Image Processing Reference
In-Depth Information
diversity to establish radio links that are more reliable and can support higher data
rates [
7
]. For example, the communication can be performed through two indepen-
dent links, namely direct and relay links. Outage event occurs when both of these
links are in deep fade, an event that is unlikely to happen. Multihop communication
can also be adopted to overcome channel fading and improve the network coverage
[
8
]. As a UE moves around, the wireless channel quality may deteriorate and at
some point the channel quality drops to a level that is unable to support the required
QoS. Or, the energy cost of the wireless link (Joule/bit) becomes too expensive,
justifying looking for an alternative energy efficient path. As such, alternative links
that are able to support the required QoS may be identified, and among them the
most energy efficient path may be replaced with the existing one. This can be done
by performing (horizontal) handover from one base station to another or vertical
handover from one RAN to another in a heterogeneous wireless network.
While cooperative communication promises energy saving for the UEs, which is
critical for users
true mobile experience, it imposes dynamic changes of commu-
nication paths between the UEs. As we referred earlier, the paths must guarantee
sufficient bandwidth to improve users
'
experience in the dynamic network envi-
ronment. Indeed, the rapid growth of video demand combined with users
'
expec-
tations for high QoS delivery has become a constant source of concern for network
operators as the network is often blamed for poor quality and video stalling. The
video stalling occurs mainly due to the lack of sufficient bandwidth guarantee on
the communication paths assigned to the services [
9
]. To this end, the Internet
Engineering Task Force (IETF) developed the Integrated Services (IntServ) [
10
]
and allowed for reserving appropriate amount of bandwidth for each service on
each network node (e.g., routers) on the communications paths to assure end-to-end
QoS delivery for the services individually. Hence, when a service terminates, the
resource reserved for it must be released for future use. The key issue here is that
resource reservation, usually resorting to the Resource Reservation Protocol
(RSVP) [
11
], involves signaling to enforce the control policies on the nodes on
the path and the related processing consumes CPU (Central Processor Unit), energy
and memory [
12
]. As it is argued in [
13
], excessive control states, signaling, long
call setup time and the related processing overhead are “Achiles
'
heel” to meet
energy efficiency, QoS and scalability targets in the future networks. Therefore,
frequent change of communication paths driven by cooperative communication
strategies deserves careful attention to prevent undue energy consumption that the
QoS enforcement processing may place throughout the network.
To reduce the performance issues raised by the IntServ signaling operations,
which are triggered upon every service request or path change, IETF introduced the
Differentiated Services (DiffServ) [
14
], as being a Class of Service (CoS) based
QoS architecture standard for the Internet. In DiffServ, the network resource is
assigned to each CoS in a static manner (e.g., a percentage of the link capacity),
which shows poor resource utilisation as traffic behaviours are dynamic and mostly
unpredictable. Therefore, resource reservation must be carried out dynamically,
taking into account network
'
s current resource conditions and changing traffic
requirements to optimize resource utilization. In this sense, IETF proposed the
aggregate resource reservation [
15
] protocol to allow for reserving more resource
'
