Fine-Grained Network Time Synchronization using Reference Broadcasts Jeremy Elson, Lew Girod, and Deborah Estrin OSDI 2002 - Boston, MA Speaker : hsiwei-Chen.

Slides:



Advertisements
Similar presentations
Christoph Lenzen Philipp Sommer Philipp Sommer Roger Wattenhofer Roger Wattenhofer Optimal Clock Synchronization in Networks.
Advertisements

The Flooding Time Synchronization Protocol
Time Synchronization - using Reference-Broadcast Synchronization
Gradient Clock Synchronization in Wireless Sensor Networks
MONET 05 Time Synchronization In Wireless Sensor Networks Anisha Menon.
April 28 th 2005 Time Synchronization in Sensor Networks 1 Time synchronization in sensor networks Presented by: Nirav Jasapara Graduate Student, USC.
Pathload A measurement tool for end-to-end available bandwidth Manish Jain, Univ-Delaware Constantinos Dovrolis, Univ-Delaware Sigcomm 02.
HIERARCHY REFERENCING TIME SYNCHRONIZATION PROTOCOL Prepared by : Sunny Kr. Lohani, Roll – 16 Sem – 7, Dept. of Comp. Sc. & Engg.
Wireless Sensor Networks Clock Synchronization Professor Jack Stankovic University of Virginia.
TELE202 Lecture 8 Congestion control 1 Lecturer Dr Z. Huang Overview ¥Last Lecture »X.25 »Source: chapter 10 ¥This Lecture »Congestion control »Source:
CPSC 689: Discrete Algorithms for Mobile and Wireless Systems Spring 2009 Prof. Jennifer Welch.
Time Synchronization for Wireless Sensor Networks
Emulatore di Protocolli di Routing per reti Ad-hoc Alessandra Giovanardi DI – Università di Ferrara Pattern Project Area 3: Problematiche di instradamento.
Distributed Systems Fall 2010 Time and synchronization.
Time Synchronization (RBS, Elson et al.) Presenter: Peter Sibley.
More routing protocols Alec Woo June 18 th, 2002.
He Huang Introduction:The Flooding Time Synchronization Protocol.
1 University of Freiburg Computer Networks and Telematics Prof. Christian Schindelhauer Wireless Sensor Networks 14th Lecture Christian Schindelhauer.
1 University of Freiburg Computer Networks and Telematics Prof. Christian Schindelhauer Wireless Sensor Networks 15th Lecture Christian Schindelhauer.
The Flooding Time Synchronization Protocol
Time Synchronization Murat Demirbas SUNY Buffalo.
Time Synchronization for Wireless Sensor Networks.
1 University of Freiburg Computer Networks and Telematics Prof. Christian Schindelhauer Wireless Sensor Networks 13th Lecture Christian Schindelhauer.
TPS: A Time-Based Positioning Scheme for outdoor Wireless Sensor Networks Authors: Xiuzhen Cheng, Andrew Thaeler, Guoliang Xue, Dechang Chen From IEEE.
Lecture 12 Synchronization. EECE 411: Design of Distributed Software Applications Summary so far … A distributed system is: a collection of independent.
Fine-Grained Network Time Synchronization using Reference Broadcasts Jeremy Elson, Lewis Girod, and Deborah Estrin U.C.L.APresenter: Todd Fielder.
Lecture 2-1 CS 425/ECE 428 Distributed Systems Lecture 2 Time & Synchronization Reading: Klara Nahrstedt.
8/18/2015 Mobile Ad hoc Networks COE 549 Synchronization Tarek Sheltami KFUPM CCSE COE 1.
1 Physical Clocks need for time in distributed systems physical clocks and their problems synchronizing physical clocks u coordinated universal time (UTC)
Timing-sync Protocol for Sensor Networks (TPSN) Presenter: Ke Gao Instructor: Yingshu Li.
Energy-Aware Synchronization in Wireless Sensor Networks Yanos Saravanos Major Advisor: Dr. Robert Akl Department of Computer Science and Engineering.
Fine-Grained Network Time Synchronization Using Reference Broadcasts Jeremy Elson, Lew Girod, and Deborah Estrin University of California, Los Angeles.
E 2 DTS: An energy efficiency distributed time synchronization algorithm for underwater acoustic mobile sensor networks Zhengbao Li, Zhongwen Guo, Feng.
Vinod Kulathumani West Virginia University
1 Pertemuan 20 Teknik Routing Matakuliah: H0174/Jaringan Komputer Tahun: 2006 Versi: 1/0.
CS450 Network Embedded Sensing Systems Week 11: Time Synchronization and Reconstruction Jayant Gupchup.
Clock Synchronization in Sensor Networks Mostafa Nouri.
Adaptive Control-Based Clock Synchronization in Wireless Sensor Networks Kasım Sinan YILDIRIM *, Ruggero CARLI +, Luca SCHENATO + * Department of Computer.
Clock Synchronization in Sensor Networks for Civil Security Farnaz Moradi Asrin Javaheri.
Improving Capacity and Flexibility of Wireless Mesh Networks by Interface Switching Yunxia Feng, Minglu Li and Min-You Wu Presented by: Yunxia Feng Dept.
Effects of Multi-Rate in Ad Hoc Wireless Networks
The Network Layer.
Load-Balancing Routing in Multichannel Hybrid Wireless Networks With Single Network Interface So, J.; Vaidya, N. H.; Vehicular Technology, IEEE Transactions.
Ryan Sites – CEG 790 Paper Presentation The Flooding Time Synchronization Protocol The Flooding Time Synchronization Protocol (SenSys ’04, November 3-5,
1 Clock Synchronization for Wireless Sensor Networks: A Survey Bharath Sundararaman, Ugo Buy, and Ajay D. Kshemkalyani Department of Computer Science University.
4: DataLink Layer1 Multiple Access Links and Protocols Three types of “links”: r point-to-point (single wire, e.g. PPP, SLIP) r broadcast (shared wire.
Outline for Today Objectives: –Time and Timers Administrative details: –Talk on learning at 4 in 130 North Building –Questions?
Wireless Sensor Networks Nov 1, 2006 Jeon Bokgyun
Time synchronization for UWSN. Outline Time synchronization knowledge Typical time sync protocol Time sync in UWSN Discussion.
Topic 2: Communications (Short Lecture) Jorge J. Gómez.
TIMING-SYNC PROTOCOL FOR SENSOR NETWORKS Presented by: Shet, Deepak Rajput, Rajiv.
Fasika Assegei Segal’s Law A man with a watch knows what time it is. A man with two watches is never sure.
Distributed Systems Principles and Paradigms Chapter 05 Synchronization.
Computer Science 1 TinySeRSync: Secure and Resilient Time Synchronization in Wireless Sensor Networks Speaker: Sangwon Hyun Acknowledgement: Slides were.
Software Development Infrastructure for Sensor Networks Operating systems (TinyOS) –Resource (device) management –Basic primitives –Protocols (MAC, routing)
Time This powerpoint presentation has been adapted from: 1) sApr20.ppt.
Tanenbaum & Van Steen, Distributed Systems: Principles and Paradigms, 2e, (c) 2007 Prentice-Hall, Inc. All rights reserved DISTRIBUTED SYSTEMS.
1 Segal’s Law A man with a watch knows what time it is. A man with two watches is never sure.
Time Synchronization Protocols in Wireless Sensor Networks.
An Energy-Efficient MAC Protocol for Wireless Sensor Networks Speaker: hsiwei Wei Ye, John Heidemann and Deborah Estrin. IEEE INFOCOM 2002 Page
UNIT IV INFRASTRUCTURE ESTABLISHMENT. INTRODUCTION When a sensor network is first activated, various tasks must be performed to establish the necessary.
Global Clock Synchronization in Sensor Networks Qun Li, Member, IEEE, and Daniela Rus, Member, IEEE IEEE Transactions on Computers 2006 Chien-Ku Lai.
6 SYNCHRONIZATION. introduction processes synchronize –exclusive access. –agree on the ordering of events much more difficult compared to synchronization.
Time Synchronization in Ad Hoc Networks Kay R¨omer MobiHOC 2001 Speaker : hsiwei.
Distributed Systems Lecture 5 Time and synchronization 1.
MAC Protocols for Sensor Networks
Outline for Today Objectives: –Time and Timers Administrative details: –About the final…
MAC Protocols for Sensor Networks
Net 435: Wireless sensor network (WSN)
Overview: Chapter 4 Infrastructure Establishment
Presentation transcript:

Fine-Grained Network Time Synchronization using Reference Broadcasts Jeremy Elson, Lew Girod, and Deborah Estrin OSDI Boston, MA Speaker : hsiwei-Chen

Outline  Introduction  Reference-Broadcast Synchronization(RBS)  Implementation on Berkeley Motes  Conclusions and FutureWork

Sources of Time Synchronization Error(1/2) Send Time Send Time  Time spent at the sender to construct the message, e.g. context switches and system call overhead  The time required to transfer the message from the host to its network interface. Access Time Access Time  Wait for the channel to be clear before transmitting

Sources of Time Synchronization Error(2/2) Propagation Time Propagation Time  The time needed for the message to transit from sender to receivers Receive Time Receive Time  Time to receive the message and notify the host of its arrival Send time & access time are most non-deterministic

Reference-Broadcast Synchronization(RBS) Traditional protocols Traditional protocols Synchronize the sender of a message with its receiver Synchronize the sender of a message with its receiver RBS RBS A set of nodes synchronize with each other (not with the sender) A set of nodes synchronize with each other (not with the sender)  reduces non-determinism

Critical Path Analysis Traditional RBS Send Time Access Time Propagation Delay Receive Time

RBS Critical Path Remove the send time & access time from the critical path Remove the send time & access time from the critical path  Receive time is independent of the sender or network load  Propagation time can be ignored in a short distance

RBS Critical Path (Cont.) Use broadcast as a relative time reference Use broadcast as a relative time reference  Broadcast packet does NOT include timestamp generated by sender  Any broadcast, e.g., RTS/CTS, route discovery packets

Simple RBS One transmitter broadcasts a reference packet to two receivers One transmitter broadcasts a reference packet to two receivers Each receiver records the time the reference packet arrived according to the local clock Each receiver records the time the reference packet arrived according to the local clock Receivers exchange the timestamps Receivers exchange the timestamps We consider the Propagation Time to be effectively 0 We consider the Propagation Time to be effectively 0 Sensitiving to differences in the receive time of messages within a set of receivers. Sensitiving to differences in the receive time of messages within a set of receivers.

Simple RBS Sender Receiver 1 Receiver 2 Broadcast exchange the timestamps

Distribution of Inter-Receiver Phase Offset Pairwise Difference in Packet Reception Time (usec) Number of Trials Phase offset recorded by a logic analyzer for 1,478 broadcast packets(total) (sender emitted 160 broadcast packets) Gaussian distribution  = 0,  = 11.1us Confidence = 99.8% 5 motes: receivers 1 mote: transmitter

Estimation of Phase Offset A sender broadcasts m reference packets A sender broadcasts m reference packets Each of the n receiver records the time at which the reference was received Each of the n receiver records the time at which the reference was received Receivers exchange timestamps Receivers exchange timestamps Receiver i computes its phase offset to any other receiver j as the average of m differences Receiver i computes its phase offset to any other receiver j as the average of m differences

Estimation of Clock Skew Clock skew due to crystal oscillators errors Clock skew due to crystal oscillators errors  Frequency error Difference between the expected and actual frequencies Difference between the expected and actual frequencies  Instability Short-term: Environmental factors, e.g., variations in temperature, supply voltage, and shock Short-term: Environmental factors, e.g., variations in temperature, supply voltage, and shock Long-term: Oscillator aging Long-term: Oscillator aging  Least square method (linear regression) Assume phase error is changing at constant rate Assume phase error is changing at constant rate Continuously adjust frequency based on recent observations relating the local oscillator to a reference Continuously adjust frequency based on recent observations relating the local oscillator to a reference

Implementation on Berkeley Motes One sender, five receivers One sender, five receivers 1000 trials 1000 trials

Receiver Group Dispersion #Reference packet broadcasts Receiver Group Dispersion (us) 25 broadcasts can reduce the dispersion below 10us 20 Receivers 2 Reeivers

Effect of Clock Skew on RBS Time (sec) Phase Offset (us) Fit Error (us)

Performance Comparisons Compare to NTP Compare to NTP UNIX implementation of RBS for fair comparisons UNIX implementation of RBS for fair comparisons  User space UNIX daemon  UDP datagram  StrongArm-based Compaq IPAQ PDA’s “Familiar” Linux (2.4 kernel) “Familiar” Linux (2.4 kernel)  Lucent Technology 11Mbit wireless Ethernet adapters 300 trials, 每個 trial 間格 8 秒鐘,Total time is 40 minuters 300 trials, 每個 trial 間格 8 秒鐘,Total time is 40 minuters

Synchronization Error with Light Network Load Error (us) Cumulative Error Prob. RBS NTP NTP-Offset Sync. Error between two IPAQ’s

Synchronization Error with Light Network Load RBS: / us RBS: / us NTP: / us NTP: / us NTP-Offset: 204 +/ us NTP-Offset: 204 +/ us RBS works better due to the reduced non- determinism RBS works better due to the reduced non- determinism

Synchronization Error with Heavy Network Load Error (us) Cumulative Error Prob. RBS NTP NTP-Offset

Synchronization Error with Heavy Network Load RBS: / us RBS: / us  With kernel support which timestamps at interrupt time, / us NTP: / us NTP: / us NTP-Offset: / us NTP-Offset: / us RBS is not affected by Medium Access Delay RBS is not affected by Medium Access Delay

Multi-Hop Time Synchronization 在 node A 面積 local 中,node 1 與 node 4 同步 2. 再以 node 4 當 getway, 與 node B 面積 local 中,node 4 與 node 7 同步 3. 即可達到 Global 同步. A 5 67 B

Multi-Hop RBS Performance Hop Mean Error( )

Conclusions and FutureWork RBS provides more precise, flexible, and resource-efficient network time synchronization than traditional algorithms RBS provides more precise, flexible, and resource-efficient network time synchronization than traditional algorithms But such as automatic, dynamic election of the set of nodes to act as beacon senders ? But such as automatic, dynamic election of the set of nodes to act as beacon senders ?

Summary 本篇作者去除掉了 Send Time and Acces 本篇作者去除掉了 Send Time and Acces Time 同步 error 的兩個來源 Time 同步 error 的兩個來源 以相對的時間代替絕對的時間 以相對的時間代替絕對的時間

Thank you