Abstract | ||
---|---|---|
Organic Computing (OC) enhances computer systems by postulating life-like properties to enable a system to self-configure, self-heal, self-optimize and self-protect. It is a solution to reduce the complexity of systems but is based on a benevolence assumption that all parts of the system are reliable and interested to further the system goal. In open and heterogeneous systems, the benevolence assumption is unrealistic, since uncertainties about the participants' behavior have to be regarded. We propose trust as a concept to cope with these uncertainties. This paper presents a trust enhancement of the self-configuration algorithm based on the well-known Contract Net Protocol. This baseline algorithm can be used in a distributed system, i.e., multi-agent system, cloud computing or grid system, to equally distribute the load of services on the nodes. However, the trust enhancement of self-configuration assigns services with different importance levels to nodes so that more important services are assigned to more reliable nodes. Evaluations have been conducted to rate the effectiveness of the algorithm when nodes are failing, i.e., the reduction of failures of important services. The results show that our self-configuration algorithm increases the availability of important services by more than 12%. To our knowledge this is the first trust integrated self-configuration process that proposes to build reliable and robust heterogeneous distributed systems in a decentralized way. |
Year | DOI | Venue |
---|---|---|
2014 | 10.1007/978-3-319-04891-8_4 | ARCS |
Keywords | Field | DocType |
contract net protocol,trust | Computer science,Real-time computing,Grid system,Organic computing,Contract Net Protocol,Cloud computing,Distributed computing | Conference |
Citations | PageRank | References |
2 | 0.39 | 15 |
Authors | ||
4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Nizar Msadek | 1 | 16 | 3.04 |
Rolf Kiefhaber | 2 | 78 | 7.84 |
Bernhard Fechner | 3 | 78 | 12.18 |
Theo Ungerer | 4 | 1262 | 136.24 |