Grid Application Deployment GAD Kit

1
Welcome to
GAD Kit
(Grid Application Deployment Kit)
Students Ho Quoc Thuan and Dudy Lim Supervisor
Dr. Ong Yew Soon
2
The problem

• How to deploy the computing code?
• How to access Grid resources?
• How to achieve high performance?

3
Objectives

• Assist users in deploying applications on Grid
  environment
• ? reduce time and cost of applications
• Target clusters of computers
• Focus on providing user four main functions
• PROVIDERS of GridRPC services
• COMSUMPTION of existing resources via GridRPC
  services
• METASCHEDULING
• LOAD BALANCING

4
Gridifying process
Service providing
Basic computing code
Grid services
Application program
Execution
Service consuming
5
Gridifying process
Service providing
GridRPC API Implementation -Metascheduler -Load
balancing
Basic Computing Code
Grid Services
Application Program
Application Execution
GAD Kit GUI -Service deployment -Resource
discovery
Service consuming
6
GAD Kits architecture
Users
Resource Monitoring
Service Consuming
Service Providing
GAD Kit
GridRPC API Metascheduler Composite Services
Computing Systems
7
GAD Kits Functions

• Function deployment
• Generates function specifications
• GridRPC Function Description (GFD)
• NetSolve Problem Description File (PDF)
• GridRPC services
• Service consuming
• Generate code from a specified template
• Compile program
• Execute application
• Running application
• Access Grid resource via GridRPC API
• Achieve high performance by Metascheduler

8
Technologies
9
GridRPC components
10
Snapshot
11
Application Airfoil Design

• Computationally expensive task!
• Gridifying time is decreased by using GAD Kit.
• High performance can be achieve by using GAD
  Kits GridRPC API implementation.

12
Experiments

• Analysis of 200 airfoil designs
• Grid resources 3 clusters

13
Experiments
14
Future work

• GAD Kit Portal
• Scientific Problem Solving Environment
• Performance Improvement

15
Q A
16

• Thanks for your attention!