Pub/Sub Systems Raphael Gomes 1 , Georgios Bouloukakis 2,3 , Fbio - - PowerPoint PPT Presentation

pub sub systems
SMART_READER_LITE
LIVE PREVIEW

Pub/Sub Systems Raphael Gomes 1 , Georgios Bouloukakis 2,3 , Fbio - - PowerPoint PPT Presentation

QoS-Aware Resource Allocation for Mobile IoT Pub/Sub Systems Raphael Gomes 1 , Georgios Bouloukakis 2,3 , Fbio Costa 4 , Nikolaos Georgantas 3 , & Ricardo da Rocha 5 Seattle, USA, June 2018 International Conference on Internet of Things


slide-1
SLIDE 1

QoS-Aware Resource Allocation for Mobile IoT Pub/Sub Systems Raphael Gomes1, Georgios Bouloukakis2,3, Fábio Costa4, Nikolaos Georgantas3, & Ricardo da Rocha5

Seattle, USA, June 2018

International Conference on Internet of Things (ICIOT)

1Instituto Federal de Goiás - Câmpus Goiânia, Brazil

2Donald Bren School of Information and Computer Sciences, UC Irvine, USA 3MiMove team, Inria Paris, France 4Instituto de Informática, Universidade Federal de Goiás, Brazil 5Instituto de Biotecnologia, Universidade Federal de Goiás, Catalão, Brazil

slide-2
SLIDE 2

Motivation

TIM system

Dynamic Heterogeneous

  • Traffic Information Management (TIM) system:

International Conference on Internet of Things (ICIOT)

  • 2
slide-3
SLIDE 3
  • 3

network discon. voluntary discon. OFF

Motivation

publishers message broker

broker 1 broker 2 broker 3

subscribers

connect connect ON ON

  • Message broker network can be allocated on Cloud:
  • Can we ensure specific end-to-end QoS between publishers and subscribers?
  • Does the Things’ intermittent connectivity affects the cost of resource allocation?

International Conference on Internet of Things (ICIOT)

slide-4
SLIDE 4

Publish/Subscribe System

  • 4

p1 p2 s2 s1 s4 s3 p3 p4

b1 b3 b2 b4 b6 b5 b7 b9 b8

subscriptions partitioning event matching

brokers publishers subscribers

1 R. Baldoni et al., “Distributed event routing in publish/subscribe communication systems:

a survey,” DIS, Universita di Roma La Sapienza, Tech. Rep, 2005. home broker home broker

event routing process

slide-5
SLIDE 5

Resource Synthesis with Intermittent Connectivity

  • 5

International Conference on Internet of Things (ICIOT)

  • How can we allocate message brokers of the pub/sub system ?
  • Use of the WS-HEU heuristic to efficiently select the resource types.
  • Additional message broker instances, if necessary.

cost ($) ? p → s QoS ?

b7 b2 b5

Q1 Q2 Q3

Multi- dimension Knapsack problem

slide-6
SLIDE 6

End-to-end QoS Estimation

  • 6

International Conference on Internet of Things (ICIOT)

  • We rely on Queueing Theory to estimate end-to-end response times between

publishers and subscribers.

  • We use 3 different queueing models:

M/M/c Queue1

𝜈 𝜈

M/M/1 Queue1 ON/OFF Queue2

𝜈

s1’s home broker

𝑞2, 𝑡1

𝜇 𝑞2 𝑡1

𝑐2 𝑐5 𝑐7

𝑐2

𝜇𝑗𝑜

𝑐2

𝜇𝑗𝑜 𝑝𝑢ℎ

  • 𝑐2

𝜇𝑝𝑣𝑢 𝑝𝑢ℎ

  • 𝜈

𝜈

𝑐5

𝜇𝑗𝑜

𝑐5

𝜇𝑗𝑜 𝑝𝑢ℎ

  • 𝑐5

𝜇𝑝𝑣𝑢 𝑝𝑢ℎ

  • 𝜈

𝜈

𝑐7

𝜇𝑗𝑜

𝑐7

𝜇𝑗𝑜 𝑝𝑢ℎ

  • 𝑐7

𝜇𝑝𝑣𝑢 𝑝𝑢ℎ

  • 𝜈

𝜈

  • End-to-end response time between p2, s1:

1 D. Gross et al., Fundamentals of queueing theory, 2008. 2 G. Bouloukakis et al., ICC 2017, ICPE 2017.

slide-7
SLIDE 7

Experimental Evaluation

  • 7

1https://github.com/raphaeldeaquino/mcloudsim

resource parameters

𝑤1 𝑤|𝑊| …

pub/sub parameters

p1 p2 s2 s1 s4 s3 p3 p4

b1 b3 b2 b4 b6 b5 b7 b9 b8 QoS estimation

Δ𝑞2, 𝑡1

𝑐2, 𝑤1 , …., Δ𝑞2, 𝑡1 𝑐2, 𝑤

|𝑊|

….

resource synthesis

end-to-end QoS requirement

Δthr

synthesis result b2 b5 b7

𝑤7 𝑤7 𝑤1 𝑤3 𝑤3 𝑤3 M-CloudSim1

simulated environment

p1 p2 s2 s1 s3 p3

slide-8
SLIDE 8

Evaluation Setup & Results (1)

  • 8

International Conference on Internet of Things (ICIOT)

event size (E) = 200 bytes λ = 0.5 to 23 events/sec

1. always connected 2. intermittent connected

VMs by Amazon, Microsoft, Google μ = 33.2 CPU instructions Δthr = 0.025 sec

  • Subscriber always connected:
slide-9
SLIDE 9

Evaluation Results (2)

  • 9

International Conference on Internet of Things (ICIOT)

  • Subscriber intermittent connected: TON = 155 sec, TOFF = 96 sec.
  • The response time is tightly coupled by the subscriber’s intermittent connectivity.
  • By slightly describing Δthr:
  • 1. the cost increases significantly.
  • 2. the response time remains at the same level.
slide-10
SLIDE 10

Next steps

  • 10
  • We allocate Cloud resources of IoT applications by considering end-to-

end QoS requirements and the Things’ intermittent connectivity.

  • Future work:
  • Introduction of additional QoS parameters: e.g., memory.
  • Estimation of end-to-end QoS between peers by using well known

pub/sub routing algorithms.

  • Resource allocation of interoperability software artifacts.

International Conference on Internet of Things (ICIOT)

slide-11
SLIDE 11

Thank you

  • 11

International Conference on Internet of Things (ICIOT) International Conference on Internet of Things (ICIOT)