05-02-2013, 11:33 AM
Capacity-Aware Routing in Multi-Channel Multi-Rate Wireless Mesh Networks
1Capacity-Aware Routing.pdf (Size: 321.15 KB / Downloads: 45)
Abstract
Employing multiple channels in wireless multi-hop
networks is regarded as an effective approach to increasing
network capacity. However, existing routing protocols may not be
able to properly utilize the advantages of multiple channels in
such networks. In this paper, we focus on IEEE 802.11-based
wireless mesh networks with stationary nodes, such as wireless
backhaul networks and community wireless networks. We
propose a new path metric called Bottleneck Link Capacity (BLC)
which accounts for the link quality, the interference among links,
and the traffic load on the links. Then, we develop a routing
protocol called Capacity-Aware Routing (CAR) which makes use
of BLC as the routing metric. Finally, we evaluate the
performance of BLC via simulations. The results show that our
path metric outperforms others in terms of system throughput
and end-to-end delay.
INTRODUCTION
Wireless mesh networks have received much attention in
recent years thanks to their desirable features, such as low
up-front cost, easy network maintenance, robustness, and
reliable service coverage [1-2]. However, in such networks,
the system capacity degrades seriously as the network grows
[3-7]. Employing multiple channels has been shown as an
effective approach to increasing the system capacity [8-11].
However, existing protocols must be redesigned properly to
take advantage of channel diversity.
The key parameter in routing protocols is the routing metric.
De Couto et al. propose a routing metric called “Expected
Transmission Count” (ETX) [12]. This metric is based on the
loss rate of broadcast packets on each link. ETX finds a path
with minimum expected number of packet transmissions
(including retransmissions) required to successfully deliver a
packet to the ultimate destination. However, when there is a tie
on the ETX values, no preference is given to the path with
higher channel-diversity. Thus, it is not suitable for
multi-channel systems even though it outperforms shortest
path routing. Richard Draves et al.
Estimating Residual Air Time
Let each node maintain two tables, namely, BusyPeriod and
ResidualAirTime. Let BusyPeriodi and ResidualAirTimei
denote these two tables at node i, respectively. BusyPeriodi[n]
specifies the aggregate length of busy periods on channel n
which is observed by node i during a predefined period Tm,
where Tm is referred to as the measurement period in this paper.
Note that a busy period includes all transmission or receiving
periods on the link over a measurement period. In addition, the
period reserved by NAV is also counted in busy periods. Let
BusyPeriod[n]c denote the latest measurement of busy periods
on channel n. We update the historical value of BusyPeriod[n]
by employing the exponential weighted averaging technique
as shown in (1).
Capacity-Aware Routing (CAR) Protocol
CAR is a routing protocol for IEEE 802.11-based
multi-channel wireless mesh networks. Like Ad-hoc
On-demand Distance Vector (AODV) routing [18] and
Dynamic Source Routing (DSR) [19], CAR builds routes
on-demand. Furthermore, CAR discovers paths on a per flow
basis. When a route is required, the source floods the ROUTE
REQUEST (RREQ) packet on the control channel to discover a
route. The RREQ packet carries the load information (the
channel map and the ResidualAirTime table) at the previous
hop and the partial path information discovered so far.
CONCLUSION
In this paper, we tackle the routing problem in IEEE
802.11-based multi-channel wireless mesh networks. We
assume nodes are stationary and each of them is equipped with
one or more NICs. We present an on-demand routing protocol
called Capacity-Aware Routing (CAR). CAR uses the
Bottleneck Link Capacity (BLC) as the route selection criteria.
The computation of the BLC metric considers the interference
among links, the link quality, and the traffic load on the links.
The goal of employing the BLC metric is to select the path
which has the highest residual capacity.