CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department...
-
Upload
caroline-mosley -
Category
Documents
-
view
217 -
download
0
description
Transcript of CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department...
![Page 1: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/1.jpg)
CENS, April 20 2007 1
Modeling Wireless Sensor Networks
Bhaskar KrishnamachariMing Hsieh Department of Electrical Engineering
USC Viterbi School of Engineering
![Page 2: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/2.jpg)
2
Overview
• Mathematical modeling provides fundamental insights into:
1. Link layer behavior
2. Protocol design
3. Scaling and architecture
![Page 3: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/3.jpg)
3
1. Modeling Link Layer Behaviorin Low Power Wireless Networks
Marco Zuniga, Bhaskar Krishnamachari, "An Analysis of Unreliability and Asymmetry in Low-Power Wireless Links", ACM Transactions on Sensor Networks, accepted to appear, 2007.
Dongjin Son, Bhaskar Krishnamachari, John Heidemann, “Experimental Analysis of Concurrent Packet Transmissions in Low-Power Wireless Networks”, Sensys ’06 + Ongoing work
![Page 4: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/4.jpg)
4
• Two simplified models form the basis of >95% of the literature on wireless networks:
X
Circular radio range with perfect reception within &zero reception outside
Collision with simultaneous transmissions within range
![Page 5: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/5.jpg)
5
Link Quality Variation with Distance
From Wooet al. ‘03
![Page 6: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/6.jpg)
6
An Explanatory Model• Basic idea: compose the following two functions (a) SNR
versus distance with (b) PRR versus SNR
Marco Zuniga, Bhaskar Krishnamachari, "An Analysis of Unreliability and Asymmetry in Low-Power Wireless Links", ACM Transactions on Sensor Networks, accepted to appear, 2007.
![Page 7: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/7.jpg)
7
Bimodal PRR Distribution• A majority of the links are either good (above 90%) or bad (below 10%), matching empirical findings (e.g., Cerpa et al. ’05)
![Page 8: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/8.jpg)
8
Expectation and Varianceof Packet Reception Rate
Justifies the presence of “long links”
![Page 9: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/9.jpg)
9
• Models Incorporated into simulators:– TOSSIM (Berkeley)– Castalia (NICTA, Australia)
• Standalone code at http://ceng.usc.edu/~anrg/downloads.html
![Page 10: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/10.jpg)
10
X
Conservative protocol assumption: always a collision
Concurrent Transmissions
Reality: SINR makes the difference
Son, Krishnamachari, Heidemann, “Experimental Analysis of Concurrent Packet Transmissions in Low-Power Wireless Networks”, Sensys ‘06
![Page 11: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/11.jpg)
11
SINR-view of Interference
)2(8)exp5.01( 10 lfSINRPRR
![Page 12: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/12.jpg)
12
Feasibility of Concurrent Transmissions
P1g11/(P2g21+N) ≥
P2g22/(P1g12+N) ≥
S1 R1
R2S2
g11g12
g21
g22
![Page 13: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/13.jpg)
13
![Page 14: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/14.jpg)
14
Linear Topology Case
Counter-intuitive“embedding” of simultaneousconversations
S1 R1 R2S2
![Page 15: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/15.jpg)
15
2. MAC Design for ScalableData Collection
Kiran Yedavalli, Bhaskar Krishnamachari, "Enhancement of the IEEE 802.15.4 MAC Protocol for Scalable Data Collection in Dense Sensor Networks", USC Computer Engineering Technical Report CENG-2006-14, November 2006.
![Page 16: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/16.jpg)
16
State of the Art: IEEE 802.15.4
• Specifies both PHY and MAC layers for low-power, low-rate embedded wireless networks.
• The MAC protocol is a slotted CSMA with binary exponential back-off
• 256 nodes allowed by standard
![Page 17: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/17.jpg)
17
p-persistent CSMA Model
…
: Idle Slot
: Collision Slot
: Successful Slot
epoch
…… … ……
![Page 18: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/18.jpg)
18
Delay and Energy Expressions
1. Average expected epoch delay
2. Average expected epoch energy consumption
ξR: Energy Consumption per node per time slot in the Receive State
ξT: Energy Consumption per node per time slot in the Transmit State
1
( 1)(1 )[ ](1 )
n
n n
L L pE Tnp p
1
2 1
( 1)(1 )[ ](1 ) (1 )
n
n R Tn n
L L p LE Ep p p
![Page 19: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/19.jpg)
19
Optimality• Delay
• Energy2
1 , 1( , )
2 ( 1)( 1), 1
( 1)( 1)
Topt
Ln
p n Ln n n L n
Ln n L
2 2 ( 1)( 1) 4 ( 1)( 1)( , ) ,
( 1)( 1) 2 ( 1)( 1)E T
optR
n n n L L n np n L
n n L L n
If ξR = ξT, the same transmission probability optimizes both delay and energy simultaneously.
![Page 20: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/20.jpg)
20
A Useful Optimality Criterion
0 10 20 30 40 50 60 70 80 90 1000.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5
Number of nodes in an epoch (n)
E[T
Idle
,n]/E
[Tn]
L = 5
poptT (n,L)
poptT (n,L) - 0.003
poptT (n,L) + 0.003
,[ ] 2 1, , ( , ) , ,[ ] ( 1)( 2 1 1)Idle nT
optn
E T L Lp p n LE T L L
When the number of contending nodes is high, this provides sensitive feedback that can be used to adapt the access rate
![Page 21: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/21.jpg)
21
Receiver Feedback Enhancement
• Receiver performs measurement and broadcast
• Window update rule:
• All contending nodes change the window size simultaneously
,
2 1( 1)( 1)( 2 1 1)
,Current
Idle Currentnext
Current
L LWTL L
WT
![Page 22: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/22.jpg)
22
Results
50 55 60 65 70 75 80 85 90 95 1000
20
40
60
80
Number of Contending Nodes
Thro
ughp
ut (K
bps)
Packet Length = 50 Bytes
IEEE 802.15.4Enhanced IEEE 802.15.4
50 55 60 65 70 75 80 85 90 95 1000
2
4
6
8
Number of Contending Nodes
Ene
rgy
(mJo
ules
) IEEE 802.15.4Enhanced IEEE 802.15.4
![Page 23: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/23.jpg)
23
3. Fair and Efficient Rate Control for Data Gathering
Avinash Sridharan and Bhaskar Krishnamachari, "Maximizing Network Utilization with Max-Min Fairness in Wireless Sensor Networks," to be presented at 5th Intl. Symposium on Modeling and Optimization in Mobile, Ad Hoc, and Wireless Networks (WiOpt), April 2007.
![Page 24: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/24.jpg)
24
Problem Formulation• Allocate rates to each
source to (a) ensure fairness, and (b) efficient use of available bandwidth.
• Closely related prior work by Rangwala et al. SIGCOMM ’06 – focuses primarily on fairness and proposes a TCP-like AIMD mechanism
![Page 25: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/25.jpg)
25
• Receiver capacity interference model – source rates from node’s sub-tree and its interfering neighbors’ sub-trees must not exceed available bandwidth
Problem Formulation
![Page 26: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/26.jpg)
26
Validating Capacity Model
![Page 27: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/27.jpg)
27
• Bottleneck rate turns out to be the minimum supply/demand ratio:
• This can be calculated easily given the tree, interference graph, and receiver bandwidths
P1: Solving for Fairness
![Page 28: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/28.jpg)
28
P2: Solving for Efficiency
• Duality-based approach based on the classic work on optimization flow control by Low & Lapsley ’99
• Introduce new dual variables (shadow prices) that weigh resource constraints
• Yields distributed algorithms with market auction interpretation
![Page 29: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/29.jpg)
29
Structure of P2’s Lagrange Dual
• Each router sets a price for its bandwidth
• The rate for each source depends on sum-price of routers affected by its flow
sum-price
![Page 30: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/30.jpg)
30
![Page 31: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/31.jpg)
31
• Increment the shadow prices in the direction of the negative sub-gradient (determined by source rates)
• Choose source-rates to maximize component function (determined by shadow prices)
• In general, this could be a very slow iterative process…
Subgradient Optimization
![Page 32: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/32.jpg)
32
Good News
• Numerical evidence: setting all shadow prices to 1 provides near-optimal solutions in one iteration!
![Page 33: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/33.jpg)
33
Resulting Heuristic
1. First determine and allocate min rate to all sources
2. Give rank to each source that is inversely proportional to the number of downstream receivers whose bandwidth it consumes;
3. Allocate saturating rates to flows, in rank order
![Page 34: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/34.jpg)
34
Simulation Results
CDF of difference from optimal solution
![Page 35: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/35.jpg)
35
Ongoing Work
• Test-bed Implementation
• Cross-layer extensions
![Page 36: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/36.jpg)
36
4. Fundamental Scaling Lawsfor Store and Query Sensor Networks
Joon Ahn and Bhaskar Krishnamachari, "Fundamental Scaling Laws for Energy-Efficient Storage and Querying in Wireless Sensor Networks", ACM MobiHoc, May 2006.
![Page 37: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/37.jpg)
37
• Race between increasing supply and demand:- Energy and storage- Application-specific event and query traffic
• The winner of this race determines scalability.
In a Nutshell
![Page 38: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/38.jpg)
38
• N nodes deployed in a 2D area with constant density for time T
• m atomic events and qi queries for the ith event, all uniformly distributed
• Can create ri replicas for event i to reduce search cost (at the expense of increased replication cost)
Preliminaries
![Page 39: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/39.jpg)
39
Data-Centric Querying Approaches
• Unstructured: expanding ring searches, random walks.
• Structured: Geographic Hash Table, DIFS, DIM
![Page 40: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/40.jpg)
40
Energy Cost Scaling
• Creplication = c1
r : # of copies of an event
N : # of nodes
• Csearch(unstructured) = c2 • Csearch(structured) = c3
EVENTEVENT REPLICATIONUNSTRUCTURED QUERYSTRUCTURED QUERY
![Page 41: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/41.jpg)
41
Energy Optimization Formulation
S : total storage sizem : the total number of eventsqi : the query rate for ith eventri : the number of copies of ith event
Cs(ri) : the expected minimum search cost of ith event
Cr(ri) : the expected replication cost of ith event
Cr(r) = c1 Cs(r) = c2
![Page 42: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/42.jpg)
42
Optimization Solution
Minimizer
The Optimized Total Cost
(inactive constraint)
(active constraint)
qi : # of queries for event i
N : # of nodesS : total storage
sizem : # of events
![Page 43: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/43.jpg)
43
Optimal Total Cost
Simplified, assuming : q : # of queries per event
N : # of nodesS : total storage
sizem : # of eventsif
if
![Page 44: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/44.jpg)
44
Illustration of Energy Scaling
m : # of eventsq : # of queries
per event
![Page 45: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/45.jpg)
45
I - Storage and Energy Scalability Results
Energy ConditionThe energy requirement per node is boundedif and only if mq1/2 = O(N1/4)
Energy constraint is stricter than storage constraint
m : # of eventsq : # of queries per eventN : # of nodes
Storage ConditionA network scales efficiently with bounded storage per node
if mq1/2 = o(N3/4)
![Page 46: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/46.jpg)
46
II - Fixed Energy Budget Results
S – successful operation region
N : # of nodese: per-node energy budget
![Page 47: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/47.jpg)
47
III - Network Lifetime Scaling Results
Network Lifetime as a function of Network Size
![Page 48: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/48.jpg)
48
Summary• Only certain classes of applications can be sustained in arbitrarily
large sensor networks.
• Specifically, if mq1/2 = O(N1/4) for unstructured networks, and mq2/3 = O(N1/2) for structured networks:
a. The network can operate with bounded energy and storage per node.
b. The network lifetime does not decrease with network size for a given energy budget.
• The results can be reinterpreted to understand how to tier sensor networks into zones with localized queries
• These results generalize in a straightforward manner to 1D and 3D deployments. 3D deployments are inherently more scalable.
![Page 49: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/49.jpg)
49
Final Thoughts
“In theory, theory and practice are the same; in practice, they’re different.”
![Page 50: CENS, April 20 20071 Modeling Wireless Sensor Networks Bhaskar Krishnamachari Ming Hsieh Department of Electrical Engineering USC Viterbi School of Engineering.](https://reader036.fdocuments.us/reader036/viewer/2022062412/5a4d1ae47f8b9ab0599785f1/html5/thumbnails/50.jpg)
50
Thanks