A Novel Dynamic Adaptive Video Streaming
Solution in Content-Centric Mobile Network
Yiran Wei, Changqiao Xu, Mu Wang, Jianfeng Guan
State Key Laboratory of Networking and Switching Technology,
Beijing University of Posts and Telecommunications, Bejing 100876, China
{wyr100, cqxu, wangmu, jfguan}@bupt.edu.cn
Abstract—The newly rising of dynamic adaptive streaming
over HTTP (DASH) enables consumers access diverse bit rate
encoded video content according to the link situation, which
improves the quality of experience (QoE) of consumers in
mobile environment. While Content-Centric Mobile Networks
(CCMNs) bring content centric design to mobile environment
and has already become a promising solution for the content-
based applications. The major challenges of DASH in CCMNs are
how to determine proper video quality in a mobile environment
and to reduce broadcasting storm. To address these problems,
we design a novel adaptive video streaming solution in CCMNs
(DAS-CCMN). We first analyze the load degree of content
carrier, and define an interest satisfied potential (ISP) concept to
reflect the ability that a content carrier can satisfy the interest
request from a consumer. In DAS-CCMN, each mobile node
shares its ISP information with neighbors. All nodes store this
information in their interest satisfied potential table (ISPT).
Based on information recorded in ISPT, a self-learning based
rate determination strategy is presented to choose video content
with proper bit rate. Moreover, an interest flooding control is
presented to solve the broadcast storming problem for adaptive
video streaming in CCMNs. Simulation results show our solution
improves the performance of current DASH service in CCMNs.
Keywords—Content-centric network, Dynamic adaptive video
streaming, Mobile ad hoc network, Rate determination, Interest
control.
I. INTRODUCTION
Recently, with the fast development of wireless communi-
cations, video streaming services in mobile ad hoc network
(MANET) has got fast development. More and more people
are using mobile terminals, such as smart phones, tablets
and car-mounted terminals, to access video content from the
Internet. However, the unstable link in MANET may lead
to frequent playback freeze and high start-up delay when
continuous using constant bit rate technology [1][2], which
decreases the QoE level of mobile consumers. In this context,
dynamic adaptive streaming via HTTP (DASH) has become a
hot issue both in academy and industry. The goal of DASH
is to distribute video content with high QoE even in dynamic
network situations. DASH technology allows consumers re-
quest video content with different bit rates dynamically, which
is based on the currently available bandwidth. Compared
with constant bit rate playback, DASH can provide smooth
playback of high quality video content, especially for mobile
environment. Therefore, DASH has become a main trend for
the future mobile Internet video streaming.
On the other hand, Content-Centric Networking (CCN, aka
NDN) is a promising future Internet architecture. Content-
Centric Mobile Networks (CCMNs) as a extension version
of CCN in mobile environment has attracted more and more
research interests for real-time applications. Owning to the
wide spread of online multimedia applications, consumers
care more about “what” the content is, instead of “where”
the content is located. The initial IP network is to transmit
content host to host, which severely restricts the content
sharing efficiency. To solve this problem, Jacobson et.al [3]
proposed CCN, a location-dependence and mobility support
architecture, which can continue exchange data even when
connectivity is rapidly changing. CCN employs ubiquitous in-
network caching and content name based routing design to
achieve content access for consumers, which helps to provide
efficient video streaming in mobile environment.
CCN and DASH have several common features such as the
client/interest request initiated pull approach, the content is
dealt with in segments for delivering and content oriented
operations. Since CCN can fully explore the advantages of
DASH according to our previous work [4], it is useful to
combine DASH technology with CCMNs to improve QoE
for delay-sensitive application like video streaming. However,
directly deploying DASH over CCMNs faces many challenges.
1) Due to in-network caching of video content with different
bit rates, it is inappropriate to use traditional end-to-end
bandwidth estimation based rate determination mechanism.
For example, when a video segment with high bit rate is
cached in a geographical closed place, the transmission delay
of this segment will be relatively low. Thus consumers will
switch to high bit rate version for next segment according to
the bandwidth estimation. If the next segment is not cached
nearby, the interest packet will be forwarded through a long
trip to the source server with long transmission delay. Then
consumer will downshift to a low bit rate content. This
repeated cycle called the bit rate oscillation [5] results in
low QoE level of consumers. 2) CCMN architecture faces
an interest flooding problem caused by its epidemic nature
[6][7]. For DASH over CCMN, interest flooding leads to
network congestions and duplication in packet transmission.
Consumers may downshift the requested video quality due to
long time delay caused by congestions. And system is not easy
to recovery after network congestion happens.