time, synchronization, and wireless sensor...
TRANSCRIPT
![Page 1: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/1.jpg)
Ted Herman/March 2005 1
Time, Synchronization, and Wireless Sensor Networks
Part II
Time, Synchronization, Time, Synchronization, and Wireless Sensor Networksand Wireless Sensor Networks
Part IIPart II
Ted HermanUniversity of Iowa
![Page 2: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/2.jpg)
Ted Herman/March 2005 2
Presentation: Part IImetrics and techniquesmetrics and techniquessingle-hop beaconsregional time zonesrouting-structure and leader clockuniform convergenceconclusion
![Page 3: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/3.jpg)
Ted Herman/March 2005 3
Multihop Synchronizationo wireless sensor networks are multihop
(sometimes ad hoc) networkso measures of quality of synchronization:
δ-difference between neighboring clocks∆-difference between basestation and any clockδ-difference along any path in routing tree
∆
δδδ
basestation
δ?
![Page 4: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/4.jpg)
Ted Herman/March 2005 4
Synchronization Techniques1. use GPS or radio beacon
requires special hardware, extra cost∆ = δ
2. use only “regional time zones”complicated time zone conversion gateways
3. use routing structure and leader clock∆ = (distance) x δbuilding, maintaining routing structure fault tolerance issues
4. use uniform convergence to maximal clockssimilar metrics to routing structure, but different fault tolerance properties
5. other: biologically-inspired methods, phase “waves”, time-flow algorithms (not yet practical)
![Page 5: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/5.jpg)
Ted Herman/March 2005 5
How to Evaluate in Practice ?o can use GPS for independent evaluation
useful to evaluate skew, not so useful for fast evaluation of offset synchronization
o self-sampling: nodes calculate difference between clock and time in a timesync message
large difference lack of synchronyo probes: single-hop broadcast, timestamped by all who
receive, then transmit recorded timestamps and observe differences in the timestamps
1st: probe broadcast
2nd: send timestamp messages
3rd: compare timestamps to infer difference in local clocks
![Page 6: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/6.jpg)
Ted Herman/March 2005 6
Presentation: Part IImetrics and techniquessinglesingle--hop beaconshop beaconsregional time zonesrouting-structure and leader clockuniform convergenceconclusion
![Page 7: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/7.jpg)
Ted Herman/March 2005 7
Single-Hop Beacono excellent performanceo single point of failureo concerns of power, legality, stealth, assurance
o practical for open area, limited scaleo special hardware: tall antenna, strong signalo basically using standard sensor hardware
![Page 8: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/8.jpg)
Ted Herman/March 2005 8
Presentation: Part IImetrics and techniquessingle-hop beaconsregional time zonesregional time zonesrouting-structure and leader clockuniform convergenceconclusion
![Page 9: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/9.jpg)
Ted Herman/March 2005 9
Regional Time Zoneso proposed for RBS
(Reference Broadcast Synchronization [Elson, 2003])
o use only “regional time zones”o conversion adds complexity --- but useful if
timesync not needed everywhere
1
3
2
A4
8
C
5
7
6B
10
D11
9
1
3
2
4
8
5
7
6
10 11
9
![Page 10: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/10.jpg)
Ted Herman/March 2005 10
RBS Statisticso multiple reference beacons, receiver-receiver
synchronization forms distribution of noise
![Page 11: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/11.jpg)
Ted Herman/March 2005 11
Noise Filteringo elimination of noise by knowledge of distribution &
error-minimizing hypotheses
![Page 12: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/12.jpg)
Ted Herman/March 2005 12
RBS Statistical Techniqueo linear regression used to obtain best offseto outlier removal would improve resultso linear regression also useful to correct skew
![Page 13: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/13.jpg)
Ted Herman/March 2005 13
Multihop RBS resultso some results after conversion over multiple regions
0
1
2
3
4
5
6
7
1 Hop 2 Hop 3 Hop 4 Hop
Erro
r (u
sec)
Std DevError
o better than worst case some errors positive, some errors negative, so some errors cancel
![Page 14: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/14.jpg)
Ted Herman/March 2005 14
Presentation: Part IImetrics and techniquessingle-hop beaconsregional time zonesroutingrouting--structure and leader clockstructure and leader clockuniform convergenceconclusion
![Page 15: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/15.jpg)
Ted Herman/March 2005 15
Rooted Spanning Treeo popular routing structure
basestation at rootselection of links in tree based on Quality metrics
o other routing types: “fat tree”, mesh, geographic
![Page 16: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/16.jpg)
Ted Herman/March 2005 16
Leader Clock at Rooto everyone follow parent in tree
periodic timesync message to neighborscollect many samples from parent (ignore others)use linear regression to follow parent offset & skew
![Page 17: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/17.jpg)
Ted Herman/March 2005 17
Leader Failureo leader doesn’t need to be basestation
if leader fails, recovery phase elects new leader
o leader election: leader is sensor node having smallest Id, parent is closest node to leader
o what happens when a node or link fails?much like routing table recovery, look for new path to leader, eventually reach “threshold timeout” and then elect a new leader
no leader …
![Page 18: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/18.jpg)
Ted Herman/March 2005 18
Evaluation of Leader Treeo generally excellent synchronization
however, strange cases can lead to δ ≈ ∆
o low overhead, simple implementationo rapid set-up for on-demand synchronization
(if we use basestation as root)
o suited to sensor networks where links are stable & failures are infrequent
o does not handle sensor mobility
![Page 19: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/19.jpg)
Ted Herman/March 2005 19
Presentation: Part IImetrics and techniquessingle-hop beaconsregional time zonesrouting-structure and leader clockuniform convergenceuniform convergenceconclusion
![Page 20: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/20.jpg)
Ted Herman/March 2005 20
Uniform Convergenceo basic idea: instead of a leader node, have all
nodes follow a “leader value”leader clock could be one with largest valueleader clock could be one with smallest valueleader value could be mean, median, etc
o local convergence global convergenceo send periodic timesync messages, use easy
algorithm to adjust offsetif (received_time > local_clock)
local_clock = received_time
![Page 21: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/21.jpg)
Ted Herman/March 2005 21
Uniform Convergence Advantageso fault tolerance is automatico each node takes input from all neighborso mobility of sensor nodes is no problemo extremely simple implementationo self-stabilizing from all possible states and
system configurations, partitions & rejoinso was useful in practice for “Line in the Sand”
demonstration
![Page 22: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/22.jpg)
Ted Herman/March 2005 22
Uniform Convergence Challengeso even one failure can contaminate entire
network (when failure introduces new, larger clock value)
o more difficult to correct skew than for treeo how to integrate GPS or other timesource?
we can use a hierarchy of clocks for application
o what does “largest clock” mean when clock reaches maximum value and rolls over?
rare occurrence, but happens somedaytransient failures could cause rollover sooner …
![Page 23: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/23.jpg)
Ted Herman/March 2005 23
Preventing Contaminationo algorithm: build picture of neighborhood
o node p collects timesync messages from all neighborso are they all reasonably close?
yes adjust local clock to maximum valueno cases to consider:
• more than one outlier no consensus, adjust to maximum value• only one outlier from “consensus clock range”
– if p is outlier, then p “reboots” its clock– if other neighbor is outlier, ignore that neighbor
o handles single-fault cases only
![Page 24: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/24.jpg)
Ted Herman/March 2005 24
Special Case: restarting nodeo algorithm: again, build picture of neighborhood
o node p joining network or rebooting clocko look for “normal” neighbors to trust
normal neighbors copy maximum of normal neighborsno normal neighbors adjust local clock to maximum value from any neighbor (including restarting ones)after adjusting to maximum, node becomes “normal”
![Page 25: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/25.jpg)
Ted Herman/March 2005 25
Clock Rollovero p’s clock advances from 232-1 back to zeroo q (neighbor of p) has clock value 232-35
question: what should q think of p’s clock?o proposal: use (<,max) cyclic ordering around
domain of values [0,232-1]
c<ba
d
he
g f<
< <<
<
<<
![Page 26: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/26.jpg)
Ted Herman/March 2005 26
Bad Case for Cyclic Orderingo network is in “ring” topologyo values (w,x,y,z) are about ¼ of 232 apart in
domain of clock values in ordering cycleo maybe, each node follows larger value of
neighbor in parallel never synchronizing!
x
w
z
y a solution to this problem
reset to zero when neighbor clocks are too far apart, use special rule after reset
![Page 27: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/27.jpg)
Ted Herman/March 2005 27
Presentation: Part IImetrics and techniquessingle-hop beaconsregional time zonesrouting-structure and leader clockuniform convergenceconclusionconclusion
![Page 28: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/28.jpg)
Ted Herman/March 2005 28
Conclusiono Part I
we saw how time sync has different needs & opportunities in wireless sensor networks than for traditional LAN/WAN/Internetpropagation delay often insignificantspecial techniques to deal with radio/MAC/system delays
![Page 29: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/29.jpg)
Ted Herman/March 2005 29
Conclusiono Part II
some quite varied alternatives for how to synchronize in multihop networkssingle-hop beacon (like GPS) good for some situationstime sync strategies can be similar to routing protocol structures (trees, zones)time sync is a “local” property, so notions like uniform convergence may be useful
![Page 30: Time, Synchronization, and Wireless Sensor Networksweb.cse.ohio-state.edu/~arora.9/788Notes/TimeSync-Two.pdf · Ted Herman/March 2005 17 Leader Failure o leader doesn’t need to](https://reader033.vdocument.in/reader033/viewer/2022050715/5f2e1e539d6ed82aac7eed9c/html5/thumbnails/30.jpg)
Ted Herman/March 2005 30
Conclusiono Some Open Problems
how to choose a timesync algorithm based on application requirements ?how to conserve energy in timesync ?are there special needs for coordinated actuation, long-term sleeping, sentries, and low duty cycles ?what kind of tools are helpful to use complicated timesync ideas, but make application design simple ?