CCLRC - EuroCRIS

obtainablerabbiΔιαχείριση Δεδομένων

31 Ιαν 2013 (πριν από 4 χρόνια και 2 μήνες)

150 εμφανίσεις

1

euroCRIS

Members Meeting

Tartu

Eddy Grąbczewski


May 2005

Brian Matthews

November 30, 2001

KM Workshop, SLE

2

Eddy Grąbczewski


Worked in IT development/support for 15 years


EU
3D MURALE

project


Brunel University


CCLRC


Data Management Team


CCLRC


Corporate Data Model designer


EU
CISTRANA

project (EU nationals)


EU

IST World

project (SME)


euroCRIS
CERIF
Task Group Leader


Technical view of the CCLRC repository


Brian Matthews

November 30, 2001

KM Workshop, SLE

3

Council for the

Central Laboratory of the

Research Councils (CCLRC)


Multi
-
disciplinary research organisation
supporting scientists and engineers worldwide


Operates world
-
class large
-
scale research
facilities


ISIS (neutron source)


Lasers (sapphire and glass lasers)


Main shareholder in Diamond Light Source (X
-
ray, UV
source)


Provides strategic advice to the UK government


Manages and participates in international
research projects

Brian Matthews

November 30, 2001

KM Workshop, SLE

4

Where is CCLRC?


Rutherford Appleton Laboratory, Oxfordshire









Daresbury Laboratory, Cheshire


Chilbolton Observatory, Hampshire

Brian Matthews

November 30, 2001

KM Workshop, SLE

5

CCLRC CDR Project


Corporate Data Repository (CDR)
project:


To address the poor database
integration affecting CCLRC business.


Create a infrastructure for future
corporate systems: SSO, CRIS, Open
Archive



Brian Matthews

November 30, 2001

KM Workshop, SLE

6

CCLRC Database Systems



Brian Matthews

November 30, 2001

KM Workshop, SLE

7


Typical questions


Why can’t I use my RAL security pass at DL?
(security access)


Why is my Windows login not transparent across
sites? (SSO)


Why are room booking services so disjointed
across sites? (integration)


How many papers has a department produced
this year?


Which EU projects to a member of staff work on?




Brian Matthews

November 30, 2001

KM Workshop, SLE

8

Causes


Heterogeneous db’s storing
similar

data on different sites:


CCLRC db’s are structurally different.


CCLRC db’s need to share data.


Heterogeneous db’s makes
developing
corporate

applications
more complex and expensive.



Brian Matthews

November 30, 2001

KM Workshop, SLE

9

Possible Approaches


A single corporate db for all applications?


Complex, difficult to agree and hence develop.



A corporate global db and local db’s for
all applications?


probably the best compromise.


We need corporate applications that work
transparently across sites as well as local
applications.


Web based applications using open interfaces.

Brian Matthews

November 30, 2001

KM Workshop, SLE

10

Workflow = Business Processes

Business
Repository
Database
Management
Systems
Databases
Rule
Management
Systems
Rules
S1
S3
S2
S4
S5
Process
Management
Systems
Processes
Process Application
Rule Application
Database Application
Brian Matthews

November 30, 2001

KM Workshop, SLE

11

Corporate Data Repository (CDR)


We use the term “repository” to mean a
common

database to complement local
databases.


The CDR contains the
Corporate Data
Model

(CDM).


Implemented in Oracle but can be ported
to PostgreSQL, MS SQL Server or
Informix.


Brian Matthews

November 30, 2001

KM Workshop, SLE

12

Corporate Data Model (CDM)


The CDM models:


People

in and outside CCLRC.


Jobs, roles and permissions of
People.


Organisation Units

in and outside CCLRC.


Projects

related to CCLRC.


Publications
in and outside CCLRC.


Consolidate data from the existing site
databases at RAL, DL and Chilbolton.


Based on euroCRIS
CERIF
standard.


Brian Matthews

November 30, 2001

KM Workshop, SLE

13

CDR Features


Ontology based (objects, entities)


ANSI/SPARC architecture


CERIF view


Temporal database (VT, TT)


Distributed architecture


Naming subschema


Support for hierarchies


Brian Matthews

November 30, 2001

KM Workshop, SLE

14

Ontology


Object at the top level


Entity is a subtype of Object


All relationships are types of entity


Brian Matthews

November 30, 2001

KM Workshop, SLE

15

ANSI/SPARC Architecture



ANSI/X3/SPARC Architecture (1975)

Brian Matthews

November 30, 2001

KM Workshop, SLE

16

CERIF (euroCRIS)



Common European Research Information Format

Brian Matthews

November 30, 2001

KM Workshop, SLE

17

CDM (CCLRC)



Brian Matthews

November 30, 2001

KM Workshop, SLE

18

CDM Subtypes



Brian Matthews

November 30, 2001

KM Workshop, SLE

19

Organization Temporal Data

Most databases only store current
facts.

What if we want to store historical
facts?


For example:


What was the organisation structure
three years ago?


What was an employee’s name before she
married?


Brian Matthews

November 30, 2001

KM Workshop, SLE

20

CDM Temporal Database


Stores
Valid Time


Stores
Transaction Time


Stores
Last Update

Time

Brian Matthews

November 30, 2001

KM Workshop, SLE

21

CDM Distributed Architecture



CF
DtbId = 3
DL
DtbId = 2
CCLRC
RAL
DtbId =
1
Brian Matthews

November 30, 2001

KM Workshop, SLE

22

Supports Many Data Models

Brian Matthews

November 30, 2001

KM Workshop, SLE

23

Naming Subschema


Based on Pierce’s ontology:


Sign


Symbol

Reference

Icon


Any

entity may have multiple signs,
symbols, references or icons.

Brian Matthews

November 30, 2001

KM Workshop, SLE

24

Support for hierarchies


Date (1986) proposed the EXPLODE
operator for SQL to view the ‘Parts
Explosion’ problem


CDM implements EXPLODE to traverse
hierarchies and simple graphs eg.
OrgUnit_OrgUnit.


CDM also implements
Transitive Closure

operators

Brian Matthews

November 30, 2001

KM Workshop, SLE

25

Summary


Multiple
-
purpose database infrastructure


Ontology based (objects, entities)


ANSI/SPARC architecture


CERIF view


Temporal database (VT, TT)


Distributed architecture


Multiple Names for
any

object or entity


Support for hierarchies


Architecture extended for CISTRANA and IST
World (programme, language & keyword
support)