according to the application demand. Scalability is one of the major advantages offered by IaaS (Infrastructure-as-a-Service) clouds. This gives rise to the idea to deploy CEP applications into IaaS clouds. But an IaaS cloud requires that CEP applications should be specifically designed for the cloud to benefit from its scalability. Thus, the problem is how to dynamically scale legacy and existing CEP applications (which have been around for years) with little or no changes to the application code.