A
Generic Framework
forQuality
-based AutonomicAdaptation
withinSensor
-basedSystems
ICSOC ASOCA 2016, Banff (Canada)
Antoine Auger, Ernesto Exposito, Emmanuel Lochin
Introduction
✓ A World of sensors
✓ Sensor-based systems (sensor middlewares and IoT platforms)
Introduction
✓ A World of sensors
✓ Sensor-based systems (sensor middlewares and IoT platforms)
➡ Observation quality is often delegated to
Introduction
✓ A World of sensors
✓ Sensor-based systems (sensor middlewares and IoT platforms)
➡ Observation quality is often delegated to
applications
➡ How to provide relevant observations to
applications that have various and evolutive needs?
Required background
What is sensor observations?
Required background
What is sensor observations?
(see also the "DIKW ladder", [Sheth2016])
Required background
What is sensor observations?
(see also the "DIKW ladder", [Sheth2016])
Sensor Raw Data
Sensor Information =
=
Sensor Raw Data + Context information
Required background
What is sensor observations?
(see also the "DIKW ladder", [Sheth2016])
Sensor Raw Data Sensor Information
Sensor Knowledge =
=
=
Sensor Raw Data + Context information
+ Semantics
Sensor Raw Data Sensor Information
Required background
What is sensor observations?
(see also the "DIKW ladder", [Sheth2016])
Sensor Raw Data Sensor Information
Sensor Knowledge =
=
=
Sensor Raw Data + Context information
+ Semantics
Sensor Raw Data Sensor Information
Data coming from sensors
Sensor observations
Required background
How to characterize observation quality? ➡ Quality of Service (QoS) but…
Required background
How to characterize observation quality? ➡ Quality of Service (QoS) but…
Other approaches:
➡ Quality of Information (QoI) [Bisdikian2009] ➡ Context information [Dey2001]
Required background
The Autonomic Computing paradigm: Autonomic Manager Managed Element Managed Element touchpoint Sensors Effectors Execute Monitor Plan Analyze Autonomic Element Knowledge
Required background
Autonomic maturity levels (from [Jacob2004]): (1) Basic (2) Managed (3) Predictive (4) Adaptive (5) Autonomic ➡ hard-coded behavior ➡ basic rules
➡ fusion, machine learning ➡ Service Level Agreements ➡ business rules
Framework proposal
Preliminary work: we studied 32 context-aware systems
• Observation characterization
(network QoS, QoI, Context)
• Observation modelling
(ontologies)
• Adaptation strategy
Temperature sensor Temperature sensor Humidity sensor Sensor middleware or IoT platform Topic: temperature Location: Banff Application
Temperature sensor Temperature sensor Humidity sensor Sensor middleware or IoT platform 1) Request Topic: temperature Location: Banff Application
Temperature sensor Temperature sensor Humidity sensor Sensor middleware or IoT platform 2) Retrieval / binding Topic: temperature Location: Banff Application
Temperature sensor Temperature sensor Humidity sensor Sensor middleware or IoT platform 2) Retrieval / binding Topic: temperature Location: Banff Application
Temperature sensor Temperature sensor Humidity sensor Sensor middleware or IoT platform 3) Response Topic: temperature Location: Banff Application
Temperature sensor Temperature sensor Humidity sensor Sensor middleware or IoT platform Topic: temperature Location: Banff … observation observation observation … Application
Temperature sensor Temperature sensor Humidity sensor Sensor middleware or IoT platform Topic: temperature Location: Banff Application
Temperature sensor Temperature sensor Humidity sensor Sensor middleware or IoT platform Topic: temperature Location: Banff
But what if…
• application needs change
Temperature sensor Temperature sensor Humidity sensor Sensor middleware or IoT platform
But what if…
• application needs change
Temperature sensor Temperature sensor Humidity sensor Sensor middleware or IoT platform
But what if…
• application needs change
Application
Temperature sensor Humidity sensor Sensor middleware or IoT platform
But what if…
• application needs change
Application
Temperature sensor Humidity sensor Sensor middleware or IoT platform
But what if…
• application needs change Application … observation observation observation … • sensor fails • poor observations
Temperature sensor Humidity sensor Sensor middleware or IoT platform
But what if…
• application needs change
Applications are asked to
adapt themselves and
Application … observation observation observation … • sensor fails • poor observations
Framework proposal
➡ How to provide relevant observations to
Framework proposal
➡ How to provide relevant observations to
applications that have various and evolutive needs?
Framework proposal
➡ How to provide relevant observations to
applications that have various and evolutive needs?
Content and quality
Framework proposal
➡ How to provide relevant observations to
applications that have various and evolutive needs?
Adaptation Content and quality
Temperature sensor Temperature sensor Humidity sensor Application API
Temperature sensor Temperature sensor Humidity sensor Application
Raw Data layer
Collection and digitization of phenomena/events
Topic: temperature Location: Banff Level: Raw Data
Temperature sensor Temperature sensor Humidity sensor Application
Raw Data layer
Collection and digitization of phenomena/events
Topic: temperature Location: Banff Level: Raw Data
API
{sensor_id: 34, value: 20}
Temperature sensor Temperature sensor Humidity sensor Application
Raw Data layer
Collection and digitization of phenomena/events
Information layer
Raw Data characterization
Temperature sensor Temperature sensor Humidity sensor Application
Raw Data layer
Collection and digitization of phenomena/events
Information layer
Raw Data characterization
API
Topic: temperature Location: Banff Level: Information
Temperature sensor Temperature sensor Humidity sensor Application
Raw Data layer
Collection and digitization of phenomena/events
Information layer
Raw Data characterization
API Topic: temperature Location: Banff Level: Information {sensor_id: 34, value: 20, unit: Celsius, location: (43.564509, 1.468910), accuracy: 0.8}
Temperature sensor Temperature sensor Humidity sensor Application
Raw Data layer
Collection and digitization of phenomena/events
Information layer
Raw Data characterization
API
Semantic layer
Temperature sensor Temperature sensor Humidity sensor Application
Raw Data layer
Collection and digitization of phenomena/events
Information layer
Raw Data characterization
API Topic: temperature Location: Banff Level: Semantic Semantic layer Semantic annotation
Temperature sensor Temperature sensor Humidity sensor Application
Raw Data layer
Collection and digitization of phenomena/events
Information layer
Raw Data characterization
API Topic: temperature Location: Banff Level: Semantic {sensor_type: temperature, value: comfort, location: room3, accuracy: good} Semantic layer Semantic annotation
Temperature sensor Temperature sensor Humidity sensor Application
Raw Data layer
Collection and digitization of phenomena/events
Information layer
Raw Data characterization
API
Semantic layer
Temperature sensor Temperature sensor Humidity sensor Application
Raw Data layer
Collection and digitization of phenomena/events
Information layer
Raw Data characterization
API Semantic layer Semantic annotation Auto. Manager Auto. Manager Auto. Manager
Temperature sensor Temperature sensor Humidity sensor Application
Raw Data layer
Collection and digitization of phenomena/events
Information layer
Raw Data characterization
API Semantic layer Semantic annotation Auto. Manager Auto. Manager Auto. Manager
Temperature sensor Temperature sensor Humidity sensor Application
Raw Data layer
Collection and digitization of phenomena/events
Information layer
Raw Data characterization
API Semantic layer Semantic annotation Auto. Manager Auto. Manager Auto. Manager
Framework usage
Guidelines to design an autonomic quality-aware sensor-based system with our framework:
1- What will be asked by applications / users? 2- How my system will adapt its behavior?
3- General use cases definition
4- What metrics for observation quality?
Conclusions
Quality of Observations (QoO) is critical within information-centric systems
Conclusions
Quality of Observations (QoO) is critical within information-centric systems
Conclusions
Quality of Observations (QoO) is critical within information-centric systems
Most of the time, delegated to applications
Generic framework to build quality-aware sensor-based systems:
✓ Autonomic adaptation based on QoO
✓ Applications can focus on their primary goal ✓ Framework generic and customizable
Perspectives
We are building an integration platform for Quality of Information Assessment as a Service (iQAS)
A Generic Framework for Quality-based
Autonomic Adaptation within Sensor-based
Systems
Antoine Auger
antoine.auger@isae.fr
References
[Sheth2016] Sheth, A.: Internet of Things to Smart IoT Through Semantic, Cognitive,
and Perceptual Computing. IEEE Intelligent Systems 31(2), 108–112 (Mar 2016)
[Bisdikian2009] Bisdikian, C., Branch, J., Leung, K., Young, R.: A letter soup for the
quality of information in sensor networks. In: IEEE International Conference on Pervasive Computing and Communications, 2009. PerCom 2009. pp. 1–6 (Mar 2009)
[Dey2001] Dey, A.K.: Understanding and using context. Personal and ubiquitous
computing 5(1), 4–7 (2001)
[Kephart2003] J. O. Kephart and D. M. Chess, “The vision of autonomic computing,”
Computer, vol. 36, no. 1, pp. 41–50, 2003.
[Jacob2004] Jacob, B., Lanyon-Hogg, R., Nadgir, D.K., Yassin, A.F.: A practical guide
to the IBM autonomic computing toolkit. IBM, International Technical Support Organization (2004)