QoS Routing and Scheduling in TDMA based Wireless Mesh

Download Report

Transcript QoS Routing and Scheduling in TDMA based Wireless Mesh

QoS Routing and Scheduling
in TDMA based Wireless
Mesh Backhaul Networks
Chi-Yao Hong, Ai-Chun Pang ,and Jean-Lien C. Wu
IEEE Wireless Communications and Networking Conference, 2007.
WCNC 2007.
Mei-zhen Chen
Outline
 Introduction
 Motivation
 The
Proposed Mechanism
 Performance Evaluation
 Conclusions
Introduction
wired connections
wireless connections
wired/wireless connections
(MR)
(WMBNs)
Introduction (cont.)

Authors present an integrated mechanism for realtime applications over TDMA-based WMBNs ,and
propose QoS-aware routing and scheduling
algorithms under the integrated mechanism.
 The algorithms can work with any type of network
topologies, and the link capacity and interference are
taken into account to guarantee QoS .
Motivation

An appropriate routing algorithm needs to find one or
more feasible routes for each pair




interference
• signal-to-interference-and-noise ratio (SINR)
Selected routed paths
• a linear-programming optimization technique
The routes shall not be restricted by a particular topology
The actual traffic demand of an MR varies as depending on
application needs of its serving clients
• a dynamic routing strategy
The Proposed Mechanism
-assumptions (1/4)

Consider a single-channel single-transceiver MAC.
 There is only one gateway in the network.
 Packet loss is assumed to be neglected.
 The TDD (time division duplexing) mode is chosen
here for uplink and downlink transmissions.

the traffic demand of an MR v, D[v]
The Proposed Mechanism
-assumptions (2/4)


The IEEE 802.16 standard provides multiple transmission
rates by applying different modulation schemes and coding
rates.
The receiver minimum input level sensitivity, RSS
The receiver
signal-to-noise
ratio (SNR)
The sampling
frequency in MHz
The number of
OFDM subcarriers
2
log2  Nused 
BW is the channel bandwidth and η is a constant sampling factor.
The number of
allocated
subchannels
The Proposed Mechanism
-assumptions (3/4)

On the other hand, RSS is also given by
where EIRP is the effective isotropically-radiated power
Gt(dBi) : transmitter antenna gain.
Gr(dBi) : receiver antenna gain.
SLt : transmitter loss
SLr : receiver loss.
TxPower(dBW) : the transmission power.
LM(dB) : the link margin.
PL(dB): the transmission path loss.
The Proposed Mechanism
-assumptions (4/4)

The MRs in WMBNs are assumed to be stationary
and communicate without obstructions, the fading
effects are negligible.
 Adopt two-ray ground reflection model as our path
loss model.
The Proposed Mechanism
gateway
-flowchart
The Proposed Mechanism
-routing algorithm (1/4)

Authors slightly modify the Dijkstra’s algorithm to
find a path from the gateway to a particular MR.
 the weights of wireless links could be changed as
the alteration of network flows.
 the modified algorithm terminates immediately
after the path from the gateway to the target MR is
discovered
The Proposed Mechanism
-routing algorithm (2/4)


A mesh network is modeled as an undirected graph G = (V,E)
 V : MRs
 E : wireless links
E[G] : set of the wireless links.
 R[i, j] : the bit rate of link (i, j) derived.
 B[i, j] : the available bandwidth of link (i, j).
 D[v] : the traffic demand of router v.
 I[v] : the granted bandwidth of router v. (initially set by 0)
 U[G] : a subset of the nodes whose demands have not been
satisfied.
 P[G] : a subset of the wireless links along the selected path.
The Proposed Mechanism
-routing algorithm (3/4)
W[i, j] : the occupancy
ratio of a link (i, j).
S[m, n] : the subtracted
bandwidth quantity of
link (m, n) interfered by
link (i, j).
The Proposed Mechanism
-routing algorithm (4/4)

to maximize the achievable traffic demand of router v,
D’[v]
The Proposed Mechanism
-scheduling algorithm

F[i, j] : whether the demand of link (i, j) is already
scheduled in the current schedule period

S[i, j] : whether link (i, j) is interfered by any other
active links
The Proposed Mechanism
-scheduling algorithm (cont.)
extract nonblocking links
Q[G]: the subset of
the links whose
demands are satisfied.
Performance Evaluation (1/5)





20MRs are randomly distributed
25 × 25 km2 area
A gateway is placed in the center of the area
The packet overheads conform to the RTP, UDP, IP
and IEEE 802.16 packet formats.
The event-driven simulation is implemented by C++
programs.
Performance Evaluation (2/5)
Performance Evaluation (3/5)
Performance Evaluation (4/5)
Performance Evaluation (5/5)
the mean traffic demand: 150Kbps
Conclusions

Authors considered the QoS problem for routing and
scheduling in TDMA-based WMBNs.
 Authors proposed an integrated routing and
scheduling mechanism to provides QoS guarantees.
 A linear programming optimization was devised to
solve the amount of non-collision bandwidth.