Querying the Physical World

1
Querying the Physical World

• Joe Hellerstein
• UC Berkeley Intel Research Berkeley
• Wei Hong
• Intel Research Berkeley
• Sam Madden
• M.I.T.

2
Motivation

• Sensor networks (aka sensor webs, emnets) are
  here
• Several widely deployed HW/SW platforms
• Low power radio, small processor, RAM/Flash
• Variety of (novel) applications scientific,
  industrial, commercial
• Great experimental platform for ubicomp
  mobility
• Cross-cutting technical challenges
• Networking, systems, languages, databases, AI,
  stats, signal proc.
• We will summarize
• The state of the art
• Our experiences building TinyDB
• Current and future research directions

Berkeley Mote
3
Some Sensornet Apps
smart cooling in data centers (hp/intel)
redwood forest microclimate monitoring
http//www.hpl.hp.com/research/dca/smart_cooling/
And More
condition-based maintenance (intel/BP)

• Homeland security
• Container monitoring
• Mobile environmental apps
• Bird tracking
• Zebranet
• Home automation
• Etc!

structural integrity (ucb/ggbd)
4
Context The Berkeley Stack
5
Declarative Queries

• Programming TinyOS Apps is Hard
• Embedded x Distributed Programming!
• Limited power budget
• Lossy, low bandwidth communication
• Require long-lived, zero admin deployments
• Distributed Algorithms
• Limited tools, debugging interfaces
• Queries abstract away much of the complexity
• Burden on the query engine developer
• Users get
• Safe, optimizable programs
• Freedom to think about apps instead of details

6
TinyDB Query Engine

• Continuous variant of SQL TinySQL
• Power and data-acquisition based in-network
  optimization framework
• Extensible interface for aggregates, new types of
  sensors

7
Agenda

• Sensornet Background
• TinyDB
• Data Model and Query Language
• Software Architecture
• TASK
• Quick overview
• A Flavor of Research Challenges

8
Sensor Networks a hot topic

• New university courses
• New conferences
• ACM SenSys, IEEE IPSN, etc.
• New industrial research lab projects
• Intel, PARC, MSR, HP, Accenture, etc.
• Startup companies
• Crossbow, Dust, Ember, Sensicast, Moteiv, etc.
• Media Buzz
• Over 30 news articles since July 2002 covering
  Intel-Berkeley/UC Berkeley sensor network
  activities
• One of 10 emerging technologies that will change
  the world MIT Technology Review

9
A Brief History of Sensornets

• People have used sensors for a long time
• Recent CS History
• (1998) Pottie Kaiser Radio based networks of
  sensors
• (1998) Pister et al Smart Dust
• Initial focus on optical communication
• By 1999, radio based networks, COTS Dust, Motes
• (1999) Estrin Govindan
• Ad-hoc networks of sensors
• (2000) Culler/Hill et al TinyOS Motes
• (2000) Bonnet/Seshadri Device Database Systems
• (2002) Madden/Franklin/Hellerstein/Hong TinyDB
• (2002) Hill / Dust SPEC, mm3 scale computing

• UCLA / USC / Berkeley Continue to Lead Research
• Many other players now
• TinyOS/Motes as most common platform
• Emerging commercial space
• Crossbow, Ember, Dust, Sensicast, Moteiv, Intel

10
Why Now?

• Commoditization of radio hardware
• Cellular and cordless phones, wireless
  communication
• Low cost -gt many/tiny -gt new applications!
• Real application for ad-hoc network research from
  the late 90s
• Coming together of EE CS communities

11
Motes
12
History of Motes

• Initial research goal wasnt hardware
• Has since become more of a priority with emerging
  hardware needs, e.g.
• Power consumption
• (Ultrasonic) ranging localization
• MIT Cricket, NEST Project
• Connectivity with diverse sensors
• UCLA sensor board
• Even so, now on the 5th generation of devices
• Costs down to 50/node (Moteiv, Dust)
• Greatly improved radio quality
• Multitude of interfaces USB, Ethernet, CF, etc.
• Variety of form factors, packages

13
Motes vs. Traditional Computing

• Embedded OS
• Lossy, Adhoc Radio Communication
• Sensing Hardware
• Severe Power Constraints

14
NesC/TinyOS

• NesC a C dialect for embedded programming
• Components, wired together
• Quick commands and asynch events

• TinyOS a set of NesC components
• hardware components
• ad-hoc network formation maintenance
• time synchronization

Think of the pair as a programming environment
15
Radio Communication

• Low Bandwidth Shared Radio Channel
• 40kBits on motes
• Much less in practice
• Encoding, Contention for Media Access (MAC)
• Very lossy 30 base loss rate
• Argues against TCP-like end-to-end retransmission
• And for link-layer retries
• Generally, not well behaved

16
Types of Sensors

• Sensors attach via daughtercard

• Weather
• Temperature
• Light x 2 (high intensity PAR, low intensity,
  full spectrum)
• Air Pressure
• Humidity

• Vibration
• 2 or 3 axis accelerometers
• Tracking
• Microphone (for ranging and acoustic signatures)
• Magnetometer
• GPS
• RFID Reader

17
Non-Volatile Storage

• EEPROM
• 512K off chip, 32K on chip
• Writes at disk speeds, reads at RAM speeds
• Interface random access, read/write 256 byte
  pages
• Maximum throughput 10Kbytes / second
• MatchBox Filing System
• Provides a Unix-like file I/O interface
• Single, flat directory
• Only one file being read/written at a time

18
Power Consumption and Lifetime

• Power typically supplied by a small battery
• 1000-2000 mAH
• 1 mAH 1 milliamp current for 1 hour
• Typically at optimum voltage, current drain rates
• Power Watts (W) Amps (A) Volts (V)
• Energy Joules (J) W time
• Lifetime, power consumption varies by application
• Processor 5mA active, 1 mA idle, 5 uA sleeping
• Radio 5 mA listen, 10 mA xmit/receive, 20mS /
  packet
• Sensors 1 uA -gt 100s mA, 1 uS -gt 1 S / sample

19
Energy Usage in A Typical Data Collection Scenario

• Each mote collects 1 sample of (light,humidity)
  data every 10 seconds, forwards it
• Each mote can hear 10 other motes
• Process
• Wake up, collect samples ( 1 second)
• Listen to radio for messages to forward (1
  second)
• Forward data

20
Sensors Slow, Power Hungry, Noisy
21
Agenda

• Sensornet Background
• TinyDB
• TinyDB Overview
• Data Model and Query Language
• Demo
• TinyDB Java API and Scripting
• Demo with TinyDB GUI
• TinyDB Internals
• Extending TinyDB
• TinyDB Status and Roadmap
• TASK
• A Flavor of Research Challenges

22
TinyDB Revisited
SELECT MAX(mag) FROM sensors WHERE mag gt
thresh SAMPLE PERIOD 64ms

• High level abstraction
• Data centric programming
• Interact with sensor network as a whole
• Extensible framework
• Under the hood
• Intelligent query processing query optimization,
  power efficient execution
• Fault Mitigation automatically introduce
  redundancy, avoid problem areas

App
Query, Trigger
Data
TinyDB
23
Feature Overview

• Declarative SQL-like query interface
• Metadata catalog management
• Multiple concurrent queries
• Network monitoring (via queries)
• In-network, distributed query processing
• Extensible framework for attributes, commands and
  aggregates
• In-network, persistent storage

24
Architecture
TinyDB GUI
JDBC
TinyDB Client API
DBMS
PC side
0
Mote side
0
TinyDB query processor
2
1
3
8
4
5
6
Sensor network
7
25
Query Language (TinySQL)

• SELECT ltaggregatesgt, ltattributesgt
• FROM sensors ltbuffergt
• WHERE ltpredicatesgt
• GROUP BY ltexprsgt
• SAMPLE PERIOD ltconstgt ONCE
• INTO ltbuffergt
• TRIGGER ACTION ltcommandgt

26
Demo Time
27
TinySQL Examples
Find the sensors in bright nests.
Sensors

• SELECT nodeid, nestNo, light
• FROM sensors
• WHERE light gt 400
• EPOCH DURATION 1s

1
28
TinySQL Examples (cont.)
Count the number occupied nests in each loud
region of the island.
29
Event-based Queries

• ON event SELECT
• Run query only when interesting events happen
• Event examples
• Button pushed
• Message arrival
• Bird enters nest
• Analogous to triggers but events are user-defined

30
Query over Stored Data

• Named buffers in Flash memory
• Store query results in buffers
• Query over named buffers
• Analogous to materialized views
• Example
• CREATE BUFFER name SIZE x (field1 type1, field2
  type2, )
• SELECT a1, a2 FROM sensors SAMPLE PERIOD d
  INTO name
• SELECT field1, field2, FROM name SAMPLE
  PERIOD d

31
Using the Java API

• SensorQueryer
• translateQuery() converts TinySQL string into
  TinyDBQuery object
• Static query optimization
• TinyDBNetwork
• sendQuery() injects query into network
• abortQuery() stops a running query
• addResultListener() adds a ResultListener that is
  invoked for every QueryResult received
• removeResultListener()
• QueryResult
• A complete result tuple, or
• A partial aggregate result, call
  mergeQueryResult() to combine partial results
• Key difference from JDBC push vs. pull

32
Writing Scripts with TinyDB

• TinyDBs text interface
• java net.tinyos.tinydb.TinyDBMain run select
• Query results printed out to the console
• All motes get reset each time new query is posed
• Handy for writing scripts with shell, perl, etc.

33
Inside TinyDB
Multihop Network
Query Processor
10,000 Lines Embedded C Code 5,000 Lines
(PC-Side) Java 3200 Bytes RAM (w/ 768 byte
heap) 58 kB compiled code (largest TinyOS
program)
Filterlight gt 400
Schema
TinyOS
TinyDB
34
Tree-based Routing

• Tree-based routing
• Used in
• Query delivery
• Data collection
• In-network aggregation
• Parent Selection
• Goals
• Ensure good connectivity
• Maintain loop-free
• Techniques
• First person you hear from
• Snoop on others, keep track of promising
  contenders, switch as appropriate

35
Power Management Approach

• Coarse-grained app-controlled communication
  scheduling

Epoch (10s -100s of seconds)
Mote ID
1
zzz
zzz
2
3
4
5
time
2-4s Waking Period
36
Time Synchronization

• All messages include a 5 byte time stamp
  indicating system time in ms
• Synchronize (e.g. set system time to timestamp)
  with
• Any message from parent
• Any new query message (even if not from parent)
• Punt on multiple queries
• Timestamps written just after preamble is xmitted
• All nodes agree that the waking period begins
  when (system time epoch dur 0)
• And lasts for WAKING_PERIOD ms
• Adjustment of clock happens by changing duration
  of sleep cycle, not wake cycle.

37
Extending TinyDB

• Why extending TinyDB?
• New sensors ? attributes
• New control/actuation ? commands
• New data processing logic ? aggregates
• New events
• Analogous to concepts in object-relational
  databases

38
Adding Attributes

• Types of attributes
• Sensor attributes raw or cooked sensor readings
• Introspective attributes parent, voltage, ram
  usage, etc.
• Constant attributes constant values that can be
  statically or dynamically assigned to a mote,
  e.g., nodeid, location, etc.

39
Adding Attributes (cont)

• Interfaces provided by Attr component
• StdControl init, start, stop
• AttrRegister
• command registerAttr(name, type, len)
• event getAttr(name, resultBuf, errorPtr)
• event setAttr(name, val)
• command getAttrDone(name, resultBuf, error)
• AttrUse
• command startAttr(attr)
• event startAttrDone(attr)
• command getAttrValue(name, resultBuf, errorPtr)
• event getAttrDone(name, resultBuf, error)
• command setAttrValue(name, val)

40
Adding Attributes (cont)

• Steps to adding attributes to TinyDB
• Create attribute nesC components
• Wire new attribute components to TinyDBAttr
  configuration
• Reprogram TinyDB motes
• Add new attribute entries to catalog.xml
• Constant attributes can be added on the fly
  through TinyDB GUI

41
Adding Aggregates

• Step 1 wire new nesC components

42
Adding Aggregates (cont)

• Step 2 add entry to catalog.xml
• ltaggregategt
• ltnamegtAVGlt/namegt
• ltidgt5lt/idgt
• lttemporalgtfalselt/temporalgt
• ltreaderClassgtnet.tinyos.tinydb.AverageClasslt/read
  erClassgt
• lt/aggregategt
• Step 3 (optional) implement reader class in Java
• a reader class interprets and finalizes aggregate
  state received from the mote network, returns
  final result as a string for display.

43
TinyDB Status

• Latest released with TinyOS 1.1.6 (5/04)
• Install the task-tinydb package in TinyOS 1.1
  distribution
• First release in TinyOS 1.0 (9/02)
• Widely used by research groups as well as
  industry pilot projects
• Ongoing deployments in Intel Berkeley Lab and
  redwood trees at UCs Sonoma Grove
• Largest deployment 80 weather station nodes
• Network longevity 4-5 months

44
The Redwood Tree Deployment

• Redwood Grove in UC Botanical Garden, Berkeley
• Collect dense sensor readings to monitor climatic
  variations across
• altitudes,
• angles,
• time,
• forest locations, etc.
• Versus sporadic monitoring points with 30lb
  loggers!
• Current focus study how dense sensor data affect
  predictions of conventional tree-growth models

45
Data from Redwoods
36m
33m 111
32m 110
30m 109,108,107
20m 106,105,104
10m 103, 102, 101
46
TinyDB Roadmap (near term)

• Support for high frequency sampling
• For a variety of deployments
• Intel Fab, BP shipboard, Golden Gate Bridge
• Store and forward
• Bulk reliable data transfer
• Scheduling of communications
• Research agenda continues
• Discussion later

47
For more information

• http//berkeley.intel-research.net/tinydb or
  http//triplerock.cs.bekeley.edu/tinydb

48
Agenda

• Sensornet Background
• TinyDB
• TASK
• SW/HW architecture
• Features
• A Flavor of Research Challenges

49
SensorNet Dilemma

• Sensors still packaged like HeathKits
• Pretty hard to cope with out of the box
• Bare metal encourages one-off applications
• Inhibits reuse
• Deployment not intuitive
• No configuration/monitoring tools
• SensorNet PhD Factor
• Today 2.5 PhDs needed to deploy a SensorNet
• Needs to be Zero

50
TASK Design Requirements

• Ease of S/W Installation
• Deployment tools
• Reconfigurability
• Health/Mgmt Monitoring
• Network Reliability Guarantee
• Interpretable Sensor Results
• Tool Integration

• Audit Trails
• Lifetime estimates

For Developers

• Familiar API
• Extensibility of S/W
• Modular services

51
Tiny Application Sensor Kit
TASK Client Tools
External Tools
TaskView
Internet
TASK Field Tools
SensorNet Appliance
TASK Server

• Simplicity vs. Functionality
• Modularity
• Remote control
• Fault Tolerant

TinyDB Sensor Network
52
SensorNet Appliance
SNA

• Intelligent Gateway
• Proxy for the sensornet
• Distributes query
• Stages results
• Manages configuration
• Components
• TASK Server
• TinyDB Client (Java)
• DBMS (PostgreSQL)
• WebServer (Apache)

http, other
TASKServer
DBMS
ODBC
TinyDB Client
SensorNet
53
Tools

• Field Tool
• In-situ diagnostics
• TaskView
• Integrated tool for management and monitoring

54
Quick TASK Demo
55
Agenda

• Sensornet Background
• TinyDB
• TASK
• A Flavor of Research Challenges

56
Sensor Network Research

• Very active research area
• Cant summarize it all
• Focus database-relevant research topics
• Some outside of Berkeley/MIT
• Other topics that are itching to be scratched
• But, some bias towards work that we find
  compelling

57
Topics

• In-network aggregation
• Acquisitional Query Processing
• Heterogeneity
• Intermittent Connectivity
• In-network Storage
• Statistical modeling and summarization
• In-network Joins
• Adaptivity and Sensor Networks
• Multiple Queries

58
Topics

• In-network aggregation
• Acquisitional Query Processing
• Heterogeneity
• Intermittent Connectivity
• In-network Storage
• Statistical modeling and summarization
• In-network Joins
• Adaptivity and Sensor Networks
• Multiple Queries

59
Tiny Aggregation (TAG)

• In-network processing of aggregates
• Common data analysis operation
• Aka gather operation or reduction in
  programming
• Communication reducing
• Operator dependent benefit
• Exploit query semantics to improve efficiency!

Madden, Franklin, Hellerstein, Hong. Tiny
AGgregation (TAG), OSDI 2002.
60
Basic Aggregation

• In each epoch
• Each node samples local sensors once
• Generates partial state record (PSR)
• local readings
• readings from children
• Outputs PSR during assigned comm. interval
• Interval assigned based on depth in tree

• At end of epoch, PSR for whole network output at
  root
• New result on each successive epoch

61
Illustration In-Network Aggregation
SELECT COUNT() FROM sensors
Interval 4
Sensor
Sample Period
Interval
Time
1
62
Illustration In-Network Aggregation
SELECT COUNT() FROM sensors
Interval 3
Sensor
2
Interval
63
Illustration In-Network Aggregation
SELECT COUNT() FROM sensors
Interval 2
Sensor
1
3
Interval
64
Illustration In-Network Aggregation
SELECT COUNT() FROM sensors
Interval 1
5
Sensor
Interval
65
Illustration In-Network Aggregation
SELECT COUNT() FROM sensors
Interval 4
Sensor
Interval
1
66
Illustration In-Network Aggregation
SELECT COUNT() FROM sensors
Interval 4
Sensor
Interval
1
67
Aggregation Framework

• As in extensible databases, TinyDB supports any
  aggregation function conforming to

Aggnfinit, fmerge, fevaluate Finit a0 ?
lta0gt Fmerge lta1gt,lta2gt ? lta12gt Fevaluate lta1gt
? aggregate value
Partial State Record (PSR)
Example Average AVGinit v ?
ltv,1gt AVGmerge ltS1, C1gt, ltS2, C2gt ? lt S1
S2 , C1 C2gt AVGevaluateltS, Cgt ? S/C
Restriction Merge associative, commutative
68
Taxonomy of Aggregates

• TAG insight classify aggregates according to
  various functional properties
• Yields a general set of optimizations that can
  automatically be applied

Drives an API!
69
Use Multiple Parents

• Use graph structure
• Increase delivery probability with no
  communication overhead
• For duplicate insensitive aggregates, or
• Aggs expressible as sum of parts
• Send (part of) aggregate to all parents
• In just one message, via multicast
• Assuming independence, decreases variance
• Recent research on convertingto robust
  dup-insensitivity
• E.g. count can be approximated robustlyin a
  dup-insensitive fashionConsidine, et al., 2003

SELECT COUNT()
70
Multiple Parents Results

• Better than previous analysis expected!
• Losses arent independent!
• Insight spreads data over many links

71
Statistical Techniques

• Approximations, summaries, and sampling based on
  statistics and statistical models
• Applications
• Limited bandwidth and large number of nodes -gt
  data reduction
• Lossiness -gt predictive modeling
• Uncertainty -gt tracking correlations and changes
  over time
• Physical models -gt improved query answering

72
Correlated Attributes

• Data in sensor networks is correlated e.g.,
• Temperature and voltage
• Temperature and light
• Temperature and humidity
• Temperature and time of day
• etc.

73
BBQ Model-based Probabilistic Querying over
Sensor Networks
Joint work with Amol Desphande and Carlos
Guestrin
Query Processor
Model
1
3
4
2
5
BarbieQ A Tiny Model Driven Query System
6
7
8
9
74
BBQ Model-based Probabilistic Querying over
Sensor Networks
Query Processor
Model
Consult Model
1
3
4
2
5
6
7
8
9
75
BBQ Model-based Probabilistic Querying over
Sensor Networks
Query Processor
Model
Consult Model
1
3
4
2
5
6
7
8
9
76
BBQ Model-based Probabilistic Querying over
Sensor Networks
Query Results
Query Processor
Model
Update Model
1
3
4
2
5
6
7
8
9
77
Challenges

• What kind of models to use ?
• Routing
• How to tour the selected set of nodes
• Optimization
• Given a model and a query, find the best set of
  attributes to observe
• Cost not easy to measure
• Non-uniform network communication costs
• Changing network topologies
• Depends on touring technique
• Large plan space
• Might be cheaper to observe attributes not in
  query
• e.g. Voltage instead of Temperature
• Conditional Plans
• Change the observation plan based on observed
  values

78
BBQ Current Prototype

• Multi-variate Gaussian Models
• Kalman Filters to capture correlations across
  time
• Handles
• Range predicate queries
• sensor value within x,y, w/ confidence
• Value queries
• sensor value x, w/in epsilon, w/ confidence
• Simple aggregate queries
• AVG(sensor value) ? n, w/in epsilon, w/confidence
• Uses a greedy algorithm to choose the observation
  plan

79
A complex aggregate Wavelets

• just like AVG or COUNT
• but fancier -)
• re-codes a signal into a sum of square waves
• biggest coefficients ? approx. reconstruction
• lossy compression
• multi-resolution

• Examples
• histograms
• 2-d or 3-d compression and reconstruction

80
computing a wavelet
35, -1, 3, 8, -4,
3, 3, 3
81
computing a wavelet
And a correspondingCommunication Graph
82
computing a wavelet
And a correspondingCommunication Graph
83
computing a wavelet
And a correspondingCommunication Graph
84
resulting comm graph
a binomial tree!
85
I never promised you a binomial

• TinyDB agg trees are not any special shape
• Parents chosen for reliability, not tree shape
• Options
• Cope with non-binomial trees
• How to merge two subtrees of different sizes?
• Can muck with the Wavelet, but it biases the
  answer
• Can forward when merge is impossible, but comm
  cost
• Develop a scheme to ensure binomial agg trees
• When is it possible?
• Distributed algorithm to do it?

86
Concluding Remarks

• Sensor nets becoming a reality
• For sensornet system hackers, TinyOS an emerging
  standard
• For sensornet app writers, TinyDB an emerging
  standard
• TASK a developing example of a turnkey solution
• Lots of room for innovation
• Applications and vertical solutions
• Systems work in OS, Networks, DBs
• Opportunities in Statistics, AI, Signal
  Processing
• This all meshes in new and interesting ways!

87
Sensornets Getting Started

• TinyDB home page
• http//telegraph.cs.berkeley.edu/tinydb
• The TinyOS home page
• http//webs.cs.berkeley.edu/tinyos
• Crossbow (motes stargates)
• http//www.xbow.com
• Moteiv (motes)
• http//www.moteiv.com
• Intel Imote
• www.intel.com/research/exploratory/motes.htm.

88
Backup Slides
89
computing a wavelet
35, 0, 3, 8, -4,
0, 0, 0
90
computing a wavelet
35, 0, 3, 8, -4,
0, 0, 0