Beruflich Dokumente
Kultur Dokumente
Deborah Estrin
IV-1
1. Collision avoidance
2. Energy efficiency
3. Scalability in node density
4. Latency
5. Fairness
6. Throughput
7. Bandwidth utilization
IV-3
(Joules/Node/Received Event)
0.14 0.018
Average Dissipated Energy
Diffusion
Average Dissipated Energy
IV-4
Identifying the Energy Consumers
15
Power (mW)
10
0
CPU TX RX IDLE SLEEP
SENSORS
RADIO
• E TX ≈ E RX ≈ E IDLE >> E SLEEP
• Need to shutdown the radio
IV-6
Sensor-MAC (S-MAC) Design
(Wei et al. 2002)
• Tradeoffs
Latency
Energy
Fairness
IV-7
IV-8
Periodic Listen and Sleep
IV-9
• Schedule maintenance
– Remember neighbors’ schedules
— to know when to send to them
– Each node broadcasts its schedule every few periods
– Refresh on neighbor’s schedule when receiving an
update
– Schedule packets also serve as beacons for new
nodes to join a neighborhood
IV-10
Collision Avoidance
IV-11
Overhearing Avoidance
IV-12
Message Passing
Fairness Energy
Msg-level latency
IV-13
• 802.11
RTS 3 Data 3 Data 3 ... Data 1
CTS 2 ACK 2 ACK 2 ... ACK 0
• Platform
• Motes (UC Berkeley) :
• 8-bit CPU at 4MHz,
• 8KB flash, 512B RAM
• TinyOS: event-driven
• Also used as NIC for 32-bit embedded
PCs
IV-15
1400
1000
• Each source node sends 800
10 messages
600
— Each message has 10 400
fragments x 40B
200
• Measure total energy
— Data + control + idle
0 2 4 6 8 10
Message inter-arrival period (second)
Message Inter-arrival period
IV-16
Adaptive Topology
IV-17
IV-18
Connectivity Measurements*
*An Empirical Study of Epidemic Algorithms in Large Scale Multihop Wireless Networks
Ganesan, Krishnamachari, Woo, Culler, Estrin and Wicker, UCLA/CSD-TR 02-0013.
IV-19
Tradeoff
• How many nodes to activate?
– few active nodes:
• distance between neighboring nodes high -> increase packet loss and
higher transmit power and reduced spatial reuse;
• need to maintain sensing coverage (see earlier session on
coverage/exposure)
IV-20
Adaptive Topology Schemes
IV-21
Help Neighbor
Messages Announcements Data
Data Message Messages Message
IV-22
State Transitions
after Tt
Test Active
neighbors < NT
neighbors > NT (high ID for ties);
and
or
• loss > LT
loss > loss T0
• loss < LT & help
after Tp
Passive Sleep
after Ts
NT: neighbor threshold
LT: loss threshold
T?: state timer values (p: passive, s: sleep, t: test)
IV-23
Testbed Platforms
• Tiered architecture: heterogeneous set of platforms (no particular
set of compromises with a unique platform).
• PC-104/iPAQ enable running tasks that are more CPU/memory
intensive.
IV-24
Energy Savings Ratio
Energy Savings Ratio (normalized to the Active case, all nodes turn on) as a
function of density. ASCENT provides significant amount of energy savings,
with a factor of 5 for high density scenarios.
IV-25
Open question: to what extent would SMAC or Geo-routing produce the same result?
IV-26
Goal Routing Assumptions
(general: energy dependency
savings)
SPAN preserve capacity of the gets connectivity matrix and 802.11 MAC with Power
raw topology neighbors from routing Savings mode
requires modifications in the
routing lookup process
STEM tradeoff latency for needs routing info to direct the 2 radios/wake-up channel
energy savings wake-up wave connectivity conditions
remain constant in sleeping
periods
IV-27
sensor network
IV-28
STEM: Sparse Topology and Energy Management
Wakeup plane: f1
Data plane: f2
IV-29
Initiator node
f1
B1
B2
1. beacon received
Train of beacon packets
TRx
2. beacon acknowledge
T
f1
Target node
IV-30
Performance Metrics
E
T S (sec) E0
T (sec) T (sec)
IV-31
Energy vs Latency
E
50%
E0
10%
1%
0. 1%
TS (sec)
IV-32
STEM Design Decisions
D
Sleep mode
IV-33
1 initiator node
IV-34
Exploiting Latency AND Density
Zzz
Active nodes
Zzz
Zzz
STEM
IV-35
Performance of STEM+GAF
E
E0
STEM GAF
alone
avg # of neighbors
IV-36
Conclusions
IV-37
• Load Balancing.
• Larger scale experiments.
• Interaction with adaptive MAC and geographic
routing
• Application defined Adaptive Fidelity
– Expanding on STEM’s data driven characteristics to
achieve more than on/off behavior
IV-38
Routing
IV-39
GRAB:
Field Based Minimum Cost Forwarding
(Lu et al 2002)
IV-40
ADV Dissemination Example
IV-41
IV-42
Data Dissemination
IV-43
IV-44
Data Dissemination Example contd.
IV-45
Routing on a Curve
(Nath et al 2002)
IV-46
TBF (Trajectory based forwarding)
• Fundamental Idea
– Route packets along a specified trajectory
• Generalization of Source Based Routing and Cartesian
routing
• Trajectory specified in the packet
IV-47
Specifying trajectory
y = ax + b, y = sin x
• Function
• Equation axx+= by
t , y += c
at =
+ b1;, xx= + y 2t , +y 1= r cos t
2
r sin
• Parametric
– we use this choice
IV-48
Features of TBF
• Basic Features
– Decouples pathname from the actual path
• Source based Routing (LSR, DSR etc) mixes
naming and route path
• Applications:
– Route around obstacles/changes/failures
– Trajectory forwarding need not have a “destination”
• Route along a line, pattern
• Applications:
– Flooding, discovery, group communication (pollination)
IV-49
Routing on a curve
IV-50
Spoke flooding
IV-51
100
90
80
70
60
Coverage
50
Packets
40
30
20
10
0
4 9
Number of Spokes
IV-52
Related Work
• MANET
– Reactive[DSR], proactive[AODV], TORA,
GPSR[KarpKung00]
• Location-aided routing
– Geocast[Navas97],Cartesian-LAR, [KOVaidya98]
• Localization techniques
– GPS, Cricket[Priyantha00],RADAR[Bahl00],
– Ahlos[bulusu00,Savides01],TdoA technique, Visual cues
IV-53
Discussion
• Trajectory modifications
• No forward progress
– Retry from source
– Local reverse gear
– Apply detour patches
– Packets allowed to backtrack
• Specifying trajectories
– Accurate trajectories
– Use fourier components
– Tradeoff between number of components and
accuracy
IV-54
In-Network Processing
The Key to Sensor Network scalability and Realization
IV-55
IV-57
IV-58
A more general look at
Data Centric vs. Address Centric
approach
(Krishnamachari et al.)
• Address Centric
• Distinct paths from each source to sink.
• Data Centric
• Support aggregation in the network where paths/trees overlap
• Essential difference from traditional IP networking
IV-59
IV-60
Comparison of energy costs
Address Centric
Shortest path data centric
Greedy tree data centric
Nearest source data centric
Lower Bound
IV-61
IV-63
Examples
• Map isotherms and other “contours”, gradients, regions
– Record images wherever acoustic signatures indicate significantly
above-average species activity, and return with data on soil and air
temperature and chemistry in vicinity of activity.
– Mobilize robotic sample collector to region where soil chemistry and
air chemistry have followed a particular temporal pattern and where
the region presents different data than neighboring regions.
IV-64
Why databases?
IV-65
Database-centric Approach
(Muntz et al.)
• Query processing
– Storage structures:
• Less choice here … traditionally, data organized on disk
• Here, data originates at the sensors’ location
– Time coherency across space
– “Stream” processing … one pass over data
• E.g., try to estimate a histogram on the fly.
– Quality of service:
• Latency vs. completeness
• Actuators: e.g., camera pan and zoom
– Tradeoffs in communication vs. computation and storage:
e.g., filter-refine: where to do what.
IV-66
Differences between databases and sensor
networks
IV-67
IV-68
Traditional Approach: Warehousing
Warehouse
Front-end
Sensor Nodes
IV-69
Sensor Sensor
DB DB
Sensor
DB
Sensor Sensor
Front-end DB DB
Sensor
Sensor DB Sensor
DB DB
IV-70
Sensor Nodes
Sensor Database System
IV-71
Performance Metrics
• High accuracy
– Distance between ideal answer and actual answer?
– Ratio of sensors participating in answer?
• Low latency
– Time between data is generated on sensors and answer is
returned
• Limited resource usage
– Energy consumption
IV-72
Representing Sensor Data and Sensor Queries
• Sensor Data:
– Output of signal processing functions
• Time Stamped values produced over a given duration
– Inherently distributed
• Sensor Queries
– Conditions on time and space
• Location dependent queries
• Constraints on time stamps or aggregates over time windows
– Event notification
IV-73
IV-74
Fjording the Stream: An Architecture for Queries
over Streaming Sensor Data
(Hellerstein et al.)
IV-75
IV-76
Hardware Architecture
• Centralized data
processing.
• Sensor proxies read and
configure sensors.
• Query processor interacts
with proxies to request and
get sensor data.
• Sensor proxies support
multiple simultaneous
queries, multiplexing the
data.
IV-77
Operators
IV-78
Sensor Sensitive Operators
IV-79
Sensor Proxy
IV-80
Building a Fjord
• For all sensor data sources, locate the proxy for the
sensor, and install a query on it to deliver tuples at a
certain rate to a push queue.
• For non-sensor data sources, set up a pull queue to
scan for data.
• Pipe the data through the operators specified by the
query.
IV-81
Query
• Find average car speeds during time window (w), for all
segments the user is interested in (knownSegments)
• More complicated queries are possible, with joins of streaming
sensor data and historical data stored in a normal database
fashion.
IV-82
Dataflow for the Query
• Data is pushed from the sensors to the user, through the filter
operator set up by the query.
• Multiple similar queries can be added to an existing fjord,
instead of creating one per query.
IV-83
Conclusion
IV-84
The Sylph Middleware
IV-85
IV-86
Sensor Module
IV-87
• Device Proxy
• Common Service Discovery Mechanisms (e.g., Jini)
• Soft-State Reliability
IV-88
Proxy Core
• Device Manager
• Sensor Proxy
– Service discovery
– Layered functionality (e.g., buffering)
• Query Processor
– Query parsing
– Directed graph of stream operators
– Query optimization
• Query Distribution - Gateways
IV-89
IV-90
At A Glance …
IV-91
IV-92
The Tiny Aggregation (TAG) Approach
(S. Madden, UCB)
IV-93
IV-94
Aggregation Functions
(Madden)
Example: Average
AVGmerge {<S1, C1>, <S2, C2>} → < S1 + S2 , C1 + C2>
AVGinit{v} → <v,1>
AVGevaluate{<S1, C1>} → S1/C1
IV-95
Query Propagation
(Madden)
IV-97
Simulation Result
(Madden, OSDI ‘02)
Some aggregates
Neighbors = ~20 80000
Total B y tes Xm itte d
70000
60000 require dramatically
50000 more state!
40000
30000
20000
10000
0
EXTERNAL MAX AVERAGE COUNT MEDIAN
Aggregation Function
IV-98
Taxonomy of Aggregates
(Madden)
IV-99
IV-100
SensorWare
IV-101
SensorWare Architecture
External user can Message
inject agent script exchanging
OS OS
Magnetic Sensor
WaveLan Card - uC Based with RS232
-IEEE 802.11b Compliant - Range of +/- 2Gausus
- 11 Mbit/s Data Rate -Adjustable Sampling Rate
- X, Y, Z output
- Device ID Management
IV-103
User Node
Script Migration
Mobile Script
- Light weight
- Programmable
Script Migration - Code Mobility
- Tcl front end
IV-104
Event-driven Model
IV-105
IV-106
Tracking: Mobile Script Flooding
User Node
“Tracking Script
Code” injecting
Sensor Node
Monitoring
Target
Video IV-107
Node
User Node
Event Sensing
Event Sensing
Event Sensing
Event Sensing
Event Sensing
Event Sensing
Sensor Node
Event Sensing
Video IV-108
Node
Tracking: Mobile Script Activation
User Node
Tracking Code
Activated
Tracking
Code
Activated
Target
Sensor Node
Video IV-109
Node
User Node
Tracking
Script Migration
Tracking
Target Position Information
Sensor Node
Script Migration
Monitoring
Position Information
Video IV-110
Node
Tracking: Position Notification
User Node
Tracking
Tracking
Script Migration
Target
Sensor Node
Tracking Monitoring
Script Migration
Position Information
Video IV-111
Node
60%
– Rapid reprogramming 40%
– Message receive and send 20%
contexts 0%
0 20 40 60 80 100 120 140 160 180 200 220 240
IV-112
Moving Forward:
Storage, Representation, Distributed processing
• Spatio-temporal Queries
– Index data for easy temporal and spatial
searching
K V K V • Interpretation of Spatially Distributed Data
K V – Provide ability to interpret data at different spatial
K V
K V and temporal extents. Per-node processing
K V K V
alone is not enough
K V
K V
• Pattern-Triggered Data Collection
K V
K V
– Should enable multi-resolution data storage and
retrieval for data collection.
• Data Centric Protocols and In network
Processing
– Network does in-network processing of data
based on distribution of data
– When data changes, portion of the network
updates understanding of the data
– When user queries, it travels to node that
maintains relevant summarized data
IV-113
summarize data
• Queries for statistics
Sensor Value
• Show degree of data
variation in an area
• Queries for maps
• Geographically show data
variation
Sensor Value
• Provide approximation of
map
ion
cat
Lo
Y
X Location
IV-114
DIMENSIONSTM: A Multi-Resolution Storage
Architectures for Sensor Networks (Ganesan et al)
IV-115
• For level=1 to N
– Cluster-Heads (i-1) send data to Cluster-Head (i)
– ClusterHead(i) aggregates data from 4 lower level clusterheads,
locally stores the deltas, and sends resulting coefficients to
ClusterHead(i+1)
• To a first order approximation, size of coefficients transmitted at each
level is constant.
IV-116
Content Addressable Storage
…aka Data Centric Storage
(Ramiswamy et al)
IV-117
8 < Temperature ≤ 8
IV-118
Index Construction
Indexes contain histograms summarizing
1 < Temperature ≤ 8
events within their value and spatial ranges 4 < Temperature ≤ 8
Event
Attributes 7 < Temperature ≤ 8
Name = EventLongitude
Value = 120
8 < Temperature ≤ 8
Attributes Coordinates
(120, 171)
Name = EventLatitude (0111 1000, 1010 1011)
Value = 171
Hash String
Attribute
Name = Temperature “Temperature 1 8”
Hash Coordinates
(121, 169)
Value = 8
Min = 1
Level
Function (0111 1001, 1010 1001)
Max = 8
Distribution = Uniform 1 of 4
IV-119
4 < Temperature ≤ 8
1 < Temperature ≤ 8
Search
• Search most constrained index nodes
for region of interest
• Search travels from constrained
semantics to constrained geography
7 < Temperature ≤ 8
(location of an event)
8 < Temperature ≤ 8
• Distributions are generated as part of
the index construction
Node Histogram Node Histogram Node Histogram
< 8 -- 752 < 7 -- 700 < 5 -- 400
> 8 -- 0 > 8 -- 0 > 8 -- 0
Region 00
Region 00 Region 00 Region 00
7 52
80
82 5 -- 100 5 -- 30
6 -- 80 6 -- 20
Region 01 7 -- 26 7 -- 13
Region 01 8 -- 1 8 -- 1
7 12
80
80 etc.
Region 10
Region 10 Region 10
7 15 Region 10
82
80 5 -- 20 5 --15
6 -- 20 6 -- 15
Region 11 7 -- 9 7 -- 4
Region 11 8 -- 0 8 -- 1
70
80
80
IV-120