Investigating Random Linear Coding from a Pricing Perspective.

Hailing Zhu,Khmaies Ouahada

doi:10.3390/e20080548

Abstract

In this paper, we study the implications of using a form of network coding known as Random Linear Coding (RLC) for unicast communications from an economic perspective by investigating a simple scenario, in which several network nodes, the users, download files from the Internet via another network node, the sender, and the receivers as users pay a certain price to the sender for this service. The mean packet delay for a transmission scheme with RLC is analyzed and applied into an optimal pricing model to characterize the optimal admission rate, price and revenue. The simulation results show that RLC achieves better performance in terms of both mean packet delay and revenue compared to the basic retransmission scheme.

Highlights

In the concept of network coding [1], encoding operations are performed on a group of data packets (e.g., by using the exclusive-OR (XOR) function) before they are forwarded towards their destinations, which departs from the conventional store-and-forward transmission paradigm
The common idea behind these applications is that a source node encodes information packets using Random Linear Coding (RLC), all intermediate nodes need to perform RLC to encode the random linear combinations that they hear from the source node or other intermediate nodes and send out these new random linear combinations until the receiver decodes the original information packets
The objective of this paper is to study the mean packet delay with RLC and quantify the impact of RLC from a pricing perspective by comparing the optimal admission rate, the optimal price, and the corresponding expected delay as well as optimal revenue obtained by a node using RLC with those obtained from the basic retransmission scheme

Summary

Introduction

In the concept of network coding [1], encoding operations are performed on a group of data packets (e.g., by using the exclusive-OR (XOR) function) before they are forwarded towards their destinations, which departs from the conventional store-and-forward transmission paradigm. It is shown that RLC can offer benefits over uncoded transmissions in terms of both throughput and delay if the blocksize is adapted to the traffic load It is suggested in [11] that bulk-service queueing , which has been applied in other works (e.g., [12,13,14]), for RLC performance analysis, is an appropriate model to capture the fact that randomly arriving packets are encoded, transmitted, and removed from the queue in groups corresponding to blocks. Geo/Geo/1 queuing model and an approximate but analytically tractable expression for the mean packet delay with RLC can be derived Based on this approximate expression, we analyze the optimal admission rate, the optimal price and the corresponding expected delay when the sender node chooses to use RLC and compare them with those obtained from the basic retransmission scheme to study the performance of RLC from an economic point of view.

Related Work

System Model

User Utility

Revenue Optimization

Numerical Examples

Conclusions