Semantic Grid Services Semantic Grid Services

1
(Semantic Grid) Services Semantic (Grid
Services)

• Professor Carole Goble
• The University of Manchester, UK
• e-Science North West Regional Centre
• myGrid, OntoGrid, Knowledge Web
• GGF Semantic Grid Research Group

2

• The ongoing convergence between Grids, Web
  Services and the Semantic Web is a fundamental
  step towards the realisation of a common
  service-oriented architecture empowering people
  to create, provide, access and use a variety of
  intelligent services, anywhere, anytime, in a
  secure, cost-effective and trustworthy way.
• Next Generation Grids 2
• Requirements and Options for
• European Grids Research 2005-2010 and Beyond
• EU Expert Group Report July 2004

3

• To realise the Next Generation Grid requires
  semantically rich information representation, the
  exploitation of knowledge, and co-ordination and
  orchestration that is aware of context and task
• David Snelling, NextGRID
• Building Intelligent Grid Services

4
Knowledge everywhere alreadyits called metadata

• State properties of a resource
• Data in a purchase order
• Current usage agreement for resources on a grid
• Metrics associated with work load or performance
  on a Web server
• Declarative descriptions of data sets, codes,
  services, workflows
• Typing and classifying service or workflow
  inputs, outputs, goals,
• Access rights to resources
• Declarative descriptions for, and records of,
  service interactions
• event notification topics, provenance trails,
  monitoring records
• Policy and profile encoding
• personal profiles and security groupings
• Used in
• job control workflow composition, semantic
  dataset integration, resource brokering, resource
  scheduling, problem solving selection,
  intelligent portals
• GGF WG-CMM, CIM, GIS, MDS, .

5

• Knowledge and the knowledge producing
  consuming protocols patterns are already in
  Grid Middleware and Grid Applications.
• Embedded in middleware code, in schemas, in
  catalogues, in applications and in practice.

6
Bringing knowledge into the light

• Managing and operating a Grid intelligently
  requires
• 1. Knowledge
• Knowledge about the state and properties of Grid
  components, and their configurations
• Mechanisms for interpreting that knowledge
• 2. Intelligently acquiring and refreshing
  knowledge
• 3. Use it practically in decision making.

7
Convergence

• Semantic Web Technologies
• Semantic Web itself

8
Semantic Web mechanisms
Trust
?p -gt ?a pa
?p -gt ?a pa
Rules SWRL

• Uniform naming scheme.
• Metadata descriptions of properties and
  content
• Metadata glue linking resources together
• Ontologies interpretation of metadata for
  people and processes.

?p -gt ?a pa
?p -gt ?a pa
?p -gt ?a pa
Ontologies OWL/RDFS
Metadata Annotation RDF
Search engines and filters
Web XML, URI, UniCode
Applications
Deep web PHP, WS
9
Making Knowledge Explicit
OWL Web Ontology Language
RDF Resource Description Framework
10

• Make knowledge explicit.
• Make knowledge protocols explicit.
• Describe some of these declaratively so they
  might be exchanged and machine processed.
• Metadata data here is what it is and/or how it
  relates to something else
• Ontologies / controlled vocabularies we
  understand each other

11
Knowledge Stakeholders
Knowledge for Grid Applications
Knowledge for the operation of the Grid
Sources of Knowledge
12
knowledge worker'sapplications and tools
Grid Domain Applications
Upper domain generic services
Collective services
Plumbing
Application Knowledge
Base services
Operational Knowledge
System services
Web Service Resource Framework Web
Service-Notification WS-I
Web Services
13
The Semantic Grid is an extension of the current
Grid in which information and services are given
well-defined and explicitly represented meaning,
better enabling computers and people to work in
cooperation
Semantics in and on the Grid
14
Time to move beyond slogans.
15
Semantic Grid roadmap

• Exploit the languages from the Semantic Web and
  other.
• Specifying and developing the architectural
  components and tools forming the infrastructure
  of the Semantic Grid and define the architecture
  of the (Semantic) Grid.
• Prototyping applications using the languages, the
  components and defining the content necessary.
• Developing in parallel, yet are interdependent.
• A maelstrom of research coupled concurrently with
  standards activity, and early experiments and
  prototypes running alongside (some) commercial
  developments.

16
Semantic Grid trajectory
SDK
Demonstration Phase
Efforts
Systematic Investigation Phase Specific
experiments Part of the Architecture
Dagstuhl Schloss Seminar Grid Resource
Ontology Many projects
Pioneering Phase Ad-hoc experiments, early
pioneers
SRB
GGF Semantic Grid Research Group Many workshops
Implicit Semantics OGSA generation
Implicit Semantics 1st generation
Time
17
Three strands
Knowledge Aware Grid Services KAGS
Grid Compliant Knowledge Services GCKS
P4
Semantic (Grid Services)
(Semantic Grid) Services
Grid Aware Knowledge Services GAKS
And how all these services play
together Profiles, Protocols, Patterns, Policies
18
Three strands
Knowledge Aware Grid Services KAGS
Grid Compliant Knowledge Services GCKS
Middleware
Knowledge Additional port types relating to
knowledge, for example discovery.
Functionality Existing operations for
interaction with a knowledge service Metadata
How fast? What language is supported? Lifetime
Management Factory methods, creation of resources
Grid Aware Knowledge Services GAKS
Use of Grid infrastructure within the
implementation of the service.
19
Grid Compliant Knowledge Services

• Take todays knowledge services from the Semantic
  web and other worlds
• What does it mean for them to be Grid Services?
• What are the state properties of an ontology grid
  service?
• What are the lifetime management properties of an
  ontology grid service?
• What is a virtualised and dynamically provisioned
  ontology service, (metadata store, metadata
  annotator, reasoner ) ?
• How will an ontology grid service and a metadata
  grid service play together?

20
Grid Compliant Ontologies

• Resource
• A distinguishable unique identity and lifetime
  (usually static)
• Maintains a specific state that can be
  materialized
• May be accessed through one or more Web Services
• Artifact - a file, XML document, database,
  usually real (could be virtual). Could be
  compound.
• Service
• Base interface for inspecting and manipulating an
  ontology
• A well defined Ask-Tell API getSubConcepts(conc
  ept), getSuperConcepts(concept), classify,
  checkSatisfiability(concept), put(conceptExpressio
  n)
• Resource a connection to the Ontology Service
• An ontology might be just a file. Or an
  application. Or embedded in an application after
  a community has thought about it for a bit.

21
Ontology as an OGSA-DAI Realization
WS-DAI Message Patterns Behavioural Properties
Provide a realization of WS-DAI with specific
ontology messages (activities)
WS-DAIR Relational
WS-DAIO Ontology
WS-DAIX XML
WS-DAIO-RDF RDF Specific
WS-DAIO-OWL OWL specific
22
RDF Annotation store as an OGSA-DAI Realization
WS-DAI Message Patterns Behavioural Properties
Provide a realization of WS-DAI for RDF
WS-RDF
WS-DAIX XML
WS-DAIR Relational
DB2
mySQL
23
Data -gt Ontology Access

• Data Access collects together messages that
  access and/or modify a resource

• Note the messages are ignorant of the query
  other than its class.
• OSGA-DIAO
• The message patterns the behavioural properties
• The API for the ontology querying
• The realisation mapping to the ontology language
  OWL, RDF, RDFS, DAG

24
Knowledge Aware Grid Services

• Take a Grid service and see how it might take
  advantage of a knowledge service or knowledge
  resource.
• Might be a base Grid service or an Application
  Service or a high level Grid service.
• What are the generic and specific knowledge
  services required for Grid?
• Two starting points
• Discovery. Registry/Brokering shared semantics
  resource annotation painless knowledge recovery.
• Debugging shared semantics knowledge
  collection knowledge recovery.

25
Semantic Web Services

• Semantic Web describing data
• Semantic Web Services describing processes.
• WSMO, OWL-S

Thierrys observations about Web Service
abstractions
26
Discovery in Taverna workflow workbench

• Taverna currently ships with access to gt1000
  publicly available bioinformatics services
• Bioinformatican chooses services when forming
  workflows, with assistance.
• A common ontology is used to annotate and query
  any myGrid object including services.
• Discover workflows and services described in the
  registry via Taverna.
• Look for all workflows that accept an input of
  semantic type nucleotide sequence

27
Semantic Discovery
Low level descriptions WSDL, Scufl
Reasoner
Feta skeletons generated by mining low level
descriptions
myGrid domain classification
Ontology editor
Feta importer
Ontologist builds myGrid Domain Ontology
Knowledge Engineer
PeDRo annotator
Feta semantic discovery engine
Annotator
Descriptions are loaded and engine initiated
Search requests
Skeletal descriptions are annotated
Taverna workbench clients
UDDI registry
Feta GUI
KAVE provenance
Resource match make
User interacts with GUI to discover resources
Annotated descriptions are stored
28
Intelligent Debugging Architecture
Acklin
29
Keeping track
Relationship BLAST report has with other
Other classes of information related to BLAST
report
Jun Zhao, Chris Wroe, Carole Goble, Robert
Stevens, Dennis Quan, Mark Greenwood, Using
Semantic Web Technologies for Representing
e-Science Provenance in Proc 3rd International
Semantic Web Conference, Hiroshima, Japan, Nov
2004
30
Grid Aware Knowledge Services

• What is the architecture of distributed knowledge
  services?
• Can Grid platforms realistically provide a robust
  distributed stateful computing platform for agent
  systems?
• OGSA-DAIS for RDF repositories.
• Replica location service for replicated knowledge
  services.
• Secure file transfer for metadata.
• Event notification for metadata or ontology
  updates.
• Authentication and authorisation for updates.
• Metadata updated by workflows
• Security and RDF!
• Distributed reasoning !!
• Depends on the availability of these Grid
  services.

31
WS-Notification and Semantic Integrity

• Subscriber an Annotation Service - indicates
  interest in a particular (semantic) topic
  Ontology Version change - by issuing a subscribe
  request
• Subscriptions are WS-Resources
• Various subscriptions are possible
• Notification may be triggered by
• WS Resource Property value changes
• Other situations
• Broker examines current subscriptions
• Brokers may
• Transform or interpret topics lt- knowledge!

subscribe
notify
Metadata service
notify
notify
subscribe
S
S
S
Publisher
notify
Ontology Service
Adapted from Dr. Daniel Sabbah, IBM, Globus
World 2004.
32
Yet Another Stack
Car repair settlement, satellite data
configuration.
Grid Application and Application Services
resource discovery, intelligent debugging,
provenance mining
OGSA OntoKit knowledge Generation services
Patterns Upper Services Semantic broker,
semantic registry, semantic logging, semantic
workflow management, vocabulary management
PATTERNS OF INTERACTION
OGSA OntoKit semantic grid services
Base services annotation management ontology
access and integration, annotation access,
reasoning, ontology alignment GRID PROPERTIES
OGSA plumbing services
OGSA OntoKit plumbing services
Resources Ontology, Knowledge Base, Registry,
Database DOMAIN MIDDLEWARE
Resources
33
Obstacles to Overcome

• Semantic what?
• Compelling use cases
• Revolution is only possible when it becomes
  inevitable
• Niche activity.
• No content or hard to get the content!
• Ontology acquisition. Pain-free metadata
  acquisition.
• Baggage of communities
• Different agendas
• Hendler Principle A little semantics goes a
  long way.
• Failure to mainstream agents
• Instability of both platforms
• Middleware hard to use and incomplete
• Off putting to the other side
• Deployment, research, development, applications
  and standardisation all happening together
• Whither Grid Architecture?

34
MDA and the Grid
Prof.dr. iga Turk
Computation Independent Model

• Where is grid?
• current grids are on a platform level
• grids compatible with service oriented
  architectures are on ASM level
• Challenge
• should grids do better than SOA based on Web
  Services?
• automatic transformation of PIM models into a
  grid specific ASMs and PSMs
• Opportunity
• transform a business level architectures to Web
  Services, Grid, whatever-comes-next platform

manual
PlatformIndependent Model
automatic
ArchitectureSpecific Model
e.g. OGSA
automatic
Platform Specific Model
e.g. GT4, gLite
semi automatic
working system
35
Map concepts between ontologies

• Unicore and GLUE have different philosophies for
  describing resources -(
• In Unicore, the resources are described in terms
  of resource requests
• In GLUE, resources are described in terms of the
  availability of resources.

36
Not all knowledge will use separate services
Use
Explicit
Ontologies Rules Non-embedded metadata
Embedded metadata
Type systems
Schemata
Implicit
Text descriptions
Shared human consensus
Implicit
Explicit
Assertion
37
Source of metadata and knowledge

• Grid Resource Ontology
• Activation Energy
• Metadata mining
• The network effect service providers rule
• Return on investment for service providers and
  users
• Applications keep it real listen to users to
  take short cuts.

38
Semantic proportionsspeculation no empirical
foundation at all
Generic Grid
Resource
Application
39
Grid
Knowledge, Agents the Semantic Web

• Knowledge aware grid services

Overcoming community divisions Growing pains of
middleware Make it easier not harder or more
interesting A little semantics goes a long
way Evolution not revolution Technology push
40
WSRF is the instruction set of the Grid
Thierry Priol
Semantic Grid Services
Grid service behaviour
WSRF
WS-I
41
Whither Grid Architecture?
42
K-WfGrid
InteliGrids
Provenance
SIMDAT
Applications Use Cases
UniGrids
SDK
Grid Architecture
Semantic Architecture
NextGRID
WSRF
WS-I
Semantic Grid Architecture
43
Summary

• What existing technologies can we harness and
  what needs to be done that is new?
• Semantic SOA what are the resources, services,
  profiles, patterns and policies?
• What are the appropriate abstractions for a
  Semantic Grid based architecture? (or a Grid
  Architecture?)
• How will semantics make the Grid more flexible
  and simpler and how do we avoid making it more
  complicated!
• How do we ensure close cooperation with design
  and development of next generation Grid research
  and next generation knowledge research?

44
Thanks

• myGrid consortium, esp. Phil Lord, Pinar Alper,
  Chris Wroe, Luc Moreau
• OntoGrid project members
• Norman Paton, OGSA-DAI
• Prof.dr. iga Turk, InteliGrids
• John Brooke, UniGrids
• Stephane Viali
• Thierry Pioli, CoreGrid
• David de Roure, GGF Sem-Grd RG
• http//www.semanticgrid.org/