Chaining environmental Web services: composing a Web service chain

armyfertileSecurity

Nov 3, 2013 (3 years and 7 months ago)

72 views

1
Chaining environmental Web services:
composing a Web service chain
Huub Scholten, Ayalew Kassahun,
Adrie Beulens (Wageningen University, Netherlands)
Arno Krause (Alterra, Netherlands)
Chaining environmental Web services:
composing a Web service chain
Huub Scholten, Ayalew Kassahun,
Adrie Beulens (Wageningen University, Netherlands)
Arno Krause (Alterra, Netherlands)
Concluding
remarks
CHESS-system
(Human)
roles
Standards
Paradigms
Initiatives
Details of
CHESS
CHESS-
definer
Problem to be
solved
3
Problems to be solved

Decision making for environmental problems is complex

Modelling requires various artefacts:

Model
(model modules),
data
,
problem
and
OS
information,
tools
, etc.

Artefacts are often
distributed

Artefacts are often
proprietary

Many owners of artefacts prefer to make them available as
Web services

Model-based problem solvers have to
find
the artefacts and
bind
them to a
single end-user application that can be
executed
to solve the problem

Solving this problem by
automatic
Web service chaining
4
CHESS –Ch
aining E
nvironmental Web s
ervices

CHESS-system consists of

CHESS-framework (that generates environmental applications)

(environmental) Web services
•Models
•Data
•Other tools

Environmental applications for decision and policy support
Overview of system to be built
Feed-back
Weather Forecast
CentreWeather forecast
Hydrological
Forecast Centre
Streamflowforecast
and pre-warning
Models and tools A
Local
river
basin
forecast
centre
Local
river
basin
forecast
centre
Flood alert -inundation mapping
Models
and tools B1
Models
and tools B2
Public
Stakeholder
Manager
Decision-maker
Local data

Flood alert -inundation mapping

Local
river
basin
forecast
centre
Expert knowledge
Uncertainties
Feedback
Expert knowledge
Uncertainties
Expe
rt knowledge
Un
ce
rtainties
Weather Forecast
CentreWeather forecast
Hydrological
Forecast Centre
Streamflowforecast
and pre-warning
Models and tools A
Local
river
basin
forecast
centre
Local
river
basin
forecast
centre
Flood alert -inundation mapping
Flood alert -inundation mapping
Models
and tools B1
Models
and tools B2
Public
Stakeholder
Manager
Decision-maker
Local data

Flood alert -inundation mapping
Flood alert -inundation mapping

Local
river
basin
forecast
centre
Expert knowledge
Uncertainties
Expert knowledge
Uncertainties
Feedback
Expert knowledge
Uncertainties
Expert knowledge
Uncertainties
Expe
rt knowledge
Un
ce
rtainties
Expe
rt knowledge
Un
ce
rtainties
Example:
environmental
simulation
modelling for
flood prediction
in France
7
(Human) roles
Feed-back
interacts with the
definer
-part of the
analyser
to
define what functionality, data, etc. is needed to
solve the environmental problem at hand
are end-users; they use the resulting
environmental
application
to get advice on the environmental
problem at hand
Web service builders:
convert existing,
relevant data, models,
tools, etc. into Web
services that are
CHESS-compliant
8
Some international initiatives to be used

SISE
: Single Information Space in Europe for the Environment

Workshops, ideas, requirements

Not implemented yet

SEIS
(INSPIRE): Shared Environmental Information System

EC sponsored initiative
to modernise and simplify the collection, exchange
and use of the data and information required for the design and
implementation of environmental policy

Technically based on INSPIRE (Infrastructure for Spatial Information in
Europe)

Mainly focusing on (real time, environmental) data
9
Existing standards / paradigms

SOA paradigm
(Service-Oriented Architecture): find, bind, execute Web services

UDDI
(Universal Description Discovery and Integration): a platform-independent
registry for businesses worldwide to list themselves on the Internet

WSDL
(Web Services Description Language): an XML-based language that
provides a model for describing Web services

OpenMI
(developed in HarmonIT-project, FP5): a standard for model linkage in
the water domain

Seamless
(based on OpenMI): integration model-modules for agricultural
simulation models

OpenMI and Seamless are not designed for distributed systems
10
Knowledge for CHESS’meta repository (1)

UDDI
helps to find Web services by registration, which consists of:

White Pages
: address, contact, and known identifiers;

Yellow Pages
: industrial categorizations based on standard taxonomies;

Green Pages
: technical information about services exposed by the business

WSDL
is a language for Web services to

Actual binding of Web services

Specifying transport (SOAP protocol)

Specifying operations to perform on Web services

Etc.
11
Knowledge for CHESS’meta repository (2)

OpenMI
repository

Standard for software (model / data)
interface specification
;

Interface acts as a
contract
between software components;

Implementing the OpenMI interface makes a component
OpenMI
compliant
.

Seamless
knowledge base:

Ontology
•Domain specific knowledge
•Semantic meta-information of software components (incl. model-modules)
•Schemata for relational data base

Semantic content

Relational database

Limited set of agricultural models included
12
Knowledge for CHESS’meta repository (3)

Furthermore we can build on the ontology framework of Scholten (2008)
model/artifact
model representation power
model representation demand
model structure
meta-data
process knowledge
other requirements eisen
problem
problem description
probleem owner
goals
OS definition
domain knowledge
existing models
requirments to models
processes
project organisation
data management
model development
model analysis
model use
etc.
meta-knowledge
(basic terminology)
terminology
terminology
terminology
uses
uses
uses
13
Some detail of the design

Definer

The
composer
has to define and specify: problem-solving process, OS, which
artefacts (model + tools), which data, etc.

Collects all requirements (process, web services, problem)

Semantic mapper
(Ayalew Kassahun)

Meta-repository

UDDI, WSDL-terminology, available services, domain knowledge

To be used by
definer
,
semantic mapper
and
chain designer

Chain designer

Workflow Engine workflow for solving specific problem with the chain

Web service Engine query services and chaining (acc. to workflow)

Environmental application
chain + user interface to solve problem at hand
14
Definer: (human) composer CHESS-IF

Content of interactions:

Questions related to
problem and object system
•On
problem
: problem subject, wanted solution(s),problemcomplexity, problem description,
problem context, problem solving methodology, problem domain(s), etc.
•On
object system
(OS): OS structure (entities, relations/processes between entities), time
frame, location, temporal resolution, spatial resolution, etc.

On
artefact(s):
models, data, tools, dynamic, static, proprietary aspects,
access issues, etc.

On
problem solving process
: what to do (by the chain of environmental
Web services), order of things to do, quality of service, etc.

Result
is: set of requirements (views and perceptions of composer on problem
and solution to be used by semantic mapper
15
Concluding remarks

Strong need by providers of environmental models / data for Web services

Too difficult and cumbersome to develop an application consisting of a chain of
(environmental) Web services for non-ICT skilled researchers

CHESS is
innovative
because of:

Automatic chaining (not existing as far as we know)

Allows non-ICT skilled domain experts to find appropriate, distributed
environmental Web services and chain these to a single application

The CHESS-approach to handle knowledge with an ontology for semantic
interoperability speeds-up developing standards

Need for funding