Leader Election in Opportunistic Networks
Radu Stefan Dragan, Radu‐Ioan Ciobanu, Ciprian Dobre · 2017
New technologies brought to life by the rapid growth in number of mobile devices are starting to be employed for creating cyberinfrastructures for smart cities. One such technology comes in the form of opportunistic networks (ONs), where mobile nodes only communicate between each other through protocols such as Bluetooth or Wi-Fi Direct. The usefulness of ONs in smart cities is high, because they are able to reduce the load on the city's infrastructure, while at the same time decreasing the response time. Multiple opportunistic nodes need to collaborate towards opportunistic computing and smart cyberinfrastructures, and for this reason we argue that the problem of consensus in ONs is of the utmost importance. Our aim is to propose a general-purpose consensus algorithm for opportunistic networks. However, to expect a given portion of all the nodes inside an opportunistic networks to agree upon a subject is certainly overly-optimistic. Thereby, we aim to group the nodes in communities based on their closeness to each other and the time they spend together. Then, a leader is elected in each community, and the leader will be in charge with the consensus decision. In this paper, we focus on grouping nodes into communities and electing the leader in each community. We propose and analyze two approaches, which we compare in order to find the most suitable one.