CrystalGrid 2004

1
CrystalGrid 2004

• Aspects of Current CDS Service
• Interactions with e-Science

2
CDS Overview

• Grant funded by EPSRC
• Based at Daresbury Lab (CCLRC)
• Present Service started 1993
• 4 staff
• Provide access to data, support and training
• Service free of charge to users
• Currently 3300 users from 100 sites

3
Database coverage

• Crystallography
• Synthetic Organic Chemistry
• Spectroscopy
• Physical Chemistry

4
Help/support

• Website - http//cds.dl.ac.uk/cds
• Phone/email us
• Manuals - mostly online
• Online help
• Online tutorials
• Flash movies

5
Registration

• Online system
• Individual ids required
• Current Rep

6
CDS Refunding

• Latest 3 years Refunding Grant began 1/4/04
• CDS continues to be supported by four staff
  members
• Boost to Physical Chemistry holding with
  successful application for funding for DETHERM
  thermophysical properties database
• Starting major publicity iniative with ambitious
  site visits programme
• Interim review of Service in 2005

7
Physical Chemistry

• DETHERM
• One of the world's largest thermophysical
  property databases of pure compounds and
  compound mixtures
• Contains 4.9 Million data sets for around 130,000
  systems
• (about 24,000 pure substances and 106,000
  mixtures)
• covering more than 500 property fields.

8
Physical Chemistry

• The recent CDS renewal grant included funding to
  acquire a full set of datasets from the supplier
  (DECHEMA e.V.) for use by the UK academic
  community
• For instance in the field of vapour-liquid-equilib
  rium data, it contains more than 95 of data
  published worldwide.
• Further details are available at the DETHERM
  pages on the CDS web site.

9
Publicity Initiatives

• In the light of discussions with the EPSRC
  following on from the meeting with the CDS Grant
  Review Panel training plans have been modified
  and enhanced
• Our Roadshow ideas have been refined and
  expanded. At each site will now give a CDS
  Overview lecture/seminar which takes place after
  a manned CDS poster and discussion session in the
  departmental foyer.
• The planned schedule of visits is advertised on
  the CDS web
• Fuller details of these and other aspects are
  given in the CDS 2003/4 Annual and Interim
  Reports

10
CrystalGrid 2004

• Interactions with e-Science
• Some Aims for the Future

11

• Current - What CDS has at the moment
• Future - Interactions with e-Science
• Example DLV
• Example Linking Databases - Crystal Web
• Metadata
• Collaboration Tools
• Archiving /Data Curation

12

• PRESENT
• A. Search individual database using proprietary
  software to see if compound, crystal, spectrum,
  data exists.
• B. Save/download/convert specific data for use
  with packages on their desktop machine.
• Or Conduct simple search of CDS databases using
  desktop package (currently only one )and then
  making use of some of that data in the package.

13

• Issues
• Proprietary databases, different systems,
• different front-ends no direct control
• by CDS.
• Data is different in each database.
• Cannot query ALL databases using
• one query.

14

• e-Science - Experience
• Integration into Problem Solving Environments
  (PSEs)
• Database access over the Grid (OGSA-DAI
  IBM Oracle)
• Data and metadata technologies for storage (XML
  etc)
• Authorisation and authentication.

15
Example DL Visualise (DLV)
Simple search of databases
Fires up computational packages
Takes crystal co-ordinates and produces displays
16

• DL Visualise
• Current
• 1. It is a kludged system.
• 2. Requires on going interactions between CDS and
  CSE
• to set up and maintain - sensitive to future
  modifications.
• Future - Input from e-Science
• Use standard protocols and definitions
• Publicised to community
• Web services - e-Science concept should do the
  job better.

17
(No Transcript)
18
Linking Databases Present - Crystal Web -
(miniGRID) able to search different
crystallography databases (cannot as yet use
drawn structure as query)
19

• Linking Databases - Future - Input from
  e-Science
• Metadata taken from disparate databases and
  merged into one database (e.g. Compound Locator
  idea MDL)
• Creates a meta data layer
• Transparently passes queries to the meta layer
• Transparently translates queries through
  metadata to
• different formats and different query types
• Transparently searches multi-data sources with
  different
• query formats/types
• Present results to the user

20

• Linking Databases - Future
• Structures
• May contain large amount of binary data.
• - Use Metadata but with links to full data.
• Requires generic chemistry format.
• XML?
• INChI? (IUPAC-NIST Chemical Identifier)

21
CDS and e-Science Collaboration Tools e.g.
Examining and manipulating datasets over the
network (e.g. rotating structure on colleagues
machine) Review data and add comments. Similar
technology to video conferencing and remote
experiments
22

• CDS and e-Science
• Archiving /Data Curation
• Data deposited and saved in one location or
  distributed around linked locations.
• Established e-Science aspiration
• Local DL expertise
• JISC Integrated Information Environment
  initiative
• Would also require incentive to users to add
  data!
• e.g. Spectral data mandatory for those with
  government grant to archive data

23
CDS and e-Science Archiving /Data Curation Some
tools already present at CDS (e.g. ISIS -
Screening Compound Database) data can be input
if users send it. Data entry tools required if
users enter data directly to local database.
Could then use batch entry to main database or
data could be harvested (whole or meta
data?) e.g. CrystalGrid elemental composition
and reduced cell data.
24
(No Transcript)
25
CCPs
CDS
Commercial databases
Communities
Training/infrastructure
?Serving structures to GUIs
Legacy formats
Internationalisation?
? building up additional data?
Added value from cross database integration
Computational codes Site Visits / PR
Integrated delivery of data compute services
Integrated delivery of data compute services
Testbed projects
Grid modalities for search / delivery
Comp Results Libraries
Hardware/software infrastructure
Project databases
Expert systems
E-Science
Extensible data representations
New mechanisms for authentication, authorisation,
eventually payment