Hongsuda Tangmunarunkit

1
Grid Infrastructure for SCEC

• Hongsuda Tangmunarunkit
• Computer Scientist
• Center for Grid Technologies
• USC/Information Sciences Institute
• Annual Meeting, 09/09/2003

2
Community Modeling Environment
KNOWLEDGE REPRESENTATION REASONING Knowledge
Server Knowledge base access, Inference Translatio
n Services Syntactic semantic translation
Knowledge Base
Ontologies Curated taxonomies, Relations
constraints
Pathway Models Pathway templates, Models of
simulation codes
DIGITAL LIBRARIES Navigation Queries Versioning
, Topic maps Mediated Collections Federated acces
s
KNOWLEDGE ACQUISITION Acquisition
Interfaces Dialog planning, Pathway
construction strategies Pathway Assembly Template
instantiation, Resource selection, Constraint
checking
Code Repositories
FSM
RDM
AWM
SRM
Users
Data Collections
Data Simulation Products
GRID Pathway Execution Policy, Data ingest,
Repository access Grid Services Compute storage
management, Security
Pathway Instantiations
Storage
Computing
3
Outline

• Introduction to Grids
• Current status of SCEC Grid Testbed
• Pathway Applications SCEC Grid
• Ongoing and Future Work

4
Why Grids?
Enabling technology for resource sharing
coordinated problem solving in dynamic,
multi-institutional virtual organizations
5
Grid Infrastructure Services

• Grid vision Provide an integrated, on-demand,
  ubiquitous access to computing, data, and
  services
• Grid technologies are far from complete or
  perfect, but are out there and evolving rapidly!
• Fundamental Grid Services for SCEC
• Authentication and authorization
• Single sign-on
• Resource discovery and monitoring
• Reliable remote service invocation
• Unified interface to various computing resources
• High-performance remote data access
• Data Discovery

6
SCEC Grid Testbed
SCEC
SDSC
USC
PSC
TeraGrid
ISI
TeraGrid facilities will be added to the SCEC
testbed in the future.
7
Execution Environment
gravity.usc.edu
Job Submission Manager
(3) Jobs are submitted to remote servers
(4) Submit request for file transfer
8
Pathway1 Seismic Hazard Map Calculation

• A map generation requires 104-106 Seismic Hazard
  Analysis (SHA) calculations
• Distribute sets of SHA (with different
  parameters) on multiple resources

Pathway1 - Run Job-1 SHA w/ params-1 on
condor.usc.edu - ... - Run Job-n SHA w/
params-n on condor.usc.edu - Transfer outputs to
gravity.usc.edu
35 mins instead of 7.5 hours for n11000!
Job Submission Manager
Submit Pathway1
SCEC Grid Resource
Generate map
condor.usc.edu
Job-1
Job-2
Job-m
9
Visualization tools

• Pathway2 Workflow (submitted from
  gravity.usc.edu)
• Transfer Problem definition to almaak.usc.edu
• Run CVM request on almaak.usc.edu
• Transfer CVM output to hpc.usc.edu
• Run Olsen simulation on horizon.sdsc.edu
• Transfer outputs to pinto.isi.edu
• Run visualization tool on pinto.isi.edu

Physics-based Solver
Problem definition
CVM Code/Data
Geometric and physical properties and max freq
Olsens code
CVM (CMU)
Outputs
Input mesh
Outputs
Storage System
Source definition (e.g. fault locations)
almaak.usc.edu
horizon.sdsc.edu
pinto.isi.edu
gravity.usc.edu
10
Ongoing Future Work

• Ontology-based Resource Matchmaker (with Stefan
  Decker, ISI)

11
Ongoing Future Work

• Ontology-based Resource Matchmaker (with Stefan
  Decker, ISI)
• Match Grid resources to users application
  requirements
• Current development Tangmunarunkit_Decker_Kesselm
  an_ISWC03
• Grid Resource Ontology
• Matchmaker as a Grid service
• Future Plan
• Pathway Application Ontology
• Interface/Integrate with workflow planner and CAT
  (pathway composition)
• Grid-based Visualization Framework
• SCEC Pathway simulations generate large amount of
  4D data
• A standalone workstation cant handle
  visualization of large data set
• Our solution leverage Grid resources to achieve
  scalable visualization
• Expand the SCEC Grid infrastructure to include
• TeraGrid facilities
• SCEC TB Storage System

12
Grid-based Visualization Framework
Grid Resources
Visualization Machine
13
13.6 TF Linux TeraGridComputing at 40 Gb/s
14
Grid Infrastructure Services

• Grid Vision provide an integrated, on-demand,
  ubiquitous access to computing, data, and
  services
• Grid technologies are far from complete or
  perfect, but are out there and evolving rapidly!
• Fundamental Grid Services
• Authentication and authorization
• Single sign-on
• Resource discovery and monitoring
• Reliable remote service invocation
• Unified interface to various computing resources
• High-performance remote data access
• Data discovery