- Home
- The OK System
- Research Highlights
- Structure-Preserving Semantic Matching
- Good-Enough Answers
- OK Kernel
- Adaptive Routing in P2P Overlays
- Architecture for P2P Reasoning
- Privacy in Resource Discovery
- Trust in P2P
- Ontology Matching Techniques
- Query Answering in Large, Open Systems
- Cautious Knowledge Sharing
- Trust in Ontological Entities
- Interaction-situated Semantic Alignment
- Massively Scalable WS Discovery
- Applications Areas
- Demonstrations
- Downloads
- Publications
- Deliverables
- People
The OpenKnowledge Kernel
One of the main results of the OpenKnowledge project is the OpenKnowledge Kernel [1]. Each Kernel contains several modules that integrates or freshly implements the research from the different partners in the project. The development of the software started right at the beginning of the project in order to maximize the chance that integration of the research goes smoothly. The Kernel software can be downloaded from the openk.org website and easily installed on any machine that runs Java.
When a Kernel is launched and correctly configured it can form or participate in an OpenKnowledge Network. Such a network is organized in a Peer-to-Peer overlay over a LAN or WAN. It can be used to search, distribute, combine and execute interactions and services, called Interaction Models(IMs) and OpenKnowledge Components (OKCs) respectively. OKCs are pieces of software that offer some kind of re-usable and stateless functionality like a WebService. IMs act like a formal protocol that describes how roles interact. OKCs are implementations of such roles. More precisely, IMs and OKCs are the main currency of the OpenKnowledge Network. They can be shared on the network via a publish functionality offered by the Kernel software. The roles in the IMs are automatically matched with available OKCs running on peers participating in the network. In order for peers to autonomously decide with whom to interact, a negotiation procedure precedes the actual execution of the roles. Such executions are monitored by Coordinators.
The case-studies in the project all make use of the Kernel software, meaning that for each of them IMs and OKCs are developed compatible to run on the OpenKnowledge Network.
Several papers are written to evaluate the properties of the modules of the Kernel. For example in [2] an extensive emulation is performed of the Discovery Service which is a distributed scalable search system for storing and retrieving IMs, OKCs, subscriptions of peers etc.
The ultimate goal of this work is to establish a global scale network where thousands of users share and run IMs and OKCs. To facilitate this, we try to make our software very easy to install and give clear examples and tool support to facilitate feeding the network.
Related Publications
- [1] Siebes R, Dupplaw D, Kotoulas S, de Pinnick PA, van Harmelen F, Robertson D. 2007. The OpenKnowledge System: an interaction-centered approach to knowledge sharing. Proceedings of the 15th International Conference on Cooperative information systems (CoopIS).
- [2] Kotoulas, S and Siebes, R. 2007. Rarity-based routing in peer-to-peer networks. COPS Workshop at IEEE WETICE.