SPICE Service Platform for Innovative Communication Environment

1
SPICE Service Platform for Innovative
Communication Environment

• Service Creation Environment for Mobile Services
• Paolo Falcarin
• Politecnico di Torino Italy
• http//softeng.polito.it

MMC Workshop, 2nd 3trd April 2008, Berlin, H.
Hertz Institute
2
Contents

• General intro on SPICE Project
• Focus on Telco Service Creation and Composition
• Approach
• Developer Studio
• Use Case
• Automatic Composition Engine
• Conclusion/Perspectives

3
IST SPICE General Introduction

4
SPICE Consortium
5
Objectives

• Architecture and framework to support
• Easy and Quick service creation
• Compose and Deploy intelligent mobile
  communication and information services
• Speed up the Service Creation Process
• Model-Driven Development
• Tool chain for fast development cycles
• Address heterogeneous execution environments
• Telco / IT
• At Server and Terminal side

MMC Workshop, 2tn 3trd April 2008, Berlin H.
Hertz Institute
6
Service Creation in SPICE- Motivation -

• Take advantage of Service Oriented Architecture
  benefits
• Services are built and made available in
  repositories
• New (composite) services can be composed out of
  existing services
• with the challenge of realizing automatic
  service composition
• Services and service request descriptions are
  necessary in order to discover, match and invoke
  services
• Address Service Creation and composition by
• Professional Service Developers
• End Users

MMC Workshop, 2tn 3trd April 2008, Berlin H.
Hertz Institute
7
SPICE Service Creation Environment

• Developer Studio

MMC Workshop, 2tn 3trd April 2008, Berlin H.
Hertz Institute
8
SPICE approach SPATEL language

• Define a high-level language (SPATEL)
• to specify Telco / IT integrated services
• Using a Meta-Model , based on EMF/MOF
• Service Interface can be annotated semantically
  and to express
• Semantic Data Types Goals
• Non functional properties (QoS, security,
  Cost,).

9
Service Description formalism
Interface of a composite component
Logic using state-machine definition
c
A Semantically annotatedservice
interface (invoked by the composite)
An 'opaque' implementation
10
Service Creation Environment (SCE)

• Service Creation Environment
• tool chain allowing editing of SPATEL definitions
• generating various implementations that can be
  deployed in different targets

Platform Independent Design
Code Generation manual coding
ExecutionFramewoks
Variability in the deployment of the
orchestration logic
11
Transformation Code Generation
State Machine Code Impl
Web Interface for quick simulation
An Interface in Phone
12
SPICE Architecture SCE viewpoint
12
13
The Developer Studio v.1StarUML add-in
14
The Developer Studio v.2Eclipse GMF-based editor
15
Ontology Viewer Semantic Editor
16
End-User Studio

• Rule driven End-User Studio
• Semantic abstract composition of functionalities
• Reuse servicesfrom SCE/Repository inthe
  End-UserStudio

17
Use Case

• A Dinner planning composite Service

18
Use Case Dinner Planning
User requests to receive a restaurantproposal
list at launch time located in the surroundings
Agenda
OrchestrationEngine
Traveller
Interest Points
3rd Party Call
19
Service Interface in SPATEL
20
Orchestration Definition in SPATEL
Email inspection
Orange Locator
Alcatel GIS
NEC Audio Conf
Yellow Pages
Google Calendar
OrangeSMS sender
21
E-Tourism Scenario
Telecommunication Services
9. SIP MessageContact Restaurant
Sip Server
JAIN SLEE (Mobicents JBoss)
SIP RA
1. SIP MessageFind Restaurant at 2000
Location Service WP2 Component
8. SIP MessageNearby restaurants list
IM RA Wrapper WP2 Component
3PCC RA Wrapper WP2 Component
5.Get EU Location
7.SIP Message to EU
2. SIP Message from EU
10. SIP Message from EU
11. Activate SIP Call
Service Composition
IT Services
BPEL Orchestra
J2EE (JBoss 4.0)
Personal Agenda WP2 Component
3. Store Info
4. At 2000, alarm
GIS WP2 Component
6. Get restaurants list nearby location
21
22
SPICE Service Creation Environment

• Automatic Composition Engine

23
Automatic Composition Engine (ACE)

• The purpose of ACE is to automatically compose a
  number of existing services into a new
  (composite) service that satisfies some user
  request
• The existing services are specified in SPATEL in
  terms of semantically annotated inputs, outputs,
  preconditions, effects, goal and non-functional
  properties
• The user request can be specified either in
  natural language or as a formal request, in terms
  of desired inputs, outputs, goals and
  non-functional properties
• When more than one service composition satisfies
  the user request the non-functional properties
  are used to rank these service compositions

24
Composition by Service Request in formal language

• Takes a service request, specified in SPATEL
  (i.e., desired inputs, outputs, goals and
  non-functional properties)
• Discovers all services that match any of the
  specified inputs, outputs and goals
• Creates all possible candidate service
  compositions from the discovered services
• Ranks the candidate service compositions based on
  their aggregated non-functional properties
• Filters out the service compositions that do not
  meet some additional user constraints
• Generates SPATEL for the service compositions
  that meet the user constraints

25
Automatic Composition Engine formal service
request
SCE
Formal Service Request (goal, cost, interactions
relations, QoS)
RequestedService
Determine equivalence
AggregatedService
3
1
2
Find service composition
Composition matches Request?
Determine aggregation of(selection of) service
properties
Service A
Service B

• 1. compose
• 2. matches
• 3. aggregate

Service C
Service composition
26
Composing from request in natural language

• Takes a service request in natural language
• Analyzes the request to find pertinent "concepts"
• Discovers all services matching these concepts
• Automatically composes the discovered services
• Generates a new composite service
• Deploys the new service on the execution engine
• Generates an application that uses the new
  service
• Executes the application and presents the results
  back to the user

27
Automatic Composition Engine natural language
Formal Service Request (goal, cost, interactions
relations, QoS)
RequestedService
Determine equivalence
AggregatedService
3
1
2
Find service composition
Composition matches Request?
Determine aggregation of(selection of) service
properties
Service A

• 1. compose
• 2. matches
• 3. aggregate

Service B
Service C
Service composition
28
Conclusion/Perspectives

• Abstraction at the interface level and at
  composition level is the key of flexibility
• Multiple implementations and deployments for the
  same service orchestration logic
• Multiple possible implementations for each used
  enabler.
• Thanks to MDA and SOA marriage
• Use of OMG standards UML, MOF, QVT and tools
  implementing the standards EMF/SmartQVT
• Sevice Creation Environment is provided as a set
  of Eclipse plugins

29
Questions
30
(No Transcript)
31
Use Case 2

• Design and implementation of a SNS with a Free
  Call function

32
Clubber SNS Scenario (1)
An end-user (Monica) is in the premises of a club
and would like to invite a friend (Mariano) to
join her in the club
32
33
Clubber SNS Scenario (2)
34
Clubber SNS Scenario (3)
Paolo chooses to accept the message receives
information about the club (name, logo, map, etc)
34
35
Clubber SNS Scenario (4)
MMC Workshop, 2tn 3trd April 2008, Berlin H.
Hertz Institute
36
Clubber SNS Scenario (5)
Paolo can then decide to join with cheaper access
the club
MMC Workshop, 2tn 3trd April 2008, Berlin H.
Hertz Institute
37
Functional view of the composition
9
Accept Call
Call Activation
Invite Mariano
8
2
Send docwith holes
3
check
Monica
4
Paolo
get
7
5
6
Decode push
get
get
get
Access Control
Photo of Monica (encoded)
Route tothe Club (encoded)
RenderManager
Call Reference (encoded)
Club Icon
Operator
MMC Workshop, 2tn 3trd April 2008, Berlin H.
Hertz Institute
38
SEE Elements

• Engines
• SPATEL Engine
• BPEL Orchestra Engine
• Enablers
• Referred call manager (on top of Asterisk)
• Localisation (Valencia simulation with SIAFU)
• Authentication
• Application components
• Mediation components (pseudonym resolution,
  encoder)
• Render manager

MMC Workshop, 2tn 3trd April 2008, Berlin H.
Hertz Institute
39
Modelling of the Orchestration
MMC Workshop, 2tn 3trd April 2008, Berlin H.
Hertz Institute
40
Results
Dynamic Desktop Client to sendthe invitation
Invitation received (in HTML format)
MMC Workshop, 2tn 3trd April 2008, Berlin H.
Hertz Institute