Efficiency of Dynamic Arbitration in TDMA Protocols

Similar documents
Multiple Access (3) Required reading: Garcia 6.3, 6.4.1, CSE 3213, Fall 2010 Instructor: N. Vlajic

Exact Response Time of FlexRay Communication Protocol

Lecture 8: Media Access Control

Department of Computer Science and Engineering. CSE 3213: Communication Networks (Fall 2015) Instructor: N. Vlajic Date: Dec 13, 2015

Utilization Based Duty Cycle Tuning MAC Protocol for Wireless Sensor Networks

Lecture 8: Media Access Control. CSE 123: Computer Networks Stefan Savage

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

SourceSync. Exploiting Sender Diversity

COSC 3213: Computer Networks I Instructor: Dr. Amir Asif Department of Computer Science York University Section B

Time Iteration Protocol for TOD Clock Synchronization. Eric E. Johnson. January 23, 1992

Optimized Schedule Synthesis under Real-Time Constraints for the Dynamic Segment of FlexRay

Lecture 23: Media Access Control. CSE 123: Computer Networks Alex C. Snoeren

Design of an energy efficient Medium Access Control protocol for wireless sensor networks. Thesis Committee

UTILIZATION OF AN IEEE 1588 TIMING REFERENCE SOURCE IN THE inet RF TRANSCEIVER

CAN for time-triggered systems

Peripheral Sensor Interface for Automotive Applications

Outline. EEC-484/584 Computer Networks. Homework #1. Homework #1. Lecture 8. Wenbing Zhao Homework #1 Review

Department of Computer Science and Engineering. CSE 3213: Computer Networks I (Fall 2009) Instructor: N. Vlajic Date: Dec 11, 2009.

Time Triggered Protocol (TTP/C): A Safety-Critical System Protocol

WirelessHART Modeling and Performance Evaluation

6.1 Multiple Access Communications

Broadcast in Radio Networks in the presence of Byzantine Adversaries

Scheduling and Optimization of Fault-Tolerant Embedded Systems

Systems. Roland Kammerer. 29. October Institute of Computer Engineering Vienna University of Technology. Communication in Distributed Embedded

Multiple Access Methods

Politecnico di Milano Scuola di Ingegneria Industriale e dell Informazione. E2 Multiplexing

GSM and Similar Architectures Lesson 08 GSM Traffic and Control Data Channels

MACHINE TO MACHINE (M2M) COMMUNICATIONS-PART II

INTRODUCTION TO WIRELESS SENSOR NETWORKS. CHAPTER 3: RADIO COMMUNICATIONS Anna Förster

Modular Performance Analysis

Scheduling and Communication Synthesis for Distributed Real-Time Systems

Clock Synchronization

Wireless Networks (PHY): Design for Diversity

The physical layer in the CAN FD world The update

Bit Reversal Broadcast Scheduling for Ad Hoc Systems

UNIT 6 ANALOG COMMUNICATION & MULTIPLEXING YOGESH TIWARI EC DEPT,CHARUSAT

p-percent Coverage in Wireless Sensor Networks

Clock Synchronization

SYSTEM SENSOR WIRELESS REMOTE INDICATOR PRODUCT SPECIFICATION

Self-Stabilizing Deterministic TDMA for Sensor Networks

DOPPLER SHIFT. Thus, the frequency of the received signal is

ECE 333: Introduction to Communication Networks Fall Lecture 15: Medium Access Control III

WiMedia Interoperability and Beaconing Protocol

ECU with emulated partial networking functionality

Peripheral Sensor Interface for Automotive Applications

Temperature-Compensated Clock Skew Adjustment

Chapter 5. North American Cellular System Based on Time Division Multiple Access

SYNCHRONIZED PHASOR MEASUREMENT TECHNIQUES. A.G. Phadke

DYNAMIC CONFIGURATION IN A LARGE SCALE DISTRIBUTED SIMULATION FOR MANUFACTURING SYSTEMS

Message Scheduling Optimization for FlexRay Protocol

Outline for February 6, 2001

Ultra-Low Duty Cycle MAC with Scheduled Channel Polling

Energy Efficient Scheduling Techniques For Real-Time Embedded Systems

DEMONSTRATIONAL SYSTEM FOR TRAINING IN FlexRay COMMUNICATION

Robust Key Establishment in Sensor Networks

Outline. Wireless Networks (PHY): Design for Diversity. Admin. Outline. Page 1. Recap: Impact of Channel on Decisions. [hg(t) + w(t)]g(t)dt.

Peripheral Sensor Interface for Automotive Applications

Feasibility Studies of Time Synchronization Using GNSS Receivers in Vehicle to Vehicle Communications. Queensland University of Technology

Synchronous Atomic Broadcast for Redundant Broadcast. Channels. Flaviu Cristian. University of California, San Diego. La Jolla, CA

Wireless ad hoc networks. Acknowledgement: Slides borrowed from Richard Y. Yale

Data and Computer Communications. Tenth Edition by William Stallings

Reti di Telecomunicazione. Channels and Multiplexing

A Review of Current Routing Protocols for Ad Hoc Mobile Wireless Networks

ICT 5305 Mobile Communications. Lecture - 4 April Dr. Hossen Asiful Mustafa

WiNRADiO WR-G35DDCi Multichannel Coherent Application Guide

A Decentralized Network in Vehicle Platoons for Collision Avoidance

Resolving State Inconsistency in Distributed Fault-Tolerant Real-Time Dynamic TDMA Architectures

A Wireless Communication System using Multicasting with an Acknowledgement Mark

Chapter 3 : Media Access. Mobile Communications. Collision avoidance, MACA

Timing Analysis of the FlexRay Communication Protocol

COMET DISTRIBUTED ELEVATOR CONTROLLER CASE STUDY

Politecnico di Milano Facoltà di Ingegneria dell Informazione. 3 Basic concepts. Wireless Networks Prof. Antonio Capone

Optimization of RRU/UE Pairs for Multi-Point to Multi-UE Coordinated (MP2MUC) Transmissions

Contents. IEEE family of standards Protocol layering TDD frame structure MAC PDU structure

Design and analysis of electric vehicle battery management system based on flexray bus

Medium Access Control

Best Fit Void Filling Algorithm in Optical Burst Switching Networks

3.5: Multimedia Operating Systems Resource Management. Resource Management Synchronization. Process Management Multimedia

USB 3.1 ENGINEERING CHANGE NOTICE

Communication Analysis

Improved Directional Perturbation Algorithm for Collaborative Beamforming

Medium Access Schemes

SYSTEM SENSOR WIRELESS REPEATER PRODUCT SPECIFICATION

Multiple Access CHAPTER 12. Solutions to Review Questions and Exercises. Review Questions

QPLL a Quartz Crystal Based PLL for Jitter Filtering Applications in LHC

Mobile Computing. Chapter 3: Medium Access Control

Dependable Communication Synthesis for Distributed Embedded Systems *

Link State Routing. Stefano Vissicchio UCL Computer Science CS 3035/GZ01

David Grandblaise Voice: +33 (0) Motorola Fax: +33 (0)

Mathematical Problems in Networked Embedded Systems

Peripheral Sensor Interface for Automotive Applications

Task Allocation: Motivation-Based. Dr. Daisy Tang

3 Definitions, symbols, abbreviations, and conventions

Bloodhound RMS Product Overview

EN50160 Individual Report Summary

2 Polling. Politecnico di Milano Facoltà di Ingegneria dell Informazione

MODULE IV. End Sem. Exam Marks. Syllabus

Luca Schenato joint work with: A. Basso, G. Gamba

Distributed Control-as-a-Service with Wireless Swarm Systems"

Multi-Channel Design Considerations

Transcription:

Efficiency of Dynamic Arbitration in TDMA Protocols April 22, 2005 Jens Chr. Lisner

Introduction Arbitration methods in TDMA-based protocols Static arbitration C1 C1 C2 C2 fixed length of slots fixed schedule of senders configured Dynamic arbitration C1 C4 communication cycle C3 C3 communication cycle slots have dynamic length schedule determined at runtime for every cycle C1 C4 C5 C5 C9 C9 (left empty) (left empty) t t

Introduction Mixed-mode protocols provide both methods static part dynamic part C1 C2 C3 C1 C4 C5 C9 (left empty) C1 C2 C3 C1 C4 C5 C9 (left empty) t communication cycle Improved flexibility Ability to send additional information Example: exception handling

Fault-tolerance Common problem: babbling idiot fault Solution: Guardians protecting the channels from faulty controllers Independent guardians Each node has one guardian for each channel. Central guardians Guardians reside on a hub and are controlled by a protocol controller.

Fault-toerance Controllers guard each other Each controller acts as guardian for its neighbour Controllers are fully independent Controllers serve different hosts Guaranteed fail-silent behaviour in case of controller faults.

The Tea Protocol Mixed-mode media access regular part (static) extension part (dynamic) Double-controller architecture Extension schedule is determined in regular part Agreement-based scheduling Fault-tolerant and efficient solution

Time Model r j Deviation from nominal frequency of oscillator j (ppm) local time fastest clock ϕ n = 1 Clock speed ϕ j = 1 + r j cycle ϕ f < 1 slowest clock ϕ s > 1 Δ min { Δ max real time

Media Access Methods Minislotting with assigned slots Pre-configured schedule: Controllers are statically assigned to minislots Controllers may send or not C i+2 is sending Slot is expanded to the end of the message C i C i+1 C i+2 C i+3 t

Media Access Methods Minislotting with assigned slots: Timing constraints for slot of controller f d max maximum signal delay Slot as seen by f Δ max Δ max Signal arrives in this interval Earliest possible sending time d max Slot as seen by s t

Media Access Methods Minislotting with assigned slots Minimum minislot length ε λ minislot 2Δ max + d max + ε small error term (rounding error, descretization, ) Slot length if controller is sending λ message (i) length of message of controller i in real time λ slot (i) = λ message(i) λ minislot +1 λ minislot 2λ minislot

Media Access Methods Minislotting with assigned slots T as Overhead caused by the arbitration method Total length of dynamic part used for transmission m last T as = m last λ minislot last minislot Length of used slots k i=1 Overhead at start and end of used slots λ slot (i) + k2δ max

Media Access Methods Minislotting with relaxed timing sender message end dmax δ final receiver Slot n Signal arrival Slot n + 1 idle state detected after tranmission end: d max + δ final + ε

Media Access Methods Minislotting with relaxed timing T ls Overall overhead k last index of used slots sum of unused minislots h number of used slots T ls = (k h)λ minislot + h(δ final + d max + ε) Improvements: δ final is expected to be lower than Δ max end of transmission detection

Media Access Methods Agreement-based scheduling dynamic slot length sender d max δ final d switch receiver Signal arrival λ next δ final not necessary, if message reaches maximum length

Media Access Methods Agreement based scheduling Schedule previously known No additional arbitration mechanisms required h k T ts = h total number of controllers sending number of controller which do not fully utilize maximum slot length (k h) ( d + d + ε) + δ max switch final k + T reg additional overhead in regular part

Fault-tolerance Tea is able to tolerate double faults Possible behaviour of... Faulty controller - fail-silent - corrupted messages - babbling idiot Faulty channel - message corruption - byzantine behaviour Messages corruption missing messages messages with invalid CRC (or similar protection)

Fault-tolerance Case 1: next sender receives transmission on faulty channel A n denies access corrupted message d max δ final next sender starts sending B (no transmission) next sender sees corrupted message other controllers follow next sender

Fault-tolerance Case 2: next sender receives no transmission on faulty channel maximum slot length timeout next sender starts sending A B (no transmission) (no transmission) other controllers follow next sender

Examples Scenario 1: Maximum utilization cycle length Δ max 160 ms 12 µs ε d max 1 µs δ final 0.44 µs 1.1 µs d switch 0.25 µs High Δ max dominates in assigned slots and relaxed timing methods Network constants d max and δ max have minimal impact on overhead Improve clock synchronization for assigned slots and relaxed timing to reduce overhead

Examples Scenario 2: Overhead caused by empty slots Overhead in assigned slots and relaxed timing grow with number of empty slots (mainly depending on Δ max in longer cycles) Constant overhead for agreement-based scheduling

Conclusion Efficient solution for mixed-mode TDMA protocols available The Tea protocol uses pre-determined schedules for dynamic arbitration Agreement method overhead depends mainly on network constants Methods with assigned slots and relaxed timing require accurate clocks Agreement methods needs additional overhead for majority voting Agreement method requires minimum overhead compared to other methods