Overhead Control with Reliable Transmission of Popular Packets in Ad-Hoc Social Networks

TL;DR

This paper introduces RTPS, a novel bandwidth allocation and acknowledgment prioritization technique for ad-hoc social networks, significantly improving transmission reliability and reducing contention compared to traditional TCP methods.

Contribution

The paper proposes RTPS, a new method that enhances data transmission reliability in ASNETs by prioritizing popular nodes and acknowledgments, addressing TCP limitations.

Findings

RTPS improves throughput and latency in ASNETs.

RTPS reduces packet loss and contention.

RTPS outperforms traditional TCP in various network scenarios.

Abstract

Reliable social connectivity and transmission of data for popular nodes is vital in multihop Ad-hoc Social Networks (ASNETs). In this networking paradigm, transmission unreliability could be caused by multiple social applications running on a single node. This leads to contentions among nodes and connection paths. In addition, congestions can be the result of multiple senders transmitting data to a single receiver and every sender waiting for a positive acknowledgment to move on. Therefore, traditional Transmission Control Protocol (TCP) performs poorly in ASNETs, due to the fact that the available bandwidth is shared among nodes using round trip time and the acknowledgment is provided individually to every data packet. To solve these issues, we propose a technique, called Overhead Control with Reliable Transmission of Popular Packets in Ad-Hoc Social Networks (RTPS), which improves transmission reliability by assigning bandwidth to users based on their popularity levels: extra bandwidth is assigned to the nodes with higher popularity and their acknowledgments are sent with higher priority. In addition, RTPS further reduces contentions and packet losses by delaying acknowledgment packet transmissions. Our detailed investigations demonstrate the excellent performance of RTPS in terms of throughput latency and overhead with different hop-distances and different numbers of concurrent TCP flows.