Information Power Grid IPG

1
Information Power Grid (IPG)

Overview Anthony Lisotta lisotta_at_nas.nasa.gov
Computer Sciences Corporation NASA Ames May 2,
2003
2
Outline

• What are Grids?
• Current State of Information Power Grid (IPG)
• Future Directions
• Wrap-up

3
Grids

• A Grid is a system that
• coordinates resources that are not subject to
  centralized control
• using standard, open, general-purpose protocols
  and interfaces
• to deliver nontrivial qualities of service.
• Grid software is middleware

This is a Grid Enabled Infrastructure
Grid Middleware
- Ian Foster, July 2002
4
Normal Grid Characteristics

• An underlying security infrastructure such as the
  Grid Security Infrastructure (GSI), which is
  based on public key technology
• Protection for at least authentication
  information as it flows from resource to resource
• Readily accessible information about the
  resources on the Grid via a single mechanism, the
  Grid Information Service (GIS)
• Single sign-on
• A seamless processing environment
• An infrastructure that is scalable to a large
  number of resources
• The ability for the grid to cross administrative
  boundaries

5
The Grid Today

• Common Middleware
• Abstracts independent, hardware, software, user
  ids, into a service layer with defined APIs
• Provides comprehensive security,
• Allows for site autonomy
• Provides a common infrastructure based on
  middleware

User
Application
The underlying infrastructure is abstracted into
defined APIs thereby simplifying developer and
user access to resources, however, this layer is
not intelligent.
Grid Middleware
Infrastructure
Network
Site A
Site B
6
Outline

• What are Grids?
• Current State of Information Power Grid (IPG)
• Future Directions
• Wrap-up

7
IPG Vision

• To make the practice of large-scale science and
  engineering, as well as other widely distributed,
  data intensive NASA activities, much more
  effective than it is today.
• Grid technology is the foundation to making this
  vision a success

8
IPG Locations
Goddard Space Flight Center
Ames Research Center
Glenn Research Center
Jet Propulsion Labs
Langley Research Center
9
Current IPG Services
Launch Pad Web Portal
10
Outline

• What are Grids?
• Current State of Information Power Grid (IPG)
• Future Directions
• Wrap-up

11
Future Directions

• Focus on Managing Complex, Distributed Data
  (MCDD)
• Intelligently manage data in a grid
• IPG Advanced Services
• easier to use
• deliver value added services
• easier to maintain

12
MCDD

• Goal Intelligently manage data in a grid
• NASA data is inherently distributed e.g., various
  Earth science archives.
• Important focus of IPG
• Access to files
• Initial use of grid-enabled Storage Resource
  Broker
• Data staging and replica management building on
  grid community research
• Need grid support for file metadata
• NASA data can be on
• Disk-resident data pools
• Tertiary storage data archives
• Will build on considerable data grid work from
  the international grid community

13
MCDD - Objectives

• Identify key data services abstractions that will
  address a wide range of large-scale, federated,
  distributed NASA data problems
• This potentially has considerable utility in many
  NASA areas as it essentially seeks to abstract,
  regularize, and componentize the general
  operations that are performed on scientific data.
• Analyze these abstractions in the context of
  services oriented Grids and determine an
  implementation and prototyping strategy
• Identify existing work in this area and design an
  implementations approach that maximizes leverage
  from current work
• Demonstrate the prototype services within the
  context of some existing NASA applications

14
IPG Advanced Services

• Goal Location-independent computing
• Intelligent resource selection
• User-specified preferences and requirements
• Resource characteristics, access, allocations,
  cost, load, performance predictions
• Create system-dependent job from
  system-independent job
• Given a system (or systems) to use
• Pick directories, set paths and environment,
  specify which executables and libraries to move,
  specify which data files to move
• Reliable execution of system-dependent job
• Application execution file management
• Dynamic access and accounting
• Run applications without a permanent local
  account
• Charging resource use to remote accounting
  systems
• Resource pricing
• Workflow management
• Specify and execute workflows

15
Concept Architecture
User Program
Higher-level Services Version 1 in use
Resource Broker Client
Higher-level Services In Development Now
Resource Broker Job

Transitioned Services
Resource Broker
Portability Manager Client
Resource Selector Client
Job Manager Client
resource requirements and preferences
Portability Manager Job
resource sets
Job Manager Job
Portability Manager
Resource Selector
IPG Job Manager
Dynamic Access Client
Performance Prediction Clients
GIS Client
GIS Client
Data Grid Client
Globus GRAM Client
Grid FTP Client
resource info
GRAM Job
3rd party copy
allocations
resource and service status
access allowed
Globus GRAM
Grid FTP
prices
Access Client
Allocation Manager
Testing and Monitoring
Resource Pricing
Access Policies
PBS
Allocation Client
access allowed map user to userid
deduct from allocation
16
Outline

• What are Grids?
• Current State of Information Power Grid (IPG)
• Future Directions
• Wrap-up

17
Summary

• Transition support of IPG to operations
• Testbed in parallel using GT3?
• Formalize NASA grid organizations
• Continue to make the Grid easier to use
• Research complex distributed
• data management

Testbed
Operations

• Shift primary work to increase usability
• Globus stabilized
• Attempt inter-grid processing
• Begin development of initial services

Stability level
Transition Period

• Persistent Grid
• Informal Organization
• Lots of Teamwork

• Testbed
• Learning Grids
• Organizing Players

1999
2000
2001
2002
2003
2004