A platform for distributing and reasoning with OWL-EL knowledge bases in a Peer-to-Peer environment Alexander De Leon 1 , Michel Dumontier 1,2,3 1 School of Computer Science 2 Department of Biology 3 Instititute of Biochemistry Carleton University, 1125 Colonel By Drive, K1S 5B6, Ottawa, Canada adlbatti@scs.carleton.ca, michel dumontier@carleton.ca Abstract. Memory exhaustion is a common problem in tableau-based OWL reasoners, when reasoning with large ontologies. One possible solu- tion is to distribute the reasoning task across multiple machines. In this paper, we present, as preliminary work, a prototypical implementation for distributing OWL-EL ontologies over a Peer-to-Peer network, and reasoning with them in a distributed manner. The algorithms presented are based on Distributed Hash Table (DHT), a common technique used by Peer-to-Peer applications. The system implementation was developed using the JXTA P2P platform and the Pellet OWL-DL reasoner. It re- mains to demonstrate the efficiency of our method and implementation with respect to stand alone reasoners and other distributed systems. 1 Introduction Scalability is the main technological challenge faced by OWL-DL reasoners. This issue is a natural consequence of the computational complexity of reasoning problems. The tableau algorithm, used by most OWL-DL reasoners, is known to be NExpTime-complete and can create data structures of double exponential size of the input [6]. In other words, the algorithm is not only CPU intensive but also requires a large amount of memory for the case of large ontologies. This paper explores the idea of implementing a distributed OWL reasoner using a peer-to-peer (P2P) approach. The goal is to distribute the tableau algorithm such that it can utilize the combined resources of the P2P network, and therefore mitigate the memory issues that arise when reasoning with large ontologies on a single machine. As part of the OWL 2 recommendation, different fragments are proposed which trade expressivity in favor of efficient reasoning. These fragments have been named OWL profiles and each of them defines a set of syntactic restrictions of the full OWL 2 language. Each profile targets a different type of application. The OWL-EL profile is a subset of OWL-DL based on the EL + + light weight description logic [4]. This profile is recommended for ontologies with a large