The Contextor Infrastructure for Context-Aware Computing

1
The Contextor Infrastructure for Context-Aware
Computing

• Gaëtan Rey, Joëlle Coutaz
• Engineering HCI research group
• CLIPS IMAG, Grenoble, France

2
I-AM
3
Outline

• Synthesis of the state of the art (architectural
  aspects)
• The contextor infrastructure

4
Synthesis of the state of the art

• No consensual definition for the notion of
  context
• A clear need for infrastructures that support the
  development of context-aware computing
• Like conventional systems heterogeneity,
  distribution, interoperability,
• Unlike conventional systems multi-scaling,
  spontaneity,

5
Architecture
Sensing numeric observables
6
Architecture
Transformation symbolic observables
Sensing numeric observables
7
Architecture
Situation and context identification
Transformation symbolic observables
Sensing numeric observables
8
Architecture
Context adapter
Situation and context identification
Transformation symbolic observables
Sensing numeric observables
9
Architecture
Context adapter
Situation and context identification
Privacy / Security / Trust
Transformation symbolic observables
Sensing numeric observables
10
Architecture
Context adapter
Situation and context identification
Privacy / Security / Trust
History
Transformation symbolic observables
Sensing numeric observables
11
Architecture
Context adapter
Situation and context identification
Privacy / Security / Trust
Discovery / Recovery
History
Transformation symbolic observables
Sensing numeric observables
12
Architecture

• For each layer
• Services are self-descriptive (meta-interface
  meta-data)
• Services are self-configurable
• Services are local to global (islands of
  interaction and existence of servers)
• Services support multiple communication
  mechanisms (publish-subscribe, request-answer)
• A uniform computational model (process/data
  oriented) and architectural style

13
Architecture
One style does not fit all A set of interoperable
infrastructures
14
The contextor infrastructure
Context adapter
Situation and context identification
Privacy / Security / Trust
Discovery / Recovery
History
Transformation symbolic observables
Sensing numeric observables
15
Outline

• Contextor
• Discovery
• Examples

16
Contextor Software abstraction

• Functional core typed communication channels

Control In
Data In

Data Out
Meta Data Out
Meta Data In
Control Out
17
Contextor Software abstraction

• Functional core
• Transformation Data (Type X) Meta Data è Data
  (Type Y) Meta Data
• Meta Data uncertainty, accuracy, latency

Control In
Data In

Data Out
Transformation
Meta Data Out
Meta Data In
Control Out
18
Contextor Software abstraction

• Functional core
• Transformation Data (Type X) Meta Data è Data
  (Type Y) Meta Data
• Control behavior adaptation
• supports reconfigurability

Control In
Data In

Data Out
Transformation
Meta Data Out
Meta Data In
Control Out
19
Dynamic Composition of Contextors
Application 1
Application 2
Data and Meta Data Flow
20
Dynamic Composition of Contextors
Application 1
Application 2
Data and Meta Data Flow
Controls Flow
21
Life Cycle of a Contextor
Creation, conception of contextor
22
Life Cycle of a Contextor
Creation, conception of contextor
Launching the contextor from its configuration
file
23
Life Cycle of a Contextor
Creation, conception of contextor
Looking for Source Contextors Send request for
service on the network
Launching the contextor from its configuration
file
24
Life Cycle of a Contextor
Creation, conception of contextor
Looking for Source Contextors Send request for
service on the network
Launching the contextor from its configuration
file
Subscribing to Source Contextors
25
Life Cycle of a Contextor
Creation, conception of contextor
Looking for Source Contextors Send request for
service on the network
Launching the contextor from its configuration
file
Subscribing to Source Contextors
Lauching the functional Core Waiting for Sink
clients
26
Life Cycle of a Contextor
Creation, conception of contextor
Looking for Source Contextors Send request for
service on the network
Launching the contextor from its configuration
file
New sink Client connection
Subscribing to Source Contextors
First sink Client Connection
Last sink Client Disconnection
Lauching the functional Core Waiting for Sink
clients
27
Life Cycle of a Contextor
Creation, conception of contextor
Looking for Source Contextors Send request for
service on the network
Launching the contextor from its configuration
file
New sink Client connection
Stop Command
Subscribing to Source Contextors
First sink Client Connection
Last sink Client Disconnection
Lauching the functional Core Waiting for Sink
clients
28
Outline

• Contextor
• Discovery Scalability
• Examples

29
P2P Network Model

• Both a client and a server
• Improves the autonomy of contextors

30
P2P Network Model

• Both a client and a server
• Improves the autonomy of contextors
• Fully Distributed Architecture
• Supports the arrival/departure of contextors

31
P2P Network Model

• Both a client and a server
• Improves the autonomy of contextors
• Fully Distributed Architecture
• Supports the arrival/departure of contextors
• Local Discovery
• Requests use UDP multicast

32
P2P Network Model

• Both a client and a server
• Improves the autonomy of contextors
• Fully Distributed Architecture
• Supports the arrival/departure of contextors
• Local Discovery
• Requests use UDP multicast
• Distant Discovery
• Hybrid Model (GloSS)
• Use of "repeaters (model based on geographical
  localization)

33
Outline

• Contextor
• Discovery
• Examples

34
Example 1 Human activities Observatory

• A dynamic Web page displays users level of
  activity

35
Example 1 Human activities Observatory

• Architecture

Web Pages
Federation of Contextors
36
Example 1 Human activities Observatory

• Architecture

Observed User
Keyboard Activity Contextor
Displays Users Activities
Servlet
Local Activity Contextor
Mouse Activity Contextor
Adapter
Local information Contextor
37
Example 1 Human activities Observatory

• public class ActivityObservatory extends
  HttpServlet
• private ActivityAdapter myAdapter
• public void init()
• myAdapter new ActivityAdapter()
• public void doGet(HttpServletRequest request,
  HttpServletResponse response) throws IOException,
  ServletException
• for (int i0 ilt myAdapter.getDataInConnections(
  "activity") i) this.displayInformation(
  myAdapter.getData("activity",i) )

38
Example 1 Observatory of Activity

• public class ActivityMouseContextor extends
  ElementaryContextor
• private Process mp null
• private String data ""
• public void init()
• mp Runtime.getRuntime().exec("MouseSensor.exe"
  )
• public void functionalCore()
• data this.mp.read()
• this.setDataOutValue(data.getBytes())

39
Example 2 I-AM

• C. Lachenal et N. Barralon Works

• A software infrastructure that supports the
  dynamic composition of heterogeneous interaction
  resources to form a unified space.
• In this space, users can distribute and migrate
  whole or parts of user interfaces as if they were
  handled by a unique computer.

40
Example 2 I-AM

• C. Lachenal et N. Barralon Works
• 2 types of contextors
• Surfaces Contextors
• Indicate the presence of each surface ( display)
  
• Indicate characteristics (size, ) of each surface

41
Example 2 I-AM

• C. Lachenal et N. Barralon Works
• 2 types of contextors
• Surfaces Contextors
• Link Contextors
• Encapsulate a physical sensor
• Indicate links (spatial relations) between each
  surfaces

PIC Microprocessor
IRDA transmitter / receiver
IRDA manager
USB connector
42
Example 2 I-AM

• C. Lachenal et N. Barralon Works
• 2 types of contextors
• Surfaces Contextors
• Link Contextors
• Encapsulate a physical sensor (currently
  simulated)
• Indicate links (spatial relations) between each
  surfaces

43
Conclusion

• Role of Contextors
• Capture and computes contextual information
• Distribution (local and global)

44
Conclusion

• Role of Contextors
• Capture and computes contextual information
• Distribution (local and global)
• Middleware
• Distributed
• Dynamic
• Autonomous

45
Perspectives

• Evaluation of architecture
• Latency
• Scalability
• Reconfigurability

46
Perspectives

• Evaluation of architecture
• Latency
• Scalability
• Reconfigurability
• Implementation of examples
• Mobility (connections / disconnections)
• Distribution (repeaters)

47
Perspectives

• Evaluation of architecture
• Latency
• Scalability
• Reconfigurability
• Implementation of examples
• Mobility (connections / disconnections)
• Distribution (repeaters)
• Contextors Upgrade
• History of the data

48
The Contextor Infrastructure for Context-Aware
Computing

• Gaëtan Rey, Joëlle Coutaz
• Engineering HCI research group
• CLIPS IMAG, Grenoble, France

49
Data et Meta DataXML and XML Schemas

• Benefits
• Portability
• Extensibility
• Uses
• Initialization using configuration files in XML
• (name, group, Data In and Data Out class )
• Communication based on XML messages