Unreliable Publish/Subscribe Protocols Georgios Bouloukakis 1,3 , - - PowerPoint PPT Presentation

▶
unreliable publish subscribe protocols
SMART_READER_LITE
LIVE PREVIEW

Unreliable Publish/Subscribe Protocols Georgios Bouloukakis 1,3 , - - PowerPoint PPT Presentation

Mar 13, 2023 277 likes •422 views

Queueing Network Modeling Patterns for Reliable and Unreliable Publish/Subscribe Protocols Georgios Bouloukakis 1,3 , Ajay Kattepur 2 , Nikolaos Georgantas 3 & Valrie Issarny 3 New York, USA, November 2018 15th EAI International Conference

slide-1
SLIDE 1

Queueing Network Modeling Patterns for Reliable and Unreliable Publish/Subscribe Protocols Georgios Bouloukakis1,3, Ajay Kattepur2, Nikolaos Georgantas3 & Valérie Issarny3

New York, USA, November 2018

15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services (MobiQuitous)

1Donald Bren School of Information and Computer Sciences, UC Irvine, USA 2Embedded Systems & Robotics, TCS Research & Innovation, India 3MiMove team, Inria Paris, France

slide-2
SLIDE 2

Motivation

  • 2

TIM system

Dynamic Heterogeneous

  • Traffic Information Management (TIM) system:

15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services (MobiQuitous)

slide-3
SLIDE 3

Publish/Subscribe (pub/sub) QoS features

  • 3

publisher 1 publisher 2 subscriber 1 subscriber 2

broker

publisher’s disconnections subscriber’s disconnections network issues disconnection

15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services (MobiQuitous)

  • Protocols and APIs: MQTT, AMQP, JMS, …
  • IoT applications often employ such protocols and their QoS features:

1. How to evaluate the peers’ end-to-end performance? 2. How to enable system parameter tuning to IoT applications running over pub/sub?

slide-4
SLIDE 4

Analysis of Pub/Sub Interactions

  • 4

15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services (MobiQuitous)

tpublish t’’publish

lifetime lifetime

tsubscribe publisher (p) subscriber (s) broker (b)

QoSu

lifetime

QoSu QoSr

tsubscribe

QoSu QoSu QoSu QoSr

t’publish

QoSr

lifetime

t’’’publish

lifetime

QoSr

lifetime

TON

s

TOFF

s

TON

s

TON

s

TOFF

s

QoSr

lifetime

QoSu

x

QoSu

x TOFF

b

TON

b

TON

b

TOFF

b

TON

b

TON

p

TON

p

TOFF

p

TON

p

TOFF

p

trace 1 trace 2 trace 3 trace 4

slide-5
SLIDE 5

Pub/Sub Interactions across Multiple Layers

  • 5

15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services (MobiQuitous)

  • We introduce a performance modeling pattern (PerfMP) with realistic

constraints found across multiple layers in the IoT

APP

connection/disconnection

PUB/SUB NET

limited data lifetime finite capacity buffers app processing delay finite capacity buffers reliable/unreliable protocols mdw processing delay finite capacity buffers transmission delay disconnections connection/disconnection

  • System designers can derive values for end-to-end response times and delivery

success rates

slide-6
SLIDE 6

Base queueing models for Pub/Sub interactions

  • 6

15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services (MobiQuitous)

𝜇𝑗𝑜

𝜈

𝜇𝑝𝑣𝑢 continuous queue 𝜇𝑝𝑣𝑢 TON / TOFF 𝜇𝑗𝑜

𝜈

intermittent (ON/OFF) queue1,2 𝜇𝑗𝑜

𝜈

𝜇𝑝𝑣𝑢

lifetime expired event valid event

messages exp. queue

  • Additional features:

𝜇𝑗𝑜

𝜈

𝜇𝑝𝑣𝑢

dropped event buffer

finite capacity queue

1 G. Bouloukakis et al., ICC, 2017 2 G. Bouloukakis et al., ICPE, 2017

  • We model the end-to-end path of a pub/sub interaction by using a

combination of different types of queueing models

slide-7
SLIDE 7

PerfMP for Pub/Sub Interactions

  • 7

15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services (MobiQuitous)

  • We model reliable or unreliable interactions by using our queueing models

UNRELIABLE

M/M/1 ON/OFF M/M/1 M/M/1

If event == expired lifetime

λapp

in

If mdw-OFF == true If subs-OFF == true If event == expired

λapp

  • ut

M/M/1 ON/OFF

PUBLISHER BROKER SUBSCRIBER

If mdw-OFF == true || event == expired

event == expired

RELIABLE

M/M/1 ON/OFF M/M/1 M/M/1

If event == expired lifetime

λapp

in

If event == expired

λapp

  • ut

M/M/1 ON/OFF event == expired

event == expired

Publisher’s or subscriber’s disconnections End-to-end disconnection pattern between publisher-broker or broker-subscriber

slide-8
SLIDE 8

Evaluation Results

  • 8
  • JINQS (Java Implementation of a Network-of-Queues Simulation):
  • pen source simulator for building queueing networks
  • We extend JINQS to implement:
  • ON/OFF queue, reliable/unreliable event transmission, other features
  • Our proposed PerfMP
  • Evaluate the trade-off between response times delivery success rates for

numerous reliable/unreliable interactions

  • System parameter tuning for TIM

1 http://xsb.inria.fr/d4d#mobilejinqs

15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services (MobiQuitous)

slide-9
SLIDE 9

p reliable to s reliable: success rates

  • 9

15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services (MobiQuitous)

TON + TOFF = 80 sec

pub pub

TON + TOFF = 30 sec

sub sub

lifetime = 10, 20 and 30 sec λapp = 2 events/sec

in

  • Success Rate 39%
  • Success Rate 63%

QoSr QoSr

slide-10
SLIDE 10

p reliable to s reliable: response times

  • 10

15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services (MobiQuitous)

  • Lower lifetime periods produce improved response time (but with lower

success rates)

  • Success Rate 39% and Response Time within 5 sec. with Prob = 0.78
  • Success Rate 63% and Response Time within 5 sec. with Prob = 0.45
slide-11
SLIDE 11

p reliable to s unreliable

  • 11

15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services (MobiQuitous)

TON + TOFF = 80 sec

pub pub

  • 1. lifetime = 10sec

λapp = 2 events/sec

in

TON + TOFF = 30 sec

sub sub

1. Success Rate 30% and Response Time within 10 sec. with Prob = 1 2. Success Rate 38% and Response Time within 10 sec. with Prob = 0.78

  • 2. lifetime = 30sec
  • QoSu feature at the subscriber side provides markedly improved response times.
  • success rates are severely bounded only by the subscriber’s disconnections.

QoSr QoSu

slide-12
SLIDE 12

TIM system tuning

  • 12

15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services (MobiQuitous)

TON = 50 sec TOFF = 40 sec

pub pub

λapp = 20 events/sec

in

QoSr QoSu TON = 20 sec TOFF = 10 sec

sub sub

  • 1. lifetime = 10sec

1. Success Rate 37% and average Response Time 2.1 sec 2. Success Rate 47% and average response time 9.49 sec

  • 2. lifetime = 30sec
  • The designer now applies finite capacity buffers (K)
  • 1. K = 10

1. Success Rate 42% and average Response Time 0.43 sec 2. Success Rate 44% and average Response Time 1.8 sec

  • 2. K = 50

lifetime = 30sec

slide-13
SLIDE 13

Next steps

  • 13
  • We introduce a performance modeling pattern (PerfMP) that captures

the application and middleware layers of pub/sub protocols.

  • System designers can rely on our PerfMP to evaluate the performance
  • f multiple IoT applications running over pub/sub.
  • Future work
  • Introduce analytical models for applications applying lifetime

periods to each published event.

  • Use our analysis for system tuning at runtime.

15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services (MobiQuitous)

slide-14
SLIDE 14

Thank you

  • 14

15th EAI International Conference on Mobile and Ubiquitous Systems: Computing, Networking and Services (MobiQuitous)