We propose in this paper a service description meta-model for describing services from a functional and non-functional perspectives. The model is inspired from the frame based modeling technique and is serialized in RDF (Resource Description Framework) using Linked Data principles. We apply this model for describing sensor services: modeling sensors and their readings enriched with non-functional properties. We also done a complete architecture for managing sensor data: collection, conversion, enrichment and storage. We tested our prototype using live streams of sensors readings. The paper also reports on the required time and storage size during the management and querying of sensor data.

1. Modelling and Querying Sensor
Services using Ontologies
Sana Baccar, Wassim Derguech, Edward Curry and Mohamed Abid
18th International Conference on Business Information Systems, BIS 2015

2. Are services’ capabilities well described?
Address check
service
Tracking service
Utility services: customer
identification, payment
processing, etc.
Rate query
service
Find services that “delivers a package” in programmableweb.com
Results under the shipping category
2

3. Sensor Data Integration and Management
• The need for the standardization of sensor data:
• Support dynamic features of sensor nodes.
• Capture relations between sensors capabilities.
• Easy to discover, re-use and exchange sensor capabilities.
• Sensor Data stream storage:
• Heavyweight and slow databases.
• Limitations in indexing and query forwarding techniques.
3

4. Objectives
1. Capability meta-model
1. Including explicit representation of the action performed
2. Independent from the capability implementation
3. Modelling functional and non-functional features
2. Sensor networks modelling and management
1. Propose an architecture for sensor data management using the
proposed model
2. Validation of the approach in a small smart environment
4

5. Objectives
1. Capability meta-model
1. Including explicit representation of the action performed
2. Independent from the capability implementation
3. Modelling functional and non-functional features
2. Sensor networks modelling and management
1. Propose an architecture for sensor data management using the
proposed model
2. Validation of the approach in a small smart environment
5

6. How can we properly describe capabilities for
human understanding and machine processing?
“Service Oriented Architecture (SOA) is a paradigm for organizing and
utilizing distributed capabilities that may be under the control of different
ownership domains.” 1
1 OASIS reference model for service oriented architecture 1.0.
Independent from their
implementation
Capabilities are stand
alone entities
Current capabilities modelling solutions model a capability as part of
concepts such as services, business processes, etc.
Capability are tight to their implementations (interaction interface)!
6

7. Solution: Frame-based modeling
• We model a capability as an action verb enriched by (zero
or many) functional or non-functional features
• An action verb is the most abstract capability
• Related features refine the given verb by giving more details
about the corresponding action.
7

8. Capability meta-model (1/2)
Attribute_value
hasValue
hasAttribute
Unit_Of_Measurement
hasUnit
Action_Verb
Capability
1..*
achieves
Attribute
1..*
8

9. Attribute_Value
Enumeration
Value
Constrained
Value
Constraint
Conditional
Value
Expression
Dynamic Value
Range Value
2..*
1
1 1
1 1
hasMin
hasMax
hasElement
constrainedBy
hasEvaluator
hasEvaluator
hasCondition
hasExpression
1
0..*
0..*
0..*
0..*
0..*
1
1
1 1
ExprType : String
ExprValue : String
Capability meta-model (2/2)
9

10. Attribute_Value
Enumeration
Value
Constrained
Value
Constraint
Conditional
Value
Expression
Dynamic Value
Range Value
2..*
1
1 1
1 1
hasMin
hasMax
hasElement
constrainedBy
hasEvaluator
hasEvaluator
hasCondition
hasExpression
1
0..*
0..*
0..*
0..*
0..*
1
1
1 1
ExprType : String
ExprValue : String
Capability meta-model (2/2)
• a Range_value specifies a range of values between a
min and max values
• used to specify flexible search requests
1 pckgWeight propertyValue weightValue.
2 weightValue a Range_Value.
3 weightValue hasMin 15.
4 weightValue hasMax 20.
5 weightValue hasStep 0,5.
10

11. Service meta-model (1/2)
NFP Service
0..*
Capability
1..*
11

12. Service meta-model (2/2)
12

13. Capability meta-model
Implementation and Use
Capability Meta‐
Model
Action Verb
Meta‐Model
Action Verbs
Capability Domain
Ontology
Capabilities
Meta‐Model:
Classes and
Properties
Action Verbs
Domain
Ontology:
Abstract
Capabilities
Capabilities
http://vocab.deri.ie/av
http://vocab.deri.ie/cap
Domain specific action
verbs
Domain specific action
set of abstract
capabilities
Capabilities’ instances
We ground our model to a set of RDF/RDFS vocabularies
13

14. Objectives
1. Capability meta-model
1. Including explicit representation of the action performed
2. Independent from the capability implementation
3. Modelling functional and non-functional features
2. Sensor networks modelling and management
1. Propose an architecture for sensor data management using the
proposed model
2. Validation of the approach in a small smart environment
14

15. High level architecture
Sensor Data
(real-time)
Raw-to-RDF
Sensor Data
Adaptation
Sensor Data
Warehousing
SPARQL querying
and Data
Visualization
Sensor
Readings
Sensor
Meta-Data
(NFP)
15

16. Actual Implementation
16

17. Actual Implementation
captures incoming data
from heterogeneous
data sources and
publish them in plain
text to a JMS
17

18. Actual Implementation
receives the collected
messages and filters
the irrelevant and
unnecessary
information
18

19. Actual Implementation
identifies possible triples in
the event-data, assigns a
URI to each identified
subject/predicate and
represents event-data in
RDF
19

20. Actual Implementation
Enriches the service with
the necessary
non-functional properties
retrieved from the Location
meta-data store
20

21. Actual Implementation
Stores the data to a
dedicate data store
Creates data aggregations
Removes old data
21

22. Output Data
22

23. Evaluation
• Real world sensors deployed within the Linked Energy
Intelligence (LEI) and Waternomics dataspace
• The dataspace has been realised in Insight building
• Sensors deployed : energy, motion detection, water
consumption, temperature and humidity
• Evaluation metrics:
• data storage space
• execution time for sensor data management
23

24. 3 Evaluation Experiments
 Size of different
sensor formats.
Experiment1 Experiment2
 Sensor Data Size
before and after
aggregation.
Experiment3
Query Execution Time:
Q1: Search per 12 minutes the
average of the water-usage in the
kitchen.
Q2: Search per minute the sum of
the water-usage in the showers.
Q3: Search the min and the max of
the water-usage captured in the
showers during the last 10Minutes.
24

25. 3 Evaluation Experiments
 Size of different
sensor formats.
Experiment1
25

26. 3 Evaluation Experiments
Experiment2
 Sensor Data Size
before and after
aggregation.
26

27. 3 Evaluation Experiments
Experiment3
Query Execution Time:
Q1: Search per 12 minutes the
average of the water-usage in the
kitchen.
Q2: Search per minute the sum of
the water-usage in the showers.
Q3: Search the min and the max of
the water-usage captured in the
showers during the last 10Minutes.
27

28. 3 Evaluation Experiments
28

29. Conclusion
Defined a new service description meta-model: capability +
NFPs.
Applying the proposed meta-model to describe sensor
services
Define and implement an architecture for sensor data
management, warehousing and manipulation
Validated the applicability and the effectiveness of our
proposed modeling approach in a small smart environment in
producing standardized data using RDF.
29

30. Future Work
Extending the proposed system to:
Handle service composition task and support complex queries
Process real-time data in large environments
Generate additional links between our data store to open data for more
knowledge discovery
Explore other big data architecture styles: Lambda Architecture
30

31. @WATERNOMICS_EU www.waternomics.eu
Project co-funded by the European
Commission within the 7th Framework
Program (Grant Agreement No. 619660)
THE RESEARCH LEADING TO THESE RESULTS HAS
RECEIVED FUNDING UNDER THE EUROPEAN
COMMISSION'S SEVENTH FRAMEWORK PROGRAMME
FROM ICT GRANT AGREEMENT WATERNOMICS NO. 619660.