Get documents about "What the Grid can do for the Semantic Web
Document Sample
What the Grid can do for the
Semantic Web
David De Roure
University of Southampton
dder@ecs.soton.ac.uk
Intersection of Communities
Grid Semantic
Semantic
Grid Web
July 2005 Dagstuhl 2
Intersection of Communities
onward and upward!
July 2005 Dagstuhl 3
Help!
Idealistically this is the Wrong Question
My papers and talks go from the user requirements to
the techniques to meet them
But perhaps it’s the right question as a mechanism
to build bridges
And this is the opposite to my normal talk
I usually explain to Grid People what the Semantic Web
(or at least machine-processable/understandable
metadata) can do for Grid middleware and apps
That’s the larger flow…
July 2005 Dagstuhl 4
Building Bridges
Web
Services
Semantic
Grid
Semantic
Grid
Web
and Agents
Theoreticians Engineers
Semantics for the Grid
Grid services for Semantic Web
Goble
July 2005 Dagstuhl 5
Semantic Grid
Interoperability
Semantic Semantic
Web Grid
Scale of
Classical Classical
Web Grid
Scale of data and computation
July 2005 Dagstuhl Based on an idea by Norman Paton 6
Semantic Grid
Interoperability
Semantic Semantic
Web Grid
Scale of
Classical Classical
Web Grid
Scale of data and computation
July 2005 Dagstuhl Based on an idea by Norman Paton 7
Let’s get the really obvious
things out the way first
Funding
Sometimes the Grid can provide funding for
Semantic Web research
Here’s how to do it…
Pay lip service to Grid
e.g. use the phrase web/grid services
Use toy examples
Don’t talk to Grid community
Duplicate (small subsets of) the Grid infrastructure
Aim to do real example at end of project and never get
round to it
Publish in Semantic Web literature
NB I am not saying that this is wrong!
July 2005 Dagstuhl 9
Big iron fat pipes
Semantic Web as Grid application
Computation could be useful for
Crunching vast quantities of data to extract knowledge
Reasoning (in various parts of lifecycle)
Doing stuff with big ontologies or lots of little ontologies
Grand challenges in AI
Actually this is quite exciting! This is stuff we
couldn’t do before.
July 2005 Dagstuhl 10
Slightly more insightful
Yolanda’s motivations
Integrating distributed info sources – needed
robustness in a dynamic, loosely coupled system,
not babysitting.
Need to detect if service dies, introspection,
persistence, shelf-life, state management.
“This is boring”
Hooray for Grid Services!
Jade etc aren’t the answer(?)
NB Good to want a distributed infrastructure - some
Semantic Web Research is very Semantic and not a lot of
Web !
July 2005 Dagstuhl 12
Robust Distributed Infrastructure
A new infrastructure to use and explore, which
supports
Both computational and data resources
Fault tolerant
QoS / QoX
Breadth and number of services
Crosses organisational boundaries
Long lived processes
Secure
Raises issues not raised by Web/Web Services
Note the things in this list which need representations
July 2005 Dagstuhl 13
Mature case studies
What does the (Semantic) Grid give the Semantic
Web community?
Interesting Semantic Web deployments!!!!
(= subwebs of The Semantic Web?)
July 2005 Dagstuhl 14
IST
July 2005 Dagstuhl 15
Semantic Grid-related projects
OntoGrid Inteligrid SIMDAT Akogrimo
DataMiningGrid K-WF Grid CoreGRID UniGrids
Provenance NextGRID China Semantic Grid Plan
CLEF Cognitive Grids COG CoAKTinG Conoise-G
Comb-e-Chem DartGrid DynamO eBank UK
FEARLUS-G Geodise GRIA GRIP HUST CGCL
Semantic Grid CO-ODE/HyOntUse MAGIK-I MIAKT
myGrid myTea PASOA Pegasus REOL SCEC/IT
SCULPTEUR A Semantic Firewall Semantic Media
VOeS WUN Grid
July 2005 Dagstuhl 16
Combe Chem pilot project
Video
Simulation
Properties
Analysis
Diffractometer
Structures
Database
X-Ray Properties
e-Lab e-Lab
Grid Middleware
July 2005 Dagstuhl 17
CombeChem Semantic Web
July 2005 Dagstuhl 18
The CombeChem Experience
Peaked at about 80 million RDF triples
3 triplestores (Jena, Kowari, 3Store)
Today 19M in Kowari, 11M in 3store, need 150M but
can’t store this
Schema reflecting practice (built units ontology)
Whole lifecycle approach
Automated capture of semantic annotation
Live, operational system
Hence a platform for Semantic Web Research and
new chemistry!
July 2005 Dagstuhl 19
A chemist speaks
The Semantic Web is an ambitious goal requiring
contributions from multiple players in order to
achieve maximum benefit. In chemistry, only a
comparatively small population is interested in any
particular area. One can conceive of free data
exchange, banishment of the proprietary file
format, but there are parties who do not want to
make data more easily available to their
competitors. However, we have demonstrated the
value of adopting this approach on the scale of
our project.
July 2005 Dagstuhl 20
A chemist speaks some more
There is another side to RDF, not often trumpeted
by web developers, and that is its use as a local
storage and reference system. Not all data needs
to be made accessible on the Web, and
intellectual property issues may prevent such
publication, but the flexibility of the RDF triple
model allows it to be applied with several key
advantages over conventional approaches.
July 2005 Dagstuhl 21
Users with requirements
www.smarttea.org
July 2005 Dagstuhl 22
Meeting Replay
Compendium
BuddySpace
Jabber
Server Compendium
BuddySpace
I-X Process panels I-X Process panels
KMI, AIAI
July 2005 Dagstuhl 23
The Global Grid Forum
Building a broad
Defining grid
international
specifications that
community for the
lead to broadly
exchange of ideas,
adopted standards
experiences,
and interoperable
requirements, and
software
best practices
Operations
Ensuring ongoing support of our mission
and communication of our progress
Leading the pervasive adoption of grid computing for research and industry
July 2005 Dagstuhl GGF 24
Developers
Active developer community, with community
processes e.g. GGF
Community forum needed to agree interop
An ontology is not useful unless shared!
Grid techies are willing to try stuff out if it’s going
to solve a real problem for them (even if it’s not
The Right Stuff!)
Provides hackers to complement SW theorists
July 2005 Dagstuhl 25
Semantic Web Services
Semantic (Web Services)
or
(Semantic Web) Services
July 2005 Dagstuhl 26
OntoGrid Objectives
Strategic Technical
Pioneer the use of Knowledge Services that
Knowledge Technologies are Grid Aware and Grid
(KT) to enhance and Compliant
extend architecture and Grid Services that are
design of Grid computing Knowledge Aware
systems
Enable Deployment of Knowledge Grid
Services Services
Knowledge Technologies
in Grid Architectures
Grid Knowledge
Deploy prototypes in the Aware Aware
context of real world
Business Applications
Goble
July 2005 Dagstuhl 27
Many new challenges
How to engineer big RDF stores
How to architect distributed RDF stores
How to do things in real time
How to work in a large scale distributed system
Describing grid services
Missing attributes, inconsistencies, conflicts
Relationship between Grid Services and agents
Agents as resources
Autonomic aspects
Virtual Organisations
July 2005 Dagstuhl 28
