Cooperative Routing Management in Wireless Mesh Networks
Authors
Abstract
Wireless Mesh Networks (WMN) provide a fexible wireless backbone for ubiquitousInternet access, and are being challenged to improve their management to
support various kinds of requirements, such as scalable multimedia applications and
integration with different wireless technologies. The multi-hop structure and lowcost
solutions used by WMN make it easier to extend their services to cover larger
areas. For this reason, scalability is a critical management issue for WMN and
therefore, it is required that WMN are enabled to handle growing amounts of traffic
load and nodes in a skilful manner. In this scenario, the routing process can serve as
one of the most useful mechanisms for managing this kind of network and meeting
the requirements of multimedia applications in a more scalable way. This process
employs methods for distributed solutions, such as routing algorithms, protocols
and metrics that work together to select the best routes to enable a performance
optimization. However, several fatoras should be taken into consideration by the
routing approaches adopted to improve the WMN scalability, such as inaccurate
routing information, high routing overhead in large networks and the congestion
areas around the gateways.
The central argument of this thesis is that although there have been several routing
proposals to improve the WMN performance, the current solutions have failed
to adopt an approach that is able to handle the three main aspects of the routing
process, i.e. inaccuracy in routing information, overloaded gateways and high overhead.
In fact, all the development has been centered on a single aspect that does
not combine dierent solutions that tackle each aspect of the routing process. The
aim of this study was to demonstrate that it is possible to create a routing approach
that allows the WMN scalability to be leveraged in an eective way. In order to
achieve this goal, this work employed an architecture, called Architecture of Routing
Management (ACRoMa), that is presented using a top-down approach in which
the main components and synergies are outlined through a detailed description.
ACRoMa has been designed to provide a routing algorithm for inter-cluster load
balancing, called Routing Algorithm for Inter-cluster Load Balancing (RAILoB),
which reduces the routing overhead and avoids overload situations in gateways and
a cross-layer routing metric, called Metric for INterference and channel Diversity
(MIND), to improve the accuracy of the routing decision through the use of precise
measures to depict interference and traffic load. RAILoB speeds up the process of
load balancing between gateways (inter-cluster). Moreover, there is an interaction
between MIND and RAILoB that enables to perform intra-cluster routing decisions.Hence, RAILoB represents the ACRoMa architecture conceptually by combining all
the components in an incremental way.
Simulation models have been carried out to validate the soundness of ACRoMa
through an in-depth evaluation, and these take into account the main factors that
influence the traffic performance (e.g. topology, applications and network size).
First of all, the evaluation of MIND shows that it outperforms several cross-layer
routing metrics in dierent conguration matrices, which is evidence that the most
accurate mechanisms employed in MIND have a benecial influence on the routing decision. Following this, there was an assessment of RAILoB in different scenarios
and applications. The results showed that RAILoB achieves higher traffic performance
than the most relevant clustering load balancing routing approach in WMN,
since it provides a more flexible and agile solution for inter-cluster routing load balancing.
Thus, the ACRoMa architecture fullled its original goals, by showing that
it is possible to enhance the WMN scalability by combining solutions in the same
approach which cooperate each other.