A Contract-Based Incentive Mechanism for Data
Caching in Ultra-Dense Small-Cells Networks
Shunfeng Chu
1
, Jun Li
1
, Tingting Liu
1,2
, Feng Shu
1
1
School of Electrical and Optical Engineering, Nanjing University of Science and Technology, Nanjing, China 210094
2
Nanjing Institute of Technology, Nanjing, China 211167
Email: shunfeng.chu@njust.edu.cn, jleesr80@gmail.com, liutt@njit.edu.cn, shufeng@njust.edu.cn
Abstract—Wireless caching is an efficient mechanism for re-
ducing downloading delay and reducing the traffic pressure over
backhual channels by caching some popular content, e.g., video
clips, in small base stations (SBSs). In this paper, we consider a
commercialized small-cell caching system consisting of a network
service provider (NSP), several video retailers (VRs) and mobile
users (MUs). The NSP leases its SBSs to VRs in order to earn
profits, while the VRs store popular videos into the lent SBSs,
thereby gaining profits from providing better services to the MUs.
We conceive the system within the framework of contract theory
by designing the optimal quality-price contract. We establish the
profit function of NSP and VRs and solve the profit maximization
problem through contract theory. Numerical results validate the
effectiveness of our incentive mechanism for the system.
Index Terms—Small-cell caching, cellular networks, stochastic
geometry, incentive mechanism, Contract Theory.
I. INTRODUCTION
Wireless data traffic is expected to increase exponentially
in the next few years driven by a staggering growth of mobile
users (MU) and their bandwidth-hungry mobile applications.
The redundancy of data transmissions can be reduced by
locally storing popular content, known as caching, into the
memory of intermediate network nodes, effectively forming
a local caching system [1]. The local caching brings content
closer to the MUs and alleviates redundant data transmissions
via redirecting the downloading requests to the intermediate
nodes. Generally, wireless data caching consists of two stages:
data placement and data delivery [2]. In the data placement
stage, popular data are cached into local storages during off-
peak time, while in the data delivery stage, requested contents
are delivered from the local caching system to the MUs.
As small-cell embedded architectures will be prevailing in
future cellular networks, known as heterogeneous networks
(HetNet) [3], caching relying on small-cell base stations
(SBS), namely, small-cell caching, is a promising trend for
the HetNets. In [4], a small-cell caching scheme, called
‘Femtocaching’, is proposed for a cellular network embedded
with SBSs, where the data placement at the SBSs is optimized
in a centralized manner for reducing the transmission delay
imposed. However, [4] considers an idealized system, where
neither the interference nor the impact of wireless channels is
taken into account. In [5], the small-cell caching is investigated
in the context of stochastic networks. The average performance
is developed via stochastic geometry theory [6], [7], where the
distribution of network nodes are modeled by Poisson point
process (PPP). However, the caching strategy in [5] assumes
that the SBSs always cache the same content.
From above discussions, considering the data placement
issue is important for current research. However, the caching
model combine many issues instead of data placement. From a
commercial standpoint, considering the topics such as pricing
on video streaming, renting local storages, is more interesting.
As video transmissions dominate the mobile data traffic, we
consider a commercialized caching system which is consisted
of video retailers (VR), network service providers (NSP) and
MUs. The VRs buy the right to the videos and publish the
videos on their web. The NSPs are typically operators of
cellular networks, because of the control of network facilities,
such as macro-cell base stations and SBSs.
In such a commercialized caching system, the VRs provide
video streaming services to MUs to earn profits. As the central
servers of the VRs, which store the popular videos, are usually
located at backbone networks and far away from the MUs, an
efficient solution is to locally cache these videos for reducing
the transmission latency, thereby attracting more customers.
These local caching demands raised by the VRs offer the NSPs
profitable opportunities from leasing their resources, i.e., the S-
BSs. Besides, by reducing redundant video transmissions over
SBSs’ back-haul channels, the NSPs can save considerable
costs. Under this circumstances, both the VRs and NSPs are
the beneficiaries from the local caching system. However, each
participant is selfish and wishes to maximize its own benefit,
leading to a competition and optimization problem.
In this paper, we research on a commercialized caching
prototype within a contract theoretic framework. The system
consists of an NSP and multiple VRs, where the NSP, as the
monopolist in the market in charge of the trading resource, i.e.,
SBSs, wishes to lease its SBSs to the VRs for the purpose
of making profits. To comply with the future trend in 5G,
we consider the ultra-dense deployment of the SBSs, i.e., the
number of the SBSs is much higher than that of the MUs.
The main contributions of the paper are as follow. First of all,
by modeling the MUs and SBSs using two independent PPPs,
we develop the probability expression of direct downloading.
Then, we formulate contracts betweeen VRs and NSP and state
the feasibility of the contract. Next, we solve the optimization
problem by a series of transformation. Finally, we provide
several results for the pricing and SBSs allocation scheme.
978-1-5090-4183-1/17/$31.00 ©2017 IEEE