Improved identification of peers using enhanced gossip++protocol in NAT

With the rise in the number of devices getting connected to the network and the limited size of IPv4 addresses, the demand for depicting a way to provide unique IP increased. Network address translation (NAT) is a common technique used to map a globally unique IP address to one or more non-unique IP addresses. In a peer to peer scenario, NAT imposes a restriction on establishing end to end connectivity between the peers. The proposed work employs an enhanced gossip algorithm which increases the exploration of private peers by accommodating the geographical properties of peers in the network. The peers only have the partial view of the network which is more refined here by making it as multiple partial views, storing public and private information separately. The reliability of the peers in the network is taken care of by participation factor. The peers thus identified are grouped based on the device specification and supported media formats thus making the system deployable in live streaming applications. The accommodation of proposed constraints in the gossip++protocol has increased the versatility of the system in a churn based dynamic scenario. Gossip++ when executed in a simulated environment exhibited 12.85% decrease in system closure time.