Upload
eric-kluijfhout
View
388
Download
0
Tags:
Embed Size (px)
DESCRIPTION
Project definition; Project set-up; Activities & outcomes; Project conclusions & HE state of the art; SURF activities 2007
Citation preview
OpenLW reference architecture project
author: Eric Kluijfhout, [email protected]
This work is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License.
To view a copy of this license, visit http://creativecommons.org/licenses or send a letter to Creative Commons, 559 Nathan Abbott Way, Stanford, California
94305, USA.
OpenLW reference architecture project
2004-2006
Eric Kluijfhout
Outline
Project definition
Project set-up
Activities & outcomes
Project conclusions & HE state of the art
SURF activities 2007
Problem 2004
VLE of the future?
Trends en new mandate HE
Technology
trends
Institutional VLE
experiences
requirements?
relevance?
new needs?
Project definition: March-June 2005
Technology-trends analysis - architectures:
• national
• international
• commercial
• non-commercial
Institutional experiences and needs – stakeholder
consultation: • Digital University consortium
• SURF
• E-merge consortium
• Apollo consortium
Trends and HE mandate – desk study + interviews
Outcomes project definition: mid 2005
VLE of the future?
Trends en new mandate HE
Technology
trends
Institutional VLE
experiences
Desk study
Architecture
analysis
Stakeholder
consultation
Stakeholder-consultation
Wide range in levels of VLE use and ambitions within and between HE institutions
Hardly any relation to educational model - VLE mainly used as a logistical system
Discrepancies between central VLE policies and local implementation
Demand for tools to communicate, plan and manage VLE use
Demand for sharing information on new trends VLE architecture is part of the overall application
architecture
Architecture analysis
Also the commercial VLE providers move
towards component-based systems and SOA –
interoperability but no interchange
SAKAI, ELF and IMS-AF internationally best
known initiatives – but important differences
Within the Netherlands a number of initiatives
towards reference architecture definition
soa as a first step towards SOA?
Desk study HE trends
The knowledge society Life long learning Globalisation Regionalisation From supply- towards demand driven The new students Towards an open HE market HE as an initiator of innovation Continued cost reductions Increasing complex society
Project credo
Instead of asking
“How to select the best VLE package for
our institution?”
We should be looking for ways
“How to manage the continuous changes
in needs and opportunities?”
Open Learning and Working Environment
Aim:
Together with HE stakeholders develop
the required knowledge, critical
mass, and required instruments to
devise an agenda for the
implementation of a service-based
architecture for OpenLW
Starting from:
Functional approach towards ‘services’ Build on what is already there within and
outside the Netherlands Limit ourselves to the educational
domain Assumptions:
• 80% of the processes within HE institutions are functionally identical
• Differences are due to implementation decisions
Project activities
1. Define functional OpenLW reference architecture to identify common services
2. Validate with HE representatives
3. Gap-analysis re. existing frameworks
4. Pilot: service definition up to technical design
5. Dissemination activities
Activity 1: functional reference architecture
Trendanalysis
Class-extensions Sub-processen
Main processenDomain model
functional service-architecture (soa)
Service definitions
HE value chain in 2015
Matching student demand and institutional
supply
Manage OpenLW
Develop educ. offerings
Instantiate educ. offerings
Exploit educ. offerings
Maintain educ. offerings
Learning act.
Learning envir.
Learning act.
Learning envir.
Learning act.
Learning envir.
Learning act.
Learning envir.
Tree general process trends to be supported through OpenLW:
1. Five main processes get more and more interwoven
2. Educational and organizational/logistical processes get interwoven
3. Role-diversification of staff and student
Domain model
OnderwijsProv ider
Persoon Kwalificatie
+ niveau:
KwalificatieStructuur
+ domein:
Begeleiding
Assessment
ContentProv ider
Prov ider
TeVerwerv enKwalificatie
+ uitleverModus: + startDatum: + doorloopTijd: + studieLast: + kosten:
Verworv enKwalificatie
+ bewijsmateriaal: l i jst
AssessmentProv ider
BegeleidingsProv ider
Dienst
PersoonProv iderOv ereenkomst
+ diensten: l i jst+ producten: l i jst+ kwalificaties: l i jst+ start: datum+ einde: datum+ uitleverModus:
BronMateriaal
Product
+ metadata: l i jst
OnderwijsEenheid{abstract}
+ programmeringsRegel: regel+ ingangsEisen: l i jst+ uitleverOpties: l i jst
Aanbod{abstract}
+ type: (ontw./expl./beheer)
Faciliteit
Prov iderProv iderOv ereenkomst
+ diensten: l i jst+ producten: l i jst+ kwalificaties: l i jst+ start: datum+ einde: datum+ uitleverModus:
Enkelv oudigeOE
SamengesteldeOE
LeerActiv iteit LeerOmgev ing
*
*
streeftna
*
*
*
*
*
*
*
levert
*
*
*
1..*
*
* bezit *
*
reguleertsamenhang
*
1..*
*
1..*
*
1..*
*
OpenLW soa definition (v 0.1)
Based on the 5 main processes:
> 80 sub-processes (UML Use Cases)
(UML) Domain model + vocabulary (±100 classes)
> 70 service-definitions (UML Sequence Diagrams)
> 40 service provider definitions
5 architecture outlines (each main process)
uitwerking
soa architectuur
uitwerking
Manual service-implementation
klant service interface service provider
formele vraag
formeel antwoord
vraag
antwoord
Computer-supported service-implementation
klant service interface
vraag formele vraag
formeel antwoordantwoord
informatiesysteem
Web-based service-implementation
klant
formele vraag
formeel antwoordwerkstation
formele vraag
formeel antwoord
informatiesysteem
web-gebaseerdinformatiesysteem
formele vraagformeel antwoord
informatiesysteem
regelaar deelvraagdeelvraagdeelantwoorddeelantwoord
Some missing computer-supported services
‘Positioning’ and validating (previously) acquired competences
Study route/planning guidance Matching student demand and institutional supply –
sector-wide Student progress tracking and mapping in open
learning settings Peer assessment and tutoring support Group and logistics management by students Personal desktop tools integration with OpenLW ………………………..
Activity 2: Validation
By small expert group
Outcomes:• Common sub-processes can be defined ‘two levels deep’
• ‘Services’ and ‘functions’ at this level are almost identical
• Difficult to take a functional – and not implementational –
view
• Requires a major effort to draw up and communicate
• Selecting the ‘critical’ common services remains a subjective
and political process
Activity 3: Gap analysis OpenLW
Open source reference models: IMS Abstract Framework E-(learning) framework implementations SAKAI
Conclusions: Process approach OpenLW versus ‘aspect’
approach other frameworks OpenLW requires more detailed elaboration
Activity 4: in-depth soa definition
Domain: Assessment• Sub-domain: peer-assessment, self-assessment, and
360o feedback• Sub-processes:
– Design assessment– Item construction– Assessment construction– Conduct assessment – Process responses– Decide
Service definition ‘Response processing’
Approach• Describe overall assessment process
• Define six sub-processes through formal Use Cases
• Compile an assessment domain model
• Service definition for sub-process ‘process response’
Validation: • Model ‘process response’ in Enterprise Architect tool
• Generate Java-classes and expert to a Java development
environment
• ‘Round trip’ generate UML models based on adapted Java
code
Required efforts
Completed:• Domain model development for assessment (outside this
project): 200 days (1 fte)
• Functional design response processing webservice (1 out of
6 sub-processes): 20 days
Still to be done:• Improve functional design (10 days); technical design (30);
implement (30 days); unforeseen (10 days): total 100 days
• For complete assessment domain: 6 x 100 days = 3 fte
• Orchestration! ???? days
Activity 5: Dissemination
Activities:• Organise two national events on VLE/SIS integration
• Smaller discussion sessions with experts
• Website and wiki
Conclusions:• Project seemed to addressed relevant issues
• But: relevance and outcomes difficult to communicate
Overall conclusions
Top-down approach to defining a reference
architecture based on functional service
definitions is possible
The applied methodology and notation (UML)
increases transparency, but not validity.
Functional sub-processes can be described in
‘free format’ WSDL
Conclusions continued
But: statelessness is problematic/ often
orchestration will be required
Detailing to the level of technical web service
designs requires a considerable effort
Crucial question remains: which are the most
relevant common services?
Within Dutch HE expertise in this area is limited
State of the art in Dutch HE
Limited number of HE institutions is implementing soa/SOA
Approach, scope and ambition vary considerably
Knowledge dissemination is slowly starting up
Many small tools available, but relatively unknown, and no
common domain model
‘Orchestration’ issue seems the next problem to solve
Coming years: ‘mixed approach’
Possibly largely supplier-driven
Other sectors may provide convincing best practice
examples
Audiences and messages
ICT departments: enterprise architecture integration (EAI)
Decision makers: planning of your institutional application
architecture + shared services
Individual users (students & staff in research, education,
admin.): web goodies / PLE
Government & funding bodies: share best practices,
standards
Economic sectors & industry: life long learning agenda
HE ICT suppliers: market needs and potential
SURF-Werkplan 2007/2010 ‘Krachtig Doen’
Realise an information-infrastructure for:• End users:
• More transparant ICT facilities
• From supply to demand driven functionality
• Individual inst./ICT dep:
• Standardise the heterogeneous information-infrastructure
• Define a future oriented information-infrastructure
• HE consortia and HE sector:
• Define shared information-architecture
• ‘Shared services’ model
Activities 2007
Knowledge dissemination and training• Suppliers conference
• ‘Best practices’ conference
• Development strategies and tools
• Website, wiki, groupwork environment?
Community-building• HE soa strategy group
• Work groups
Needs assessment and business case development• For shared services
• What is available/can be adjusted
• Elaborate on OpenLW reference model??
• Explore Archimate modelling approach
……… continued
Proof of concept pilots• ‘Shared services’ pilot implementations (workflow)
• PLE compilations (indiv. webservices)
‘SOA Route Planner’ together with HE
institutions, partners and suppliers
SOA methods & tools selection/development
Technology scouting en standards
Within SURF: create synergy between platforms