community manager a dynamic collaboration solution on heterogeneous environment 2006.06.26 hyeonsook...

Community ManagerA Dynamic Collaboration Solution on Heterogeneous Environment

2006.06.26

Hyeonsook Kim [email protected]

2006 CUS. All rights reserved.

21C R&DProject

ICPS ‘06

2

Agenda

Introduction Our Approach Experiments Conclusion & Future Work

3

Agenda

Introduction - Motivation - Community Computing - Related Works

Our Approach Experiments Conclusion & Future Work References

4

Introduction/ Motivation

* Problems 1. Heterogeneity 2. Mobility3. Adaptation

How can we reduce the complexity of collaboration service development in the pervasive computing environment?

Pervasive Computing EnvironmentIn distributed environment, services and computing devices have to collaborate autonomously and continually to achieve a goal

High level heterogeneity of device, service, communication protocol, network and etc.

High level dynamism and unpredictability Intelligent environment, Computationally enhanced environment

Services adapting to user intention and Environment change

5

Introduction/ Community Computing

Community Computing

Abstract Layer

Implemental Layer

Runtime Adaptation* New Computing Paradigm 1. Abstracted Collaboration

Model 2. Model Based Service

Description3. Automatic transformation into

real world

Image Display

ScannerAuthentication

For the consistent collaboration service among the diverse middlewares, services and devices, the service model need to be separated and abstracted from the runtime environment.

6

Introduction/ Community Computing

Community ComputingAn abstract and autonomic collaboration model for pervasive fusion service in pervasive computing environment.

3. Create community

2. Set up goal

5. Learn the results

1. Sense situation

4. Collaborate inter/intra community

Community Metaphor

7

Introduction/ Related Works

MDA (Model Driven Architecture) supports a methodology for the automatic transformation of

platform independent service model into platform dependent model

has limits for collaboration modeling cannot support dynamic collaboration service which has to

adapt itself to the runtime environment dynamically

SWS (Semantic Web Services) add semantic knowledge to existing web services discover and composite services dynamically support just one type service description language cannot support integration of various domains and environments

8

Agenda

Introduction Our Approach - Community Model & Description - Community Life Cycle & Member Binding - Community Manager

Experiments Conclusion & Future Work References

9

Our Approach/ Community Model & Description

Community Model

Community Skin

Sec

urity

User Perference

QoS

Situation-Aware

Autonom

ic

member

member

member

member

member

Encapsulation using community skin provides information hiding and reuse per a community.

Policy Flesh

Member

Policies for organization and management of a community are plugged in and out dynamically.

Members [services/devices/ men] collaborate and cooperate in a community.

Goal

Community Goal

Community Action

Actions which community members have to do for community goal

An abstract goal has to be achieved by a community

10

Our Approach/ Community Model & Description

CDL (Community Description Language)

11

Our Approach/ Community Instantiation

Community Instantiation

12

Our Approach/ Community Life Cycle & Member Binding

Creation Phase:

Organization Phase:

Execution Phase:

Organization

Activation

TEXT_VIEW{CLOSEST, 2}

USER_RESPONSE{DEPENDENT_MEMBER, 1}

STREAM_VIEW{DEPENDENT_MEMBER, 1}

TEXT_VIEWLCD_Display

USER_RESPONSEGetResponse

AUTHENTICATIONgetAuthenKey

TEXT_VIEWsetCaption

STREAM_VIEWMoviePlayer

TEXT_VIEWLCD_DisplayLCD_Display on Cup

USER_RESPONSEGetResponseGetResponse on PDP

AUTHETICATIONgetAuthenKeygetAuthenKey on Authenticate Center

TEXT_VIEWsetCaptionsetCaption on PDP

STREAM_VIEWMoviePlayerMoviePlayer on PDP

Community Member Binding on Community Life Cycle

Creation Phase When describing the community, the members

are handled with service types which imply

meta-services Organization Phase The members are handled with service names

of the services which satisfy service types and

constraints in the execution environment. Execution Phase After a community is activated, the members

are handled with service ids of the service instances which run on the real device

Organization

Execution Dormancy

reorganize

activate

deactivateactiv

ate

deactivate

reorg

anize

reorganize

Creation

Exterminationachieve goal

organize

13

Our Approach/ Community Manager

Community Manager

Situation Generator

generates situation message

Community Organizer

interpret community template and manage community life cycle

Community Executor

discovery and select member instances

Situation Rule Repository

Community Repository

Meta Service Repository

Mid

dlew

areC

om

mu

nity M

anag

er

Community Repository

Meta ServiceRepository

LookupTable

Service Ontology Repository

Situation Rule Repository

Community Organizer

CommunityExecutor

Web Services

Applications

Agents

Situation Generator

Service Advertisement

Community: predefined situation based collaboration service model

Protocol Translation

Situation Transmition

Community Interpretation

Member Recommendation

Mem

ber D

isco

very

14

Our Approach/ Community Management System

Community Interpretation

ContextDB RDF FileModel

In memoryRDF File

Dev-Time: Protege Run-Time: Jena

Load/Query

Modify/WriteTranslate

Encoding

FOAF

family

community

Include

prefer

Community Script Community

Object In Memory

Load/Unmarshal

Query/Answer

Interprete

Runnable Community

15

Our Approach/ Community Management System

Dynamic Plug-in Middleware Adaptor

Controller

Container(Spring Framework)

Controller


Tibco/ Rendezvous

SinkPipes

HmeStation

Front-end uT Adaptor

Back-enduT Adaptor

PipesHealthCare

TibrvSource

TibrvSinkSocketSourceSocketSink

SocketSourceSocketSink

Front-end uT Adaptor

Back-enduT Adaptor

ELAB#1 Test Environments

Source

uTAdaptorWrapper

Source

Sink

Source

Sink

Source

Sink

Sink

Source

Sink

Source

TibrvSource

TibrvSink SocketSource SocketSink

SocketSource SocketSink

TCP/ IP

TCP/ IPuTAdaptorWrapper

PipesSource

Sink

Sink

Source

Controller


MOM (Message Oriented Middleware) Based

Runtime adaptor injection

Spring Framework

JAXB Framework

Dynamic reconfiguration of communication channels (Source, Pipes, Sink)

16

Agenda

Introduction Our Approach Experiment - Demo Scenario - A Sample Community Template - Community Viewing

Conclusion & Future Work References

17

Experiment/ Demo Scenario

SituationGenerator

Library

Mother

Mother

Smart Cup

Smart Belt

StreamingServer

Smart Street Lamp

Offline Community

Stream Request/User Authentication

Current Location

Stream Transmition

Stream Translation

User Interaction

Situation Notification/User Response

Situation Notification/User Response

DaughterDaughter

Daughter

Camera Relay

Living room

Community Organizer/Executor

Home Media

Daughter

Scenario Flow

18

Experiment/ A Sample Community Template

Notify L2 Community

19

Experiment/ Community Viewing

Notify L2 Community

Community Organization

Member Discovery:

- Member Type: TEXT_VIEW

20

Agenda

Introduction Our Approach Experiment Conclusion & Future Work - Major Contributions - Future Works

References

21

Conclusion/ Major Contributions

Advantages

Ability of collaboration in heterogeneous dimension• Adaptive• Reconfigurable

Encapsulation (Abstraction)• Balancing of transparency and awareness

Reuse• Reusing the unit of a community with a goal, members and policies

Dynamic Decision Making• Externalized logic

Sharing Context in Community

22

Conclusion/ Future Works

Future works

Community Management• Policy Based Member Discovery• Community Action Evolution by Learning

Member Monitoring • Member Defect Detection• Automatic Healing

Meta Service Ontology • Standardization of Specific Domain Service Interface• Service Mapping on Syntax and Semantic Understanding

Stronger Security/Privacy methodology Inter Community Management

23

Agenda

Introduction Our Approach Experiment Conclusion & Future Work

References - Lists

24

References/ Lists

Puneet Gupta, "Evolving a pervasive IT infrastructure: a technology intefration approach”, Pers Ubiquit Comput 2004, 8: 31

Object Management Group, Model Driven Architecture Guide, 2003 M. Paolucci, K. P. Sycara. Autonomous Semantic Web Services. IEEE Inte

met Computing. 7(5): 3441, 2003. http://www.w3.org/submission/OWL-S R Sivashanmugam, Framework for Semantic Web Process Composition. Inl

emational Journal of Electronic Commerce, 2004 B. Benatallah, Environment for Web Services Composition. IEEE Internet

Computing. 17(1):40-48, 2003 J. Michael Yohe, “Community Comuputing and the Computing Communit

y", Proc. of ACM SIGUCCS Conference on User Services, October 1994

25

Thank you!