The myGrid Information Model for myGrid v2
Many of the detailed documents on the deployment of the
InformationModel are available as attachements to the Wiki page
MyGridInformationRepository.
Information Model initial notes Jul 2003
These notes arise from a preliminary discussion between
MarkGreenwood,
ChrisWroe and
NickSharman on 17 Jul 2003. They and
PeterLi will 'own' and coordinate work on the information model for the next phase of myGrid.
At
IntegrationFest5, we decided to expand the group to:
The current vesion is at
infomodel-1.3.doc (MS Word format). You can track the evolution of the information model (including a Together UML version) in our
CVS repository.
The Purpose of an Information Model
An information model aims to describe the entities of interest to a particular information system and/or in a particular application domain. Information models are also known as analysis or domain models. The information model should be rich enough that we can express all our requirements of the application in terms of its entities and their attributes and relationships.
It should be the basis of design for the application, and many of its entities/attributes/relations will have analogues in interface specifications, implementation class diagrams, database schemas and code.
In our case, myGrid is the information system and its application domain is e-Science (and particularly e-biology/bioinformatics).
Inputs to the Information Model
The overall requirement for myGrid is to support e-Science; from the original proposal, we take this to include:
- Supporting the scientific process
- Suppporting the scientist's use of the community's information
- Supporting scientific collaboration
We believe that this will require support for a range of
metadata to describe documents, services, workflows and other user-defined entities, and especially to support questions of the
provenance of these entities.
Knowledge of the application domain
We need to be aware of and can make use of the work of other groups in this area. Some relevant projects include:
- The Scholarly Ontologies Project has developed explicit models of scientific claim and refutation that we can exploit in modelling experimental structure and building provenance records
- To support collaboration, we need to model organizations, teams and individual users. The AKT Project has developed ontologies (especially the 'portal' ontology) that address this area
Our work to date
We will build on earlier work on the information model (see
MyGridInformationRepository and
InformationDescription). An initial review has identified the following areas of the information model as needing attention:
Since we think that the services developed so far play help us support e-Science, we can analyse them (and particularly their interfaces) to discover elements that should appear in the information model.
The requirements
To make the requirements concrete and their satisfaction visible, we aim to support the work of the Graves' Disease project at Newcastle.
PeterLi and
AnilWipat have been working with the GD team to develop the
GravesDisease scenario.
Robert Stevens has identified a set of
questions that a scientist might reasonably ask of a myGrid system;
AlansLabBookStoryBoard is also a valuable source of requirements.
--
NickSharman - 18 Jul 2003, 22 Sep 2003, 26 Feb 2004