2. CoAP-CTX: A Context-Aware CoAP Extension for Smart Objects Discovery in Internet of Things
COMPSAC 2017
Authors:
Msc. Felipe Mota Barreto
PHD Student Paulo Artur Duarte
Prof. Dr. Marcio Espíndola Freire Maia
Prof. Dra. Rossana Maria de Castro Andrade
Prof. Dr. Windson Viana de Carvalho

3. Federal University of Ceará
Fortaleza, Ceará, Brazil
More than 2 millions habitants
UFC – Federal University of Ceará

4. Agenda Contextualization Smart Object (SO) Discovery CoAP CoAP-CTX
Proof of Concept
Evaluation
Final Considerations
Explicação bem rápida sobre os tópicos da Agenda, apenas citando os tópicos.

5. Motivation Scenario
Smart House

6. Motivation Scenario User TV Sport fan
Watch football games on Thursday night TV
Brasil x Argentina

7. Motivation Scenario

8. Motivation Scenario

9. Motivation Scenario

10. Problem Statement
The primary goal of Smart Object (SO) discovery services is to make them reachable
Environment remote controller will find ALL available SO

11. Problem Statement

12. Problem Statement

13. Oportunity Context-awareness is a reality
The discovery process ought to consider
The functional or non-functional characteristics of each SO
The user and devices’ contexts
Reducing the search space and improves the energy footprint

14. Research main goal
Create a Smart Object Discovery service that is able to find the SOs that are relevant to an user in a given context.

15. Target Application: Environment Browsers
Traditional SO Discovery
Context-Aware SO Discovery

16. Agenda Motivation Smart Object Discovery CoAP CoAP-CTX
Proof of Concept
Evaluation
Final Considerations
Explicação bem rápida sobre os tópicos da Agenda, apenas citando os tópicos.

17. Smart Object Discovery1
Requirements for SO Discovery by Guinard et al.
R1 - Low overhead to register an SO
R2 - Low overhead in the SO discovery
R3 - Context-aware discovery
R4 - On-demand allocation 1.
D. Guinard, V. Trifa, S. Karnouskos, P. Spiess, and D. Savio. Interacting with the soa-based internet of things: Discovery, query, selection, and on-demand provisioning of web services. IEEE Transactions on Services Computing, 3(3):223{235, July 2010

18. Smart Object Discovery1
Requirements for SO Discovery by Guinard et al.
R1 - Low overhead to register an SO
R2 - Low overhead in the SO discovery
R3 - Context-based discovery
R4 - On-demand allocation
CoAP 1.
D. Guinard, V. Trifa, S. Karnouskos, P. Spiess, and D. Savio. Interacting with the soa-based internet of things: Discovery, query, selection, and on-demand provisioning of web services. IEEE Transactions on Services Computing, 3(3):223{235, July 2010

19. Smart Object Discovery1
Requirements for SO Discovery by Guinard et al.
R1 - Low overhead to register an SO
R2 - Low overhead in the SO discovery
R3 - Context-based discovery
R4 - On-demand allocation
CoAP-CTX 1.
D. Guinard, V. Trifa, S. Karnouskos, P. Spiess, and D. Savio. Interacting with the soa-based internet of things: Discovery, query, selection, and on-demand provisioning of web services. IEEE Transactions on Services Computing, 3(3):223{235, July 2010

20. CoAP – Constrained Application Protocol

21. Agenda Contextualização Descoberta de Objetos Inteligentes CoAP
CoAP-CTX
Prova de Conceito e Avaliação
Considerações Finais
Explicação bem rápida sobre os tópicos da Agenda, apenas citando os tópicos.

22. Context-Aware Discovery Process

24. COAP-CTX Architecture

25. Agenda Contextualization Smart Object Discovery CoAP CoAP-CTX
Proof of Concept and Evaluation
Final Considerations
Explicação bem rápida sobre os tópicos da Agenda, apenas citando os tópicos.

26. Proof of Concept Android Application
Based on location and history
Arduinos used to enable the communication with real devices
2 TVs; 1 Air Conditioning; 1 Luminosity Sensor
CoAP Server based on the open java implementation called jCoAP
Context
The application users generally wants to access the actuators as they arrive at home (if=“actuator”). Thus, the sensors (if=“sensor”) can enter in an IDLE mode to save resources. 1 1.
Available in:

27. Android Application (a) Performing Discovery (b) SO Discovered
(c) TV Screen
(d) AC Screen

28. Evaluation Objectives Materials and Methods
Analyze the scenarios where CoAP-CTX managed to reduce the total number of messages exchanged on the network
Measure the overhead on the discovery time brought by CoAP-CTX with respect to traditional CoAP.
Materials and Methods
Cooja Network Simulator
Linux Virtual Machine
OI simulated using Tmote Sky nodes
Communication module CC2420 (compatible with IEEE standard)

29. Evaluation Questions
EQ1: In which scenarios CoAP-CTX reduces the total number of messages exchanged on the network?​
EQ2: The overhead on the discovery time makes the CoAP-CTX infeasible?

30. Evaluation Scenarios Ambients Smart House Smart Car Smart Building
Cases
Best Case => just one final SO
Intermediate Case => depends on context
Worst Case => all SO discovered
10 simulations of 15 minutes each for every case of every ambient.

31. Evaluation Scenarios Ambients Smart House Smart Car Smart Building
Cases
Best Case => just one final SO
Intermediate Case => depends on context
Worst Case => all SO discovered
10 simulations of 15 minutes each for every case of every ambient.
Simulations
Worst Case
Intermediate Case
Best Case
Total
House 10 30
Car
Building 90

32. Intermediate Cases Configurations House Car Building CoAP Clients 1
CoAP Servers 3
Smart Objects 20 10
100
Pre-selection 15 4 70
Active SO
IDLE SO 7 85

33. Networks Simulated CoAP Server CoAP Client Smart Objects
(a) Smart Home
(b) Smart Car
(c) Smart Building
CoAP Server
CoAP Client
Smart Objects

34. Smart Home

35. Smart Home
+100 ms
-41%

36. Smart Car

37. Smart Car
+63 ms
-57%

38. Smart Building

39. Smart Building
+1,5s
-80%

40. Agenda Contextualization Smart Object Discovery CoAP CoAP-CTX
Proof of Concept and Evaluation
Final Considerations
Explicação bem rápida sobre os tópicos da Agenda, apenas citando os tópicos.

41. Final Considerations
The CoAP-CTX extends the CoAP discovery service to add support to:
Context-Aware Smart Objects Discovery
IDLE mode to Smart Objects that are not of the user’s interest.
EQ1: In which scenarios CoAP-CTX reduces the total number of messages exchanged on the network?​
EQ2: The overhead on the discovery time makes the CoAP-CTX infeasible?

42. Final Considerations
EQ1: In which scenarios CoAP-CTX reduces the total number of messages exchanged on the network?​
In scenarios where the user’s interest can be well defined, allowing the contextual rules to better identify the target SO for the discovery.
This leads to a high number of SO in IDLE mode

43. Final Considerations
EQ2: The overhead on the discovery time makes the CoAP- CTX infeasible?
The overhead on the discovery time was very dependent on the number of smart objects pre- selected.
The more generic an ambient is, more SO needs to be verified which is very time consuming.
The maximum overhead was 1.5 seconds, which is an acceptable value for non real time applications

44. Future Work
Allow definition of contextual rules to be extended and modified by the application in the development time
CoAP-CTX as a framework
Migration of contextual rules to the CoAP Server. The server itself will perform the rules for pre-selection
This will reduces dramatically the discovery time overhead in scenarios where the number of SO is high
This approach modifies the CoAP Server specifications, which leads to standard compatibility issues.

45. Thanks!