IEEE Communications Magazine - June 2017 - page 7

IEEE Communications Magazine • June 2017
5
T
he
P
resident
s
P
age
(50 billion devices estimated by 2020, 500 billion by 2025).
These connected devices, triggered by the emergence of IoT,
open the door to innovations that facilitate new interactions
among “things” and humans, and provide new opportunities
for applications, infrastructures and services. Unfortunately,
most of the IoT devices are designed for convenience and
functionality without taking into consideration security. There-
fore, securing massive IoT devices becomes a critical task.
Quantum Security and Cryptography:
With the advance
of quantum computation and communication technology,
the landscape has changed in the fields of communication
security and cryptography. People are starting to look into
the potential threat posed to encryption by quantum com-
puting. It is crucial to have secure crypto systems in the
post-quantum era.
I
nternet
Internet Security:
From the Internet security TC,
a promising direction has been proposed involving the
usage of Blockchain approaches to secure control-plane
authentication operations, by means of fully decentralized
consensus. The basic idea is to try to no longer rely on
authentication servers to validate transactions in network
operations, but use Blockchain as in the bitcoin cryptocur-
rency. Essentially, a transaction is validated if the majority
of nodes in a large network of nodes validate it. The chal-
lenge here is to define protocols to speed up the valida-
tion process.
Networked Games:
Online gaming is one of the appli-
cations stressing the network most in terms of perfor-
mance requirements. Automating the mobile application
task of offloading as a function of the network conditions
from the user or a group of users to a remote server is a
challenge being discussed for the network gaming use-
case, especially in the area of mobile Internet and edge
computing environments.
Network Data Analytics and Machine Learning:
Increasing attention has been observed toward inte-
grating traffic and mobility analytics engines, not only
in user’s computing service management, but also in
network provisioning primitives. This trend suggests that
there is an interest in going beyond common network
management practices that consider a network configu-
ration policy to be good if it is stable and does not need
to change often. Behind the term network analytics there
is the idea to profit from machine learning algorithms in
the implementation of online clustering solutions.
G
reen
C
ommunication and
C
omputing
Green 5G Wireless Communication Systems and Applica-
tions:
The dramatic increase of data traffic, system scale, and
applications in the incoming fifth generation (5G) wireless
communications has introduced significant Green challeng-
es, including increasing energy and resource consumption,
and negative environmental impacts that deserve intensive
research and development efforts.
Big Data and Green Challenges:
The Big Data era has
been recently found to have high correlations to Green chal-
lenges. There are two aspects: how to green Big Data sys-
tems themselves, and how to apply big data technologies
to general Green objectives in various applications. Both
present challenging research and development problems that
must be addressed.
C
ognitive
N
etworks
Machine Learning for Cognitive and Flexible Wireless
Networks:
Machine learning can provide promising solutions
to build a truly “smart” cognitive radio network. It will also
help develop cognitive networks in the broad sense, address-
ing issues arising in context-aware networks, augmented real-
ity, and autonomous systems.
Data-Driven Network Cognition Analysis and Design:
Data-driven spectrum sensing and sharing for IoT and
mm-wave applications are receiving strong interest from both
academia and industry. Data analytics can help understand
what is the network’s performance bottleneck and what kind
of cognition is needed most.
Techno-Economic Regulatory Framework for Spectrum
Sharing:
Such a framework can help bridge the gap among
the technical aspects of spectrum sharing research, the eco-
nomic viability of various technical solutions, and the current
or future-foreseeable reality in terms of what will be allowed
or is feasible according to national and international spec-
trum regulations.
Software-Adaptability for Spectrum Flexibility and Shar-
ing:
A software design/engineering approach is profound-
ly important in determining the radio flexibility that can be
achieved, and the speed and protocol with which adaptabil-
ity and spectrum sharing opportunities are obtained. Hence,
the optimal design of software in software radios for tailored
flexibility must be considered to maximally leverage viable
spectrum sharing concepts and opportunities.
O
ptical
N
etworks
Software Defined Networking-Based Techniques and
Related Network Virtualization (NV)/Network Functions
Virtualization:
Data networks today support programmability
through software defined networking (SDN) and network
functions virtualization (NFV). While optical system vendors
advertise some form of SDN in their product offerings, in
reality, the functionality provided by this SDN interface is
severely limited. New research directions should address
this shortcoming and also explore any performance and/
or stability issues that programmability introduces on optical
networks.
Elastic Optical Networking:
Elastic optical networks have
the potential to overcome the fixed, coarse granularity of
existing WDM technology, and are expected to support
flexible data rates, adapt dynamically to variable bandwidth
demands by applications, and utilize the available spec-
trum more efficiently. New research in spectrum manage-
ment techniques and bandwidth-variable transponders and
cross-connects is necessary to realize the potential of this
technology.
Optical Inter- and Intra-Data Center Networks:
The explo-
sive growth in demand for computation, storage and data
transfer has imposed great challenges in intra- and inter-data
center network design. Ethernet-based electrical data center
network designs face issues related to massive and compli-
cated wiring, resource over-provisioning, and high power
consumption. Therefore, hybrid architectures using both
electrical and optical network devices and/or all-optical data
center network designs are necessary to address these issues
while achieving scalability.
Integration of Optical Wired and Wireless Networks:
Free
space optics (FSO) is a mature technology that may provide
very high data rate (tens of gigabits per second) over long
1,2,3,4,5,6 8,9,10,11,12,13,14,15,16,17,...228
Powered by FlippingBook