Transcript Development of Front End tools for Semantic Grid Services

Development of Front End tools for
Semantic Grid Services
Review Meeting – March 24, 06
Dr.S.Thamarai Selvi,
Professor & Head,
Dept of Information Technology,
Madras Institute of Technology, Anna University,
Chennai.
March 23, 2006
M.I.T., Anna Univ, Chennai
1
Objective
To develop a Front End Tools for Semantic Grid Services that enables

the Service Requester to search a particular Grid Service/Resource
Semantically.

the Service Provider to describe a Grid Service/Resource
Semantically.
March 23, 2006
M.I.T., Anna Univ, Chennai
2
Activities

Study of current version of Globus Toolkit and study of
Semantic Grid Services



Understand the grid architecture and study of globus toolkit.
Study of languages needed to implement semantic grid services.
A prototype model for semantic grid services


Extending the UDDI registry to include semantic advertisements using TModels.
Design and Development of algorithms for intelligent discovery of grid services.

Design and Development of Grid Resource Portal

Functional testing and optimization of implementation
March 23, 2006
M.I.T., Anna Univ, Chennai
3
Road Map

Understanding various Components of Globus Toolkit 4.0

Understanding Semantic Web Services.

Understanding the technology used to develop Semantic Web
Services.

Understanding Semantic Grid Services.

Developing a typical prototype for Semantic Grid Services.
March 23, 2006
M.I.T., Anna Univ, Chennai
4
Summary of last review meet – Sep 02, 2005
Report

Layered Architecture of semantic grid service

Concepts and Tools to build semantic web applications that
includes:-
*Protégé, an OWL editor.
*Algernon, an inference engine
*OWLS for service descriptions

The concept of Matchmaking of services using OWLS

An application has been built that retrieves information from
resource ontology using algernon.
March 23, 2006
M.I.T., Anna Univ, Chennai
5
Summary of last review meet – Sep 02, 2005
Observations and Suggestions

Creation of resource ontology may be automated
* We have developed a java module using protégé-OWL APIs to develop
and manage ontology. We can use this module for managing resource ontology
* We are also developing a tool to convert WSDL into OWLS with which we
can create service ontology without manual intervention.

Aggregation of resource information may be automated ( may use
RFC 2016 and RFC 2608)
* We made thorough literature survey of GIIS of Globus toolkit and we
identified wsrf-query and grid-info-search toolS in Globus toolkit with which we
can aggregate the grid resource information
March 23, 2006
M.I.T., Anna Univ, Chennai
6
Status as on March 20, 2006

Implementation Issues have been identified

Several Approaches for implementing Knowledge layer
have been identified.

Implementation of Semantic Grid Architecture using
proposed approaches.

Proposal of Versatile Knowledge layer.
March 23, 2006
M.I.T., Anna Univ, Chennai
7
Focus Today

Semantic Grid Architecture

Difficulties with Conventional Mechanisms

Proposed Approaches for Knowledge Layer

Semantic Grid Architecture using Protégé Enabled Globus toolkit(PEG)

Semantic Grid Architecture using Resource Ontology Template

Further Scope

Conclusion
March 23, 2006
M.I.T., Anna Univ, Chennai
8
The Semantic Grid is an extension
of the current Grid in which
information is given a welldefined meaning, better enabling
computers and people to work in
cooperation
March 23, 2006
M.I.T., Anna Univ, Chennai
Semantic
Grid
9
Semantic Grid Architecture
This layer act as an infrastructure to support
the management and application of scientific
knowledge to achieve particular types of goal
and objective shared and Maintained
This layer deals with the way resources
are represented, stores, shared and
Maintained
This layer Manages allocation of
computational resources, Job Execution,
Secure Access to grid resources
Knowledge Layer
Information Services Layer
Data Services Layer
Computation Services Layer
Resources Includes Supercomputers, clusters
Workstations etc.,
March 23, 2006
Distributed Resources
M.I.T., Anna Univ, Chennai
10
Related Tools

Ontology
You need an Editor to Create Ontology

Inference Engine
To retrieve Knowledge from Ontology
March 23, 2006
M.I.T., Anna Univ, Chennai
11
Ontology


Ontologies are used to capture knowledge about some domain of
interest.
Ontology describes the concepts in the domain and also the
relationships that hold between those concepts

Complex concepts can therefore be built up in definitions out of
simpler concepts.

Ontology Web Language (OWL) is widely used to create
Ontology
Ex : Protégé, an OWL editor
March 23, 2006
M.I.T., Anna Univ, Chennai
12
Limitation of OWL
Though OWL has a well-defined semantics, but
it is not sufficiently expressive to characterize and describe
services
So, OWL-S, OWL for Service
March 23, 2006
M.I.T., Anna Univ, Chennai
13
OWL-S
OWL-S is an OWL-based web service ontology,
which supplies a core set of markup language
constructs for describing the properties and
capabilities of web services in unambiguous,
computer interpretable form
March 23, 2006
M.I.T., Anna Univ, Chennai
14
Limitations of OWLS


Though OWLS has WSDL2OWLS, but it cannot convert
Grid WSDL to OWLS.
It cannot recognize WSRF specific WSDL elements. Hence
we need to compromise while using the tool WSDL2OWLS
March 23, 2006
M.I.T., Anna Univ, Chennai
15
Challenges in OWLS
Difficulties
Currently there is no tool available to create Grid Service Ontology
automatically from its WSDL file
Solution
• We need to create Service Ontology using Protégé Editor
• Need to identify a tool to convert Grid WSDL into OWLS descriptions.
• Need to automate semantic descriptions of resource
March 23, 2006
M.I.T., Anna Univ, Chennai
16
Proposed Approach
Semantic Description and Discovery of Grid Services

We propose and implement Semantic Grid Architecture by integrating protégé editor with
Globus Toolkit and implements Parameter Matchmaking Algorithm for semantic discovery of
services.
Semantic Description and Discovery of Grid Resources

We propose a five layered semantic grid architecture using Gridbus broker that addresses the
need of semantic component in the grid environment to discover and describe the grid
resource semantically
Also

It is decided to devise a knowledge layer for semantic description of resources and its
retrieval, for semantic description of services and matchmaking of advertised grid services
against the requested ones.
March 23, 2006
M.I.T., Anna Univ, Chennai
17
Semantic Description and Discovery of Grid Services
March 23, 2006
M.I.T., Anna Univ, Chennai
18
Protégé Enabled Globus Toolkit

Globus Toolkit (GT) lacks a component to describe concepts semantically.

In PEG, protégé has been integrated into Globus Toolkit 4.0


It addresses the demands of a single toolkit to build grid infrastructure as
well as for semantic description and representation of services and
resources.
Globus user now can develop ontology for their services.
March 23, 2006
M.I.T., Anna Univ, Chennai
19
Architecture of Semantic Grid Service using PEG
Semantic
Discovery portlet
Data Management
Knowledge Layer
Service Ontology
File/Data
Protégé_3.1
MDS
GRAM
Application Layer
Application Portlet
Semantic
Component
Tokenizer
Information
Grid Information Portlet
Computational Grid
Services (High level
Grid Services)
GridFTP
Authentication
Authorization
GT4 Middleware
Grid Middleware
Services
GSI
R2
R1
March 23, 2006
Resources
Fabric Layer
R3
R4
M.I.T., Anna Univ, Chennai
20
Service Matchmaking

Refers to capability matching which means to compare requested service
description with the advertised service description to determine how similar
they are.

Matchmaking algorithm uses Inputs and Outputs for matching and does
not consider Preconditions and Effects as they are not sufficiently
standardized to be considered for matchmaking

On an optional basis, other properties can also be taken into account
assuming they have been described using any specific ontology language
such as OWL.
In our Parameter Matchmaking Algorithm, We propose to use Inputs,
Outputs and Functionality for matchmaking of services.
March 23, 2006
M.I.T., Anna Univ, Chennai
21
Parameter Matchmaking Algorithm

We introduce Parameter Matchmaking Algorithm that computes
degree of matching of service advertisement (A) and request (R)

The algorithm compares IOF of A and that of R, computes various
degree of matches namely:Exact:
A and R exactly matches A(IOF) ≡ R(IOF)
Plug-in: A offers more functionalities than R A(IOF) ≥ R(IOF)
Subsume:
R requests more functionalities than advertised A
A(IOF) ≤ R(IOF)
Intersection:
Not all functionalities matches A(IOF) ∩ R(IOF)
Disjoint: A and R does not match A(IOF) ≠ R (IOF)
March 23, 2006
M.I.T., Anna Univ, Chennai
22
Processes Involved
R(I)
A(I)
R(O)
Input
Matching
A(O)
R(F)
Output
Matching
Ir
A(F)
Functionality
Matching
Or
Intermediate Ranks after
Comparing I, O, F individually
Fr
Aggregate
Module
Computes Final ranked
Degree of match
Ranked Degree of Match
Parameter Matchmaking Algorithm
March 23, 2006
M.I.T., Anna Univ, Chennai
23
Functional Model of the Knowledge Layer
March 23, 2006
M.I.T., Anna Univ, Chennai
24
Implementation
Service Provider

A GridService that implements four functionality namely Addition, Subtraction,
Multiplication and Division.

Service Ontology has been created using Protégé editor to describe the service.
A sequence diagram of service provider
March 23, 2006
M.I.T., Anna Univ, Chennai
25
Service Requester

The requester submits Query.

The semantic component extracts F from the query (RF) and also from service
ontology (AF).

It compares RIOF with AIOF and computes ranked degree of match
{exact, plugin, subsume, intersection, disjoint}
Sequence diagram of service requester
March 23, 2006
M.I.T., Anna Univ, Chennai
26
Experimental Results
Sl.No
Capability Requested
Degree of
Match
Possibility of
service
invocation
1
Addition and Subtraction
Plug in
True
2
Addition, Subtraction, Multiplication and
Exact
True
Intersection
True
Division
3
Addition, Subtraction and Reversal of
string
4
Squaring and Temperature service
Disjoint
False
5
Addition, Subtraction, Multiplication and Division,
Temperature Service
subsume
True
6
Multiply, add, divide
Plug in
True
7
Square service
Disjoint
False
8
Addition and factorial
Intersection
True
9
Add, sub, divide and Multiply
Exact
True
March 23, 2006
M.I.T., Anna Univ, Chennai
27
Snapshots
March 23, 2006
M.I.T., Anna Univ, Chennai
28
Parameter Matchmaking Algorithm
Algorithm: Parameter Matchmaking Algorithm
Input: Advertised_Ontology A, Requester_query R
Output: Degree_of_Match M
Rank: input_rank,output_rank,functionality_rank
parse Ainto A(I1,I2,..Im),A(O1,O2,..On) and A(F1,F2,..Op)
parse Rinto R(I1,I2,..Ir),R(O1,O2,..Os) and R(F1,F2,..Ot)
c1=0, c2=0,c3=0
for each parsed A( I1,I2,..Im), A(O1,O2,..Om), A( F1,F2,.Fm)
do
if A(Ii)==R(Ij) then c1++;
if A(Oi)==R(Oj )then c2++;
if A(Fi)==R(Fj) then c3++;
end if
end for
March 23, 2006
input_rank=compute_intermediaterank(m,c1,r)
output_rank=compute_intermediaterank(n,c2,s)
functionality_rank=compute_intermediaterank(p,c3,t)
M=leastof(input_rank, output_rank, functionality_rank)
Rank compute_intermediaterank(i,c,j)
{
if(i==c==j) then R=1;
if(i>c=j), then R=0.75;
if(i=c<j), then R=0.50;
if(i>c<j), then R=0.25;
if(i!=c!=j), then R =0;
}
M.I.T., Anna Univ, Chennai
29
The Scenario of plug in match
March 23, 2006
M.I.T., Anna Univ, Chennai
30
The Scenario of Exact match
March 23, 2006
M.I.T., Anna Univ, Chennai
31
The Scenario of Intersection match
March 23, 2006
M.I.T., Anna Univ, Chennai
32
The Scenario of Disjoint match
March 23, 2006
M.I.T., Anna Univ, Chennai
33
Semantic Description and Discovery of Grid Resources
March 23, 2006
M.I.T., Anna Univ, Chennai
34
Objective
To propose a five layered semantic grid architecture with
knowledge layer at the top of gridbus broker

Knowledge layer – Semantic grid resource description using
adaptive ontology template and Knowledge discovery using
Algernon inference engine.
March 23, 2006
M.I.T., Anna Univ, Chennai
35
Motivation

Conventional mechanisms

UDDI

MDS

They offer searching mechanism based on keywords.

The node providers need to agree upon attribute names and values.

In grid like environment, where resources come and go there is always a
demand for framework to support semantic description and discovery of
services and resources.
March 23, 2006
M.I.T., Anna Univ, Chennai
36
A Five Layered Architecture of Semantic Grid Services
March 23, 2006
M.I.T., Anna Univ, Chennai
37
Knowledge Layer

Comprises two modules – Semantic Description and Discovery
Semantic Description

Domain Knowledge of grid is represented in ontology template

MDS is used to ‘plug’ grid resource information

Protégé-OWL APIs are used to build knowledge base of the grid using ontology
template
Semantic Discovery

Algernon inference is used to retrieve resource information
Job Descriptor

Creates Application Description File and Resource Description File to run the broker
March 23, 2006
M.I.T., Anna Univ, Chennai
38
Ontology Template
Definition – 1
Any resource can be modeled as an instance of a specific class provided that the
resource can be described using the properties defined in that class.
Definition – 2
An ontology template is the domain specific ontology that provides hierarchy of
classes with properties to define characteristics.

Protégé-OWL APIs are used to describe grid resources in the ontology
template.
March 23, 2006
M.I.T., Anna Univ, Chennai
39
Resource Ontology Template
March 23, 2006
M.I.T., Anna Univ, Chennai
40
Grid Resource Knowledge base
March 23, 2006
M.I.T., Anna Univ, Chennai
41
Semantic Component
Job execution
To user …
GridBus
Broker
Resource
Description
MDS
Res. Des File
Results
App. Des File
Job Descriptor
Submit
Job
Resource
Information
User
Resource
Semantic
Description
Semantic
Repository
Resource
Discovery
Query
Generator
User
Querying
OWL file
March 23, 2006
M.I.T., Anna Univ, Chennai
Algernon
Query
Request
42
Semantic Description
GIIS
service runs on globus machine will retrieve resource information of the
local host and stores it in LDAP server from where we can query the
information.
Protégé-OWL
provides versatile libraries with which one can manage ontology
and knowledge base. With those APIs insertion and removal of resources are
possible
OWLNamedClass computerC=owlmodel.getOWLNamedClass("WorkStation");
OWLDatatypeProperty hasIP = owlModel.getOWLDatatypeProperty("hasIP");
cpuI.addPropertyValue(owlModel.getOWLObjectProperty("hasCPUVendor"),cVendorI);
computerI.addPropertyValue(owlModel.getOWLObjectProperty("hasCPU"),cpuI);
March 23, 2006
M.I.T., Anna Univ, Chennai
43
Semantic Discovery

We use Algernon Inference Engine to retrieve information semantically.

This module accepts user query in the form of A:opB and converts it into
Algernon query to interact with the knowledge base.

Once suitable resource is discovered, user’s job will be submitted to gridbus
broker for execution.

This Knowledge Layer is implemented in Gridbus Broker, it can support most of
the popular middlewares including Globus, Alchemi etc.,
March 23, 2006
M.I.T., Anna Univ, Chennai
44
Snapshots
March 23, 2006
M.I.T., Anna Univ, Chennai
45
Protégé Ontology Editor
March 23, 2006
M.I.T., Anna Univ, Chennai
46
March 23, 2006
M.I.T., Anna Univ, Chennai
47
March 23, 2006
M.I.T., Anna Univ, Chennai
48
March 23, 2006
M.I.T., Anna Univ, Chennai
49
March 23, 2006
M.I.T., Anna Univ, Chennai
50
March 23, 2006
M.I.T., Anna Univ, Chennai
51
March 23, 2006
M.I.T., Anna Univ, Chennai
52
Further Scope
Semantic Grid Architecture using PEG

Currently, service descriptions are done manually. We are in the verge of
developing a tool to convert WSDL of WSRF services into OWS
descriptions thereby overcoming the limitation of human intervention for
creating service ontology.

Wide literature survey of domain ontology is required.

Ontology clustering can be implemented to improve the performance of
matchmaking.

Storing OWLS descriptions into UDDI registry is to be resolved for better
management of semantic descriptions
March 23, 2006
M.I.T., Anna Univ, Chennai
53
Semantic Grid Architecture using Resource Ontology
Template

The discovery module relies on the power of Inference engine used to
retrieve information semantically from the Knowledge base. Since Algernon
is an rule based inference engine, we need to implement rules to improve
the efficiency of searching Mechanism

Workflow Engine
Integrating Workflow component with the knowledge layer.
March 23, 2006
M.I.T., Anna Univ, Chennai
54
Extended Knowledge Layer

With these observations and scope, we present a versatile knowledge layer
that performs,

Semantic Description of resource using ontology template

Semantic description of services using GridWSDL2OWLS

Managing OWLS descriptions in UDDI registry

Clustering the OWLS descriptions using appropriate clustering mechanism

Implementation of QoS based Matchmaking Algorithm with the help of
domain ontology.

March 23, 2006
Implementation of Rule based semantic search engine
M.I.T., Anna Univ, Chennai
55
Extended Knowledge Layer
March 23, 2006
M.I.T., Anna Univ, Chennai
56
Conclusion

The semantic grid architecture using PEG enables the service providers to describe
their grid services semantically. Whereas, the architecture using Gridbus broker,
provide semantic descriptions of grid resources using grid resource ontology
template.

We also identify the necessity of GridWSDL2OWL-S tool and is being developed in
our Grid Computing Laboratory. We made a wide literature survey of ontology
clustering with which the performance of ontology matchmaking can be improved.

With these observations, we propose a versatile knowledge layer which can be
implemented in the grid architecture that performs semantic descriptions of grid
resources, WSDL description of WSRF services into OWL-S descriptions, Discovery
of Suitable Grid resources, Ontology clustering and QoS based Matchmaking
algorithm. With these sophisticated features implemented in architecture will result in
versatile front end for implementing semantic grid services.
March 23, 2006
M.I.T., Anna Univ, Chennai
57
Appendices
March 23, 2006
M.I.T., Anna Univ, Chennai
58
Courtesy: Global Grid Forum 16
Athens, Greece, February 13-16, 2006
Ontology Framework
uses
Product/
Process
Model
Organisational
Ontology
ref
Business
Process
Ontology
uses
OWL
Service
Ontology
ref
represents
descr
User/Role
Authorisations
OWL
Distributed
run-time
environment
Auth.
Service
RBAC
Ontology-Based
Virtual User Desktop
represents
descr
WSRF
Web
Services
Web
Services
Services
Services
OWL-S
WSDL
run on
uses
Resource
Ontology
OWL
descr
System
Resources
represents
?
grid
March 23, 2006
M.I.T., Anna Univ, Chennai
59
Life without Broker
March 23, 2006
Courtesy: University of Melbourne,
Gridbus Broker Presentation
M.I.T., Anna Univ, Chennai
60
Courtesy: University of Melbourne,
Gridbus Broker Presentation
Life with Broker
?
March 23, 2006
Scheduling
M.I.T., Anna Univ, Chennai
61
Questions
March 23, 2006
M.I.T., Anna Univ, Chennai
62
Thank You
March 23, 2006
M.I.T., Anna Univ, Chennai
63