IMPLEMENTATION OF LAMPORT’S SCALAR CLOCKS

Surekha Busa

Contents

Introduction Implementation Rule Experimental Setup Result Analysis Observation Future Work

INTRODUCTION

To show casualty in distributed system Each process Pi has a logical clock Ci assigned with

an integer the assigned value is the timestamp of this event,

denoted C(a) The timestamps are monotonically increasing consistency condition

consistency: if ab, then C(a) C(b) if event a happens before event b, then the clock value

(timestamp) of a should be less than the clock value of b strong consistency: if C(a)C(b) then ab

scalar clocks are not strongly consistent

IMPLEMENTATION RULE

Outgoing messages tagged with Sender’s clock value

Upon receive of message: myclock = MAX(message, myclock)

EXPERIMENTAL SETUP

Base Algorithm : Random Flood. Whenever the clock of the receiver take the value

of the clock of the message we say the clock is updated.

Total No. of updates per process(number of updates/no of processes) is calculated by increasing the number of processes from 10-100.

P1e11

(1)

e12

(2)

e13 e14

(4)

e15

(5)

P2e21

(1)

e22

(2)

e23

(4)

(3)

Clock is Updated

Clock is not Updated

RESULT ANALYSIS: No. Of Updates and Messages Varying No. Of Processes

RESULT ANALYSIS: No. Of Updates and Messages Per Process Varying No. Of Processes

RESULT ANALYSIS: No. Of Messages Versus No. Of Clock Updates

No. Of Processes

No. Of Updates

No. Of Messages

Updates Per Process

Messages Per Process

10 13 46 1.3 4

20 41 197 2.05 9

30 65 442 2.17 14

40 95 829 2.37 20

50 131 1230 2.62 24

60 169 1724 2.81 28

70 206 2509 2.94 35

80 240 3238 3 40

90 278 4081 3.08 45

100 319 4938 3.19 49

OBSERVATION

No. Updates and messages per process increases with increase in No. of Processes.

No. of clock updates increase with increase in No. of Messages.

FUTURE WORK

To observe its performance over various topologies.

To test my program with much larger no. of processes.

Test my program on a real distributed network.

REFERENCE

[1] Leslie Lamport (1978). ”Time, clocks, and the ordering of events in a distributed system”. Communications of the ACM 2 (7): 558-565.

[2] Mukesh Singhal,Niranjan G. Shivaratri.“Advanced Topics in Operating Systems”. McGRAW-Hill international edition:Chapter 5.

Code Defence

THANK YOU..!!