On using P2P technology to enable opportunistic management in DTNs through statistical estimation

摘要

Disruption-Tolerant Networks (DTNs) are characterized by long delays and constant disconnections among nodes. These networks have management needs analogous to those of traditional computer networks, but, their intrinsic characteristics (e.g., intermittent connections) hinder the execution of network management tasks. The employment of P2P technology is an alternative for managing DTNs since this technology promotes the autonomy of management entities (i.e., management peers). This autonomy enables the use of local logic and data in order to perform opportunistic management tasks. We propose a solution to estimate the future contacts of nodes through distributed statistical analysis of past contacts. Besides that, nodes can share information about their contacts to improve contact estimation using a P2P management overlay. The proposed solution was implemented using an open source P2P-Based Network Management (P2PBNM) system, ManP2P-ng. Furthermore, the solution was evaluated through experiments performed using publicly available DTN traces. The results show that the proposed solution can improve the execution of management tasks in DTNs.