IEEE Network - March / April 2017 - page 6

IEEE Network • March/April 2017
4
The Scanning the Literature column provides concise summaries of selected papers that are recently published in the field
of networking. Each summary describes the paper’s main idea, methodology and technical contributions. The purpose of the
column is to bring the state-of-art of networking research to readers of
IEEE Network
. Authors are also welcome to recommend
their recently published work to the column, and papers with novel ideas, solid work and significant contributions to the field are
especially appreciated. Authors wishing to have their papers presented in the column should contact the editor.
Xiaohua Tian, Shanghai Jiao Tong University
Recent years have witnessed the efforts of integrating
intelligence into the network, for which the concepts such
as active networks, application oriented networks, software
defined networks, and named data networks are proposed.
The nature of machine intelligence is to leverage the device’s
computing capability to take automatic, adaptive, and appro-
priate actions in heterogeneous scenarios. In this perspective,
the column in this issue focuses on recent advances in intelli-
gent networks. In particular, the work tackling network securi-
ty, efficiency and robustness issues with the device’s enhanced
intelligence.
With the proliferation of mobile applications, co-existence of
heterogeneous wireless devices in the common physical space
becomes inevitable. Operating at the same unlicensed Indus-
trial Scientific Medical (ISM) bands, devices following different
communication standards (e.g. WiFi, Bluetooth, ZigBee) cannot
directly share information with each other, but can incur serious
interference and degraded performance. Co-existence of active
wireless devices that are with the same frequency and different
protocols thus becomes a hot research topic in both academia
and industry. Meng Jin
et al
. propose an innovative solution
called Smoggy-Link to deal with the issue of cross-technology
interference.
Smoggy-Link: Fingerprinting Interference for Predictable Wireless
Concurrency
Meng Jin, Yuan He, Xiaolong Zheng, Dingyi Fang, Dan Xu, Tianzhang Xing, and
Xiaojiang Chen,
Proc. IEEE ICNP 2016
, pp. 1–10
Differing from conventional interference-avoidance based
approaches, Smoggy-Link exploits potential concurrency for
adaptive ZigBee transmissions under harsh interference. Smog-
gy-Link maintains an accurate link model to describe and trace
the relationship between interference and link quality of a
sender’s outbound links. Smoggy-Link can obtain fine-grained
spatio-temporal fingerprints of any interference, using a low-
cost interference identification method. The fingerprints of
interference are further utilized for adaptive link selection and
intelligent transmission scheduling. According to the real-world
implementation and evaluation results, Smoggy-Link has con-
sistent improvements in both throughput and packet reception
ratio, compared to the state-of-the-art interference-avoidance
based approaches. This work shows the authors’ creative think-
ing toward interference issues.
As the increasing mobile traffic compels the mobile access
network infrastructure to be increasingly chaotic and dense,
the software defined network (SDN)-enabled mobile access
network is bound to have a good prospect in the future intelli-
gent network, due to its excellent ability to support scalable and
flexible network applications. However, one important problem
needs to be satisfactorily solved, i.e., how to efficiently forward
data flows within the mobile access network infrastructure.
The term efficiently is of high importance, because the rule
space of SDN devices is limited. In order to improve network
performance with a limited rule space, Mianxiong Dong
et al
.
propose SRCMN in:
Rule Caching in SDN-Enabled Mobile Access Networks
Mianxiong Dong, He Li, Kaoru Ota, and Jiang Xiao,
IEEE Network
, vol. 29, no. 4, pp.
40–45, July 2015
The authors design a two-layer rule space, managed by the
SDN controller in each SDN device. With the consideration of
the user mobility, they also model the rule optimization prob-
lem and implement a cache prefetching scheme. The authors
conduct simulations that demonstrate the proposed structure
and mechanism outperform traditional rule management meth-
ods under various network settings. This work makes a contri-
bution to solving a significant problem in SDN enabled mobile
access networks.
Meanwhile, some work studies the malfunctioning of SDN.
Controllers in SDN dictate the forwarding elements: which
packets are processed by which tables in what order. When
a packet cannot be matched, it is essential for a controller to
be able to verify the forwarding plane topology and identify
link failures at forwarding plane speeds. Existing approaches
employing protocols like Link Layer Discovery Protocol could
result in control plane congestion, and could not work under
control plane partitions. Ulaş C. Kozat
et al
. propose a novel
method to verify the forwarding plane topology and identify
link failures in:
On Optimal Topology Verification and Failure Localization for
Software Defined Networks
Ulaş C. Kozat, Guanfeng Liang, Koray Kökten, and János Tapolcai,
IEEE/ACM Trans-
actions on Networking
, vol. 24, no. 5, pp. 2899–2912, Oct. 2016
In this work, the authors show that topology verification can
be achieved optimally in terms of using a minimum number of
static forwarding rules and control messages. They present a
solution based on shortest possible walk that visits every link in
the forwarding plane at least once. This solution iterates over an
initially computed shortest walk to reduce the number of for-
warding rules while preserving the walk length using a greedy
heuristic. To locate an arbitrary single link failure, the authors
present order optimal solutions. They also show how to trade
off the number of control messages and static forwarding rules
to speed up failure localization. The results of this paper are
convincing, for the proposed approach is tested over real-world
topologies.
Dynamic adaptive streaming (DAS), especially the standard
MPEG-DASH, has been widely adopted by multimedia contents
providers to distribute real-time streaming. This type of stream-
ing approach describes the way multimedia contents are seg-
mented, and uses an adaptation mechanism at the client side
to decide which representation of these segments are to be
sent to clients. This technology has been well developed in the
classical IP-based networks. With the rise of Information-Centric
Networking (ICN) and Named Data Networking (NDN), which
transmit data based on the contents instead of the hosts, it is
valuable to study the performance of DAS in NDN. Benjamin
Rainer
et al
. investigate this problem in the following paper:
SCANNING THE LITERATURE
1,2,3,4,5 7,8,9,10,11,12,13,14,15,16,...100
Powered by FlippingBook