Supporting Real-Time Monitoring in Criminal Investigations
Being able to analyze information collected from streams of data, generated by different types of sensors, is becoming increasingly important in many domains. This paper presents an approach for creating a decoupled semantically enabled event processing s
- PDF / 128,205 Bytes
- 5 Pages / 439.37 x 666.142 pts Page_size
- 116 Downloads / 252 Views
stract. Being able to analyze information collected from streams of data, generated by different types of sensors, is becoming increasingly important in many domains. This paper presents an approach for creating a decoupled semantically enabled event processing system, which leverages existing Semantic Web technologies. By implementing the actor model, we show how we can create flexible and robust event processing systems, which can leverage different technologies in the same general workflow. We argue that in this context RSP systems can be viewed as generic systems for creating semantically enabled event processing agents. In the demonstration scenario we show how real-time monitoring can be used to support criminal intelligence analysis, and describe how the actor model can be leveraged further to support scalability. Keywords: Semantic event processing · Event processing · RDF stream processing · Actor model · Criminal intelligence
1
Introduction
Semantic Web (SW) technologies provide flexible tools for working with heterogeneous data, and Linked Data principles enable information to be shared by explicitly articulating the underlying schemas and ontologies. Traditional SW technologies have been developed to support slowly evolving (or static) data, and scale quite poorly when data is highly dynamic. In recent years, a number of RDF Stream Processing (RSP) systems have therefore been developed to support streaming Linked Data, focusing on timely execution of continuous queries over streams. Unlike most types of event processing approaches, such as Drools fusion1 and ESPER2 , RSP systems use the Linked Data principles to leverage the semantics in the streaming data. The available RSP systems, however, provide only a limited set of features out-of-the-box. For example, in the available versions of C-SPARQL [3], CQELS [5], INSTANS [8], and ETALIS/EP-SPARQL [1], streams, queries, and result listeners are closely coupled with their respective engine. This can make them difficult to use in settings where streams are not 1 2
http://www.drools.org/. http://esper.codehaus.org/.
c Springer International Publishing Switzerland 2015 F. Gandon et al. (Eds.): ESWC 2015, LNCS 9341, pp. 82–86, 2015. DOI: 10.1007/978-3-319-25639-9 16
Supporting Real-Time Monitoring in Criminal Investigations
83
under the direct control of the system itself, or when other technologies need to be included in the event processing pipeline. In this paper we present an approach for creating decoupled semantically enabled event processing systems by leveraging existing technologies, and demonstrate its applicability in a criminal intelligence scenario.
2
Related Work
The Streaming Linked Data framework, based on C-SPARQL, allows publishers to stream data to a central server, where the data can be queried, stored, replayed, decorated, and republished as new streams [2]. This drastically improves the flexibility of the RSP system, making it possible to provide APIs and add functionality to the standard C-SPARQL system. The Super Stream Collider is
Data Loading...
