MARIO CANNATARO

1
Knowledge Discovery and Ontology-based Services
on the Grid

• MARIO CANNATARO
• University Magna Græcia of Catanzaro, Italy
• cannataro_at_unicz.it

Chicago, October 5, 2003
2
OUTLINE

• Exploiting data on the Grid
• Services for future Grid
• Knowledge Discovery and Ontologies
• Building knowledge services the KNOWLEDGE GRID
• KNOWLEDGE GRID Architecture and Services
• A case study ontology-based application design
• Conclusions and future work

3
EXPLOITING DATA ON THE GRID

• Main challenges for future Grids will be
• the exploitation of overwhelming amount of data
  produced by applications and Grid operation
• the exploitation of semantic of Grid resources
  and services when using them
• Grid Intelligence reflects how data and
  information available over Grids can be
  effectively acquired, represented, exchanged,
  integrated, and converted into useful knowledge
• Transforming challenges into opportunities
• Knowledge discovery and knowledge management
  services to enable new applications and enhance
  Grid management

4
INTELLIGENT SERVICES FOR FUTURE GRIDS

• Ontologies and metadata are the basic elements
  through which new Grid services can be built
• Ontologies enable semantic modeling of Grid
  resources, services, and users tasks/needs
• Resource ontologies and metadata allow
  intelligent searching and browsing
• Data Mining and Knowledge Management could enable
  high level services based on the semantic of
  stored data for
• application services implementing data analysis
  (DataCentric Grids)
• Grid management services based on analysis of
  usage data
• Peer-To-Peer and Ubiquitous Computing may be the
  orthogonal key technologies to build basic
  services
• presence management, resource discovery and
  sharing, collaboration and self-configuration

5
GRID KNOWLEDGE BASE

• Grid knowledge base is used here to indicate all
  the data stored, maintained, and updated by the
  Grid, both for application and operation purposes
• e.g. Globus metadata, Grid services usage data,
• application data, etc.
• A challenge will be their seamless integration
  and utilization.
• metadata and ontologies define semantic views on
  the Grid knowledge base
• each object on the Grid is classified through
  one or more ontologies into the knowledge base
• Needed building and integrating resource/services
  ontologies

6
ONTOLOGY-BASED SERVICES

• Leverage the semantic models of Grid objects to
  navigate and search the Grid knowledge base
• e.g. ontology-based Grid programming
• choosing, accessing and using software and data
  components, classified through ontologies, to
  build distributed component-based applications on
  Grid
• e.g. request-resource matchmaking
• find the best match (not necessarily identity)
  between job requests and Grid resources, both
  modeled through ontologies

7
KNOWLEDGE DISCOVERY SERVICES

• They are used to extract knowledge from the data
  stored into the Grid knowledge base
• ontologies could guide the selection data of
  useful
• data mining is applied to selected data
• e.g. choosing the best GridFTP connection
  parameters (TCP buffer size, parallel streams)
  by
• clustering GridFTP usage data and
• classifying new transfer requests with data
  mining techniques
• e.g. Grid-based document management service
• that classify Grid documents
• using ontologies and text mining functions

8
SEMANTIC COMPRESSION AND SYNTHESIS

• Semantic (lossy or lossless) compression and
  synthesis could be used
• to offer different views of the Grid knowledge
  base
• depending on many factors user/service goals,
  scope of resource information
• Contents can be reorganized according to
• some aggregating functions (schema extraction and
  restructuring),
• resulting in a synthetic (compressed) yet
  meaningful version
• Different views of Grid resources exposing
  different levels of details could be provided to
  different classes of users/services

9
CONTEXT-AWARENESS

• When new devices and resources are allowed to
  enter / exit the Grid in a very dynamic way
• new services able to adapt themselves to the
  environment have to be developed
• Ubiquitous Computing and Adaptive Hypermedia
  techniques could be used
• to face the environment dynamicity and
• to adapt services

10
PARALLEL DISTRIBUTED KNOWLEDGE DISCOVERY
SYSTEMS ON GRIDS

• When
• large data sets are coupled with
• geographic distribution of data, users, and
  systems,
• To mine such data it is necessary to implement
• parallel and distributed knowledge discovery
  systems (PDKD).
• The basic principles that motivate the
  architecture design of grid-aware PDKD systems
• Data heterogeneity and large data size
• Algorithm integration and independence
• Grid awareness
• Openness, Scalability, Security and data privacy

11
THE KNOWLEDGE GRID

• KNOWLEDGE GRID - a PDKD architecture that
  integrates data mining techniques and
  computational Grid resources.
• In the KNOWLEDGE GRID architecture data mining
  tools are integrated with lower-level Grid
  mechanisms and services and exploit Data Grid
  services.
• A KNOWLEDGE GRID application uses
• A set of KNOWLEDGE GRID-enabled computers -
  K-GRID nodes
• declaring their availability to participate to
  some PDKD computation, that are connected by
• A Grid infrastructure
• offering basic grid-services (authentication,
  data location, service level negotiation) and
  implementing the KNOWLEDGE GRID services.

12
KNOWLEDGE GRID ENVIRONMENT
KNOWLEDGE GRID services
Basic Grid Infrastucture
K-GRID tools
K-GRID tools
Grid Middleware
Grid Middleware
LAN
Cluster Element
Cluster Element
Cluster Element
Grid Middleware
K-GRID node
Cluster containing data sets and/or DM algorithms
Generic Grid node
K-GRID node
13
KNOWLEDGE GRID SERVICES

• The KNOWLEDGE GRID services are organized in two
  hierarchic layers
• Core K-Grid layer and
• High-level K-Grid layer.
• The former refers to services directly
  implemented on the top of generic Grid services.
• The latter is used to describe, develop, and
  execute PDKD computations over the KNOWLEDGE
  GRID.

14
KNOWLEDGE GRID ARCHITECTURE
KNOWLEDGE GRID
15
APPLICATION COMPOSITION STEPS
Metadata about K-grid resources
KMRs
Search and selection of resources
DAS / TAAS
Metadata about the selected K-grid resources
TMR
Design of the PDKD computation
EPMS
Execution Plan
KEPR
16
APPLICATION EXECUTION STEPS
17
OBJECTS and LINKS
Objects
Links
Objects represent resources
Links represent relations among resources
18
VEGA
Hosts pane
Resources pane
19
VEGA
A KGrid application can be composed of several
workspaces
20
APPLICATION EXECUTION
21
ONTOLOGY-BASED APPLICATION DESIGN

• Problem making easier the development of
  distributed Data Mining applications on the
  Knowledge Grid
• many available pre-existing software tools
• many different implementations and methodologies
• Solution develop a Domain Ontology for the Data
  Mining domain, and using it to enhance the design
  of component-based distributed data mining
  applications
• DAMON DAta Mining ONtology

22
DAMON GOALS

• For what we are going to use DAMON
• Classify Data Mining (DM) Software on the basis
  of some parameters useful to select the more
  suitable ones to solve a KDD problem
• The Data Mining task performed by the software
• The methodologies that software uses in the data
  mining process
• The kind of data sources the software works on
• The degree of required interaction with the user
• For what type of questions the ontology should
  answer
• to assist the user suggesting him/her the
  software to use on the basis of the users
  requirements/needs
• to allow the semantic search (concept-based) of
  data mining software and others data mining
  resources

23

• DAMON taxonomies
• Task the knowledge discovery goal that the
  software is intended for
• Method a DM methodology used to discover the
  Knowledge
• Algorithm the way through which a DM task is
  performed
• Software an implementation of a DM algorithm
• Suite implements a set of DM algorithms
• Data Source the input on which DM algorithms
  work to extract new knowledge
• Human Interaction specifies how much human
  interaction with the discovery process is
  required and/or supported

24
DAMON A VIEW OVER THE GRID KNOWLEDGE BASE

• The Data Mining knowledge base used to support
  knowledge discovery programming has two
  conceptual layers
• at the top layer the DAMON ontology gives general
  information about the Data Mining domain,
• at the bottom layer specific information about
  installed software components and data sources
  are maintained where resources resides.
• From an architectural point of view the ontology
  is a central resource, whereas specific metadata
  are distributed ones.
• As an example, DAMON stores the fact that the
  C5.0 Software implements the C5 Algorithm, that
  uses the Decision Tree method, that is a
  Classification method.
• The C5.0 Software node of ontology contains the
  URLs of the KNOWLEDGE GRID metadata files
  describing details about all the installed
  instances of that software.

25
ONTOLOGY-BASED SERVICES ACCESSING THE GRID
KNOWLEDGE BASE

• DAMON is used as an ontology-based assistant that
  helps the KNOWLEDGE GRID application designer in
• application formulation and design,
• selection and composition the most suitable Data
  Mining components for a specific knowledge
  discovery process
• semantic search of data mining software.

26
SEMANTIC SEARCH OF DATA MINING SOFTWARE

• Ontology-based resources selection.
• Browsing and searching the ontology allows a user
  to locate the more appropriate tasks, methods,
  algorithms and finally data mining software to be
  used in a certain phase of the KDD process.
• The user can query very detailed information
  about Data Mining resources, using several kinds
  of inference that can broaden queries..
• Metadata access.
• The ontology return the metadata URLs of all
  selected resources available on the KNOWLEDGE
  GRID nodes
• metadata are used to access and use software
  (technical parameters, policies, location and
  configuration, etc.).

27
KNOWLEDGE GRID WITH ONTOLOGY SERVICES
KNOWLEDGE GRID
28
CONCLUSIONS AND FUTURE WORK

• The increasing volume of data available on the
  Grid is a challenge but also an opportunity
• Ontology and Knowledge Discovery are key
  technologies to extract and synthesize useful
  information for applications and Grid management
• The KNOWLEDGE GRID allows to develop distributed
  data mining applications on the Grid
• ontology-based services are used in components
  search and selection by using a domain ontology
  DAMON
• We are defining a set of Grid Services using the
  OGSA conventions and mechanisms that export
  functionalities and operations of the KNOWLEDGE
  GRID

29
THANKS
30
MAIN REFERENCES

• M. Cannataro, D. Talia,  The Knowledge Grid,
  Communications of the ACM, 46(1), 2003.
• M Cannataro, D. Talia, P. Trunfio, Distributed
  Data Mining on the Grid, Future Generation
  Computer Systems, 18(8), 2002.
• M. Cannataro and C. Comito, A Data Mining
  Ontology for Grid Programming, 1st Int. Workshop
  on Semantics in Peer-to-Peer and Grid Computing
  May 2003