Self organization in content delivery networks

The Content Delivery Network paradigm with a centralized approach shows its limits in large and dynamic systems. Decentralized algorithms and protocols, such as peer-topeer (P2P) and multi agent systems, can be usefully employed. In this work, an algorithm exploiting bio-inspired agents to organize the contents in Content Delivery Networks, is proposed. Agents move and reorganize the metadata that describe the contents to improve information retrieval. Preliminary results confirm the efficacy of the approach. © 2012 IEEE.