A P2P Traffic Management Model Based on an
ISP Game
Chunzhi Wang, Shuping Wang, Hui Xu, *Hongwei Chen
School of Computer Science, Hubei University of Technology, Wuhan, China
Email: chw2001@sina.com
Abstract—While P2P applications enrich the network
application, they consume huge network bandwidth and
have a great impact on ISP. As for current traffic
optimization problem, maybe the most effective analysis tool
is the game theory. This paper proposes an ISP-involved
P2P network traffic management framework, builds a game
model and its equilibrium solution, then from the
perspective of evolution, performs convergence analysis on
the equilibrium solution, based on this, generates a traffic
management optimization algorithm, and discusses the
fairness of the algorithm. Finally, the simulation
experiments show that, the model can reach the purpose of
optimizing traffic management.
Index Terms—Traffic Management, ISP, P2P, Game Theory
I. INTRODUCTION
P2P (Peer-to-Peer) is a distributed network, and a peer
in the P2P network acts as the role of both a server and a
client. While P2P applications enrich the network
application, they consume huge network bandwidth and
have a great impact on ISP (Internet Service Provider). In
addition, the mismatch between the overlay networks and
underlay networks leads to large redundant traffic, which
strengthens the tension between P2P content providers
and ISPs.
Many researchers try to solve this problem with such
methods as cache management and traffic localization.
GuoQiang Zhang etc. survey the P2P traffic optimization
technologies from three aspects: P2P cache, traffic
locality-awareness and data scheduling [1]. Literature [2]
summarizes ISPs’ P2P traffic management schemes: p2p
blocking, p2p caching, Localization (peers), Localization
(ISPs). In order to minimize the total amount of P2P
traffic, Noriaki Kamiyama etc. present an optimum
design for capacity and location of caches based on
dynamic programming method, assuming that a transit
ISP provides caches at transit links to access ISP
networks [3]. Miyoshi etc. present a new method for P2P
traffic localization, featuring the insertion of an additional
delay into each P2P packet based on the geographic
location of its destination [4]. Byungryeol Sim etc. have
assessed the impacts of ALTO (Application-Layer Traffic
Optimization Protocol) on P2P applications from the
respects of network traffic optimization [5]. However,
researches show that residential ISPs can actually lose
money when localization is employed, and some of them
will not see increased profitability until other ISPs
employ localization [6]. So, it’s necessary to reduce
traffic and increase ISP profit through cooperation
between P2P and ISPs. Recently, some scholars study
P2P traffic optimization problem from the cooperation
between P2P and ISPs. Literature [7] has studied whether
a cooperative caching scheme could help ISPs to decrease
traffic costs caused by P2P applications. Literature [8]
provides a ranking service that applies the ISP’s own
policies to the P2P peer selection flexibly. Through the
peer selection policy, it can effectively control download
traffic. Peng Yang etc. propose a rate allocation
mechanism for achieving a balance between the cross-ISP
P2P traffic and the P2P streaming performance [9].
As for current traffic optimization problem, maybe the
most effective analysis tool is the game theory. From a
theoretical perspective, it is feasible to achieve the
purpose of traffic optimization if the ISP manages content
through certain game strategy. There are many
researchers apply game theory to P2P and ISP. Literature
[10,11] describe a game theoretic framework for scalable
video streaming over a P2P network. Literature [12]
optimizes the non-cooperative P2P network from the
game theory point of view. Literature [13] provides a new
framework based on spatial evolutionary game theory for
incentive mechanism to encourage cooperation among
peers in P2P networks. Literature [14] presents a game
theoretic framework to help the design of techniques
encouraging the ISP cooperation in P2P streaming
applications and decreasing unnecessary inter-ISP
streaming traffic. Literature [15] formulates the
interaction among ISPs and subscribers in a local market
with two ISPs competing with each other as a two-stage
game, and studies the influence of different traffic
patterns on the Nash Equilibrium of the market.
Literature [16] studies two games that model the adoption
of ISP-driven locality promotion and of ISP-owned
caches that intervene in the overlay.
In this paper, we apply game theory to P2P traffic
optimization. The remaining of this paper is organized as
follows: Section 2 proposes a P2P traffic management
framework involved with ISPs. Section 3 establishes a
game theory model and solves the Nash equilibrium, and
analyzes convergence property of the Nash equilibrium
from respect of evolution. On the basis of above, a P2P
traffic optimization algorithm is presented, and its
fairness is discussed. In section 4, the simulative result
shows that the model can achieve the effect on traffic
JOURNAL OF COMPUTERS, VOL. 9, NO. 6, JUNE 2014
doi:10.4304/jcp.9.6.1478-1483