Transcript Slide 1

CSI 2012:
47 Annual National Convention of CSI
International Conference on Intelligent Infrastructure
Program
Fluctuating Interference Tolerant
Routing Metric for Wireless Mesh
Network
Kuldip Acharya#1
Abhishek Majumder#2
Sudipta Roy*3
#Department of Computer Science &
Engineering , Tripura University
*Department of Information Technology
Assam University
1
Wireless Mesh Networks
This includes mesh routers that form an infrastructure for the clients. With
gateway functionality mesh routers can be connected to the Internet.
Infrastructure meshing
2
Problem Statement
To support multimedia applications, Quality of Service (QoS) provisioning
has become important in WMN. In this context, selection of path is a critical
factor.
In wireless network nodes access shared wireless media, the neighboring
node may compete for same bandwidth.
Various routing metrics like ETX, ETT, WCETT have been proposed.
MIC metric capture both type of interference but unable to capture sudden
and very temporary change in interference.
For this reason a new routing metric WMIC have been proposed.
3
Related work
ETX (expected transmission count)
It’s returns the average Number of MAC layer transmissions including
retransmission that is needed to successfully deliver a packet over a
wireless link.
10 packets sent to Node S from node N
8 packet successfully reached to node S from node N
LQ=0.8
S
N
Packet loss= 2/10 = 0.2 = 20 %
Probability of successful transmission from neighbor to
ourselves is 8/10 = 0.8= 80 %
4
ETX (expected transmission count)
10 packets sent to Node N from node S
6 packet successfully reached to node N from node S
S
N
NLQ=0.6
Packet loss= 4/10 = 0.4 = 40 %
Probability of successful transmission from neighbor to
ourselves is 6/10 = 0.6= 60 %
ETX= 1/ (LQ * NLQ)= 1/ 0.8*0.6 = 1/ 0.48 =2.08
5
ETT (EXPECTED TRANSMISSION TIME)
ETT returns the expected time a packet takes to successfully be sent.
0.60 ms
B
E
D
S
C
A
0.55 ms
ETT = ETX * L/R
L= Packet size
R= data rate
ETT on Route 1 = S-B-E-D = 0.70+ 0.60 + 0.65 = 1.95 ms
ETT on Route 2 = S-A-C-D = 0.60 + 0.55 + 0.60 = 1.75 ms
6
Weighted Cumulative ETT (WCETT)
Returns intra-flow interference
ETT=0.476
CH = 2
S1
ETT=0.2
CH = 1
ETT=0.2
CH = 1
A
ETT=0.5
CH = 2
S2
ETT=0.2
CH = 3
ETT=0.5
CH = 3
B
ETT=0.5
CH = 3
C
D
T
ETT=0.5
CH = 4
Fig: Non-isotonicity of WCETT
The minimum path from S1 to T is (S1, B, T)
Due to non-isotonic property of WCETT node S1 incorrectly chooses
S1,S2,C,D,T as minimum weight path.
Running dijkstra algorithm on node S1, The minimum path from S1 to
node B = S1—A---B. Though S1-A-B-T > S1-B-T
7
Weighted Cumulative ETT (WCETT)
ETT=0.476
CH = 2
S1
ETT=0.2
CH = 1
ETT=0.2
CH = 1
A
ETT=0.5
CH = 2
S2
ETT=0.2
CH = 3
ETT=0.5
CH = 3
B
ETT=0.5
CH = 3
C
D
T
ETT=0.5
CH = 4
Fig: Non-isotonicity of WCETT
This incorrect minimum weight path between S1 and T can
cause forwarding loops.
When node S2 Calculates its path to T, Dijkstra's algorithm Correctly
indicate that (S2, S1, B, T) IS THE MINIMUM PATH.
8
MIC (METRIC OF INTERFERENCE & CHANNEL SWITCHING COST)
BASED ON TWO COMPONENTS:
1. CSC ( CHANNEL SWITCHING COST )
2. IRU (INTERFERENCE AWARE RESOURCE USAGE )
• CSC from Node 1 to Node 5
NODE 1
C 1,6
W=3, C=1
NODE 3
C 1,11
NODE 2
C 1,6
W=3, C=1
NODE 5
C 1,6
NODE 4
C 1,6
W=2, C=6
W1=1, W2=13
Shortest Path 1->2->3->4->5
= 15+1+1+1= 18
9
IRU (Interference-aware Resource Usage)
D
A
C
B
T
S
E
F
An example for inter-flow interference
IRU values are estimates total amount of wasted
transmission time:
N(l) Denotes Number Of Neighbors With Which The
Transmission On Link L Interfaces.
10
Metric of Interference and Channel-switching
(MIC)
Here N is the total number of nodes in the network
min (ETT) is the smallest ETT in the network
l is the number of links in a routing path p.
i is the node.
CSC=18, IRU=6
Let, min(ETT)= 2 and N= 5, then
MIC(p)=1/5*2=1/10=0.1*(18+6)=0.1*24=2.4
11
PROPOSED METRIC:
WMIC (Weighted Metric Of Interference And
Channel- Switching )
A good link may get affected and becomes
highly fluctuated by the interference for a short
period of time.
A new metric WMIC is introduce to handle
sudden and very temporary change in
interference in a link.
12
MIC (Metric of Interference and Channel-switching)
A
C
S
D
B
F
Assume that LINK B-C is a very good link. But due to bad weather
effect link B-C suddenly get highly fluctuated for a short period of
time. At that time MIC of link B-C will found link B-C as a bed link.
13
WMIC ( Weighted Metric of Interference and Channel-switching )
Temporary interference caused by the environment will not have immediate
effect on the measured value of link quality
A
C
S
D
B
F
WMIC = (α) MIC current+ (1-α) WMIC previous
WMIC of link S-A = 0.3 * 3 + (1-0.3) * 0 = 0.9 + 0= 0.9
WMIC of link A-B = 0.3 * 4 + (1-0.3) * 0.9 = 1.2 + 0.63 =1.83
WMIC of link B-C = 0.3 * 7 + (1-0.3) * 1.83 = 2.1 +1.28= 3.38
14
WMIC Without Virtual Network
Let, link (A.B,1) < LINK (A,B,2)
However, due to reuse of channel 1 on path (A, B, 1)  (B, C, 1)
WMIC ((A,B,1)  (B,C,1)) > WMIC ((A,B,2)  (B,C,1)).
15
Virtual Network of WMIC
IRUAB (1)
0 A (1)
e
A
Bi (1)
0
IRUAB (2)
Ae (2)
w2
IRUBC(1)
Be(1)
Ci (1)
0
C-
w1
Bi (2)
Real path (A,B,1)  (B,C,1) decomposed
Real path (A,B,2)  (B,C,1) decomposed
Minimum path between Ae(1) and Ci(1)
Running Dijkstra’s algorithm A returns
16
Optimized Link State Routing Protocol
Proactive & table-driven
Each node periodically floods
status of its links
24 retransmissions to diffuse
a message up to 3 hops
Originator node
1 hop
Forwarder node
2 hop
3 hop
Receiver node
17
Optimized Link State Routing Protocol
Only selected neighbors
retransmit messages
Select MPRs such that they
cover all 2hop neighbors
11 retransmission to diffuse a message up to 3
hops
Originator node
Forwarder node
Receiver node
18
Neighbor sensing
E
A
E
D
G
B
C
Each node periodically broadcasts Hello message: List of neighbors with
bi-directional link
19
MRP selection in OLSR
Node S will select C as its MPR So all the other nodes know that
they can reach S via C
K->S route is K-C-S, whose bottleneck BW is 3
Optimal route (i.e., path with maximum bottleneck bandwidth:
K-F-S (bottleneck bandwidth of 10)
20
21
Throughput vs. Network load
As network load increases WMIC metric gives higher
throughput than ETX, ETT and MIC.
22
Delay vs. Network load
It can be observed that with the increase in network
load use of WMIC metric leads to lower delay
compared to ETX, ETT and MIC.
23
Normalized routing overhead vs. Network Load
It can be observed that WMIC has lower Routing overhead
compare to rest three, as the packet flow rate increases. The
reason behind this is that, cost incurred for route setup is
reduced.
24
Packet loss rate vs. Network load
With the increase in network load, loss rate also
increases but in case of WMIC, the packet loss rate
is the least among the rest three routing metrics.
25
CONCLUSION
Performance analysis shows that if WMIC is used for path
selection in OLSR, it gives higher throughput, lower average
delay and lower packet loss rate compared to ETX, ETT and
MIC.
26
REFERENCES
[1] Yan Zhang, Jijun Luo, Honglin Hu, B.S.Manoj and Ramesh R.Rao, Wireless
Mesh Networking: Architectures, Protocols and Standards, Auerbach
Publication, 2006.
[2] Miguel Elias M. Campista, Pedro Miguel Esposito, M.Morares, Luis
Henrique M.K.Costa,"Routing Metrics and Protocols for Wireless Mesh
Networks", IEEE Network, vol. 22, No.1, pp. 6-12, 2008
[3] Yaling Yang, Jun Wang, Robin Kravets, “Designing Routing Metrics for Mesh
Networks”, in Proc. IEEE Work shop on Wireless Mesh Networks, (invited
paper), 2005
4] Vinicius C.M. Borges, Daniel Pereira, Marilia, Curado and Edmundo
Monteiro, Routing metric for interference and channel diversity in multiradio wireless mesh networks, ADHOC-NOW, LNCS 5793, pp. 55 – 68, 2009
[5] Richard Draves, Jitendra Padhye, Brian Zil, “Routing in Multi-Radio, MultiHop Wireless Mesh Networks”, in Proc. MobiCom, pp. 114-128, 2004
[6] The network simulator ns-2, Available: http://www.isi.edu/nsnam/ns
27
[7] Pedro Miguel Esposito, Miguel Elias M. Campista, Igor M. Moraes,
“Implementing the Expected Transmission Time Metric for OLSR Wireless
Mesh Networks”, in Proc. IFIP Wireless Days Conference, pp.1-5, 2008
[8] Yaling Yang, Jun Wang, “Design Guidelines for Routing Metrics in Multihop
Wireless Mesh Networks”, in Proc. INFOCOM, pp. 1615 – 1623, 2008
[9] Ying Ge, Thomas Kunz, Louise Lamont, “Quality of Service Routing in AdHoc Networks Using OLSR”, in Proc. 36th Annual Hawaii International
Conference on System Sciences, vol. 9, pp-300.2, 2003.
[10] Yaling Yang, Jun Wang, Robin Kravets, “Interference-Aware Load
Balancing for Multihop Wireless Networks”, University Of Illions, UrbanaChampaign, Technical Report, 2005 [11] Manolis Genetzakis and Vasilios
A.Siris, “A Contention-Aware Routing Metric for Multi-Rate Multi Radio
Mesh Networks”, in proc. 5th Annual IEEE Communications Society
Conference on Sensor, Mesh and Ad Hoc Communications and Networks,
pp. 242-250, 2008
[11] Thomas Clausen, Philippe Jacquet, " Optimized Link State Routing
Protocol (OLSR) ," IETF Internet Draft , July 3 2003.
28
[12] http://hipercom.inria.fr/olsr/
THANK YOU
29