The Cactus Portal

1
The Cactus Portal

A Case Study in Grid Portal Development
Michael Paul Russell Dept of Computer Science The
University of Chicago russell_at_cs.uchicago.edu


2
Grid Portals Defined
A short description
Grid Portals are a class of www application
servers that provide a secure online environment
for gathering information about Grid services and
resources as well as provide tools for utilizing
these Grid services and resources to perform
useful tasks.
3
Example Portals
Just a few examples

• The Astrophysics Simulation Collaboration Portal
• A Cactus Portal Application
• https//www.ascportal.org/asc/
• SDSC Hot Pages
• A Web-based Interface to NPACI Computing
  Resources
• http//hotpage.paci.org/
• WebSubmit
• A Web-based Interface to High Peformance
  Computing Resources
• http//www.itl.nist.gov/div895/sasg/websubmit
  /websubmit.html

4
Grid Portal Properties
What Grid Portals have in common

• Highly-available multi-user www application
  servers.
• Accessible with common www thin-client
  technology.
• Provide secure remote access to Grid resources
  services.
• Provide consistent interfaces to diverse sets of
  services.
• Provide a comprehensive view of Grid resources
  services.
• Often serve the needs of a specific community of
  Grid users.

5
Additional Functions
Grid Portals Grid-Enable

• Secure online access to remote system tools.
• A platform for distributing and running software
  on the Grid.
• A platform for publishing information about
  resources.

6
Grid Portal Provisions
Typical online tools
Grid Security Management
Tools for secure access to stored user proxies
for authenticating to multiple Grid sites
services.
Information Queries
Tools for defining queries to determine resource
availability status, dataset location,
monitoring jobs, etc.
Grid Resource Management
Tools for configuring user resources such as
specifying appropriate location for mass
secondary data storage on remote file systems.
Grid Software Management
Tools for checking out, compiling, deploying and
configuring software on remote sites, running and
monitoring software during execution.
Grid Data Management
Tools for common file management on remote sites
tools for locating and accessing large shared
datasets, etc.
7
Architecture Overview
Grid Portal Site boundary
Secure HTTPD
Persistent storage systems
8
Key Properties
Architectural points

• Built upon common www application client-server
  technology.
• Leverage Grid technology for access to data and
  computing.
• Utilizes Grid Security Infrastructure (GSI)
  software.

9
Application Services
Commonly accessed services
10
Grid Architecture
Comp
Comp
The Application Grid is the set of services
available on sites distributed throughout the
Grid. Grid Portals utilize these services as well
as provide users with this view of the Grid.
Data
GIS
WWW Client
Data
Data
WWW Portal
WWW Client
GIS
Comp
WWW Client
Soft
Soft
GIS
GIIS
Proxy
MyProxy Service
Portal GIIS
11
Software Requirements
Grid Portal Client Software

• Internet browser (Netscape 4.0 or Internet
  Explore 4.0)

GSI Client software

• MyProxy Client Software (http//www.globus.org/co
  g)

12
A Case Study
The Cactus Portal
The Cactus Portal is an online environment for
building, distributing, running and monitoring
Cactus applications. More importantly, it
provides tools for helping users to effectively
utilize the entire set of Grid computational and
data management services to which they have
access.
13
Cactus Defined
What is Cactus?
Cactus is a freely available, modular, portable
and manageable environment for collaboratively
developing high- performance multidimensional
simulations
14
Cactus Portal Provisions
Look familiar?
Grid Security Management
Tools for secure access to stored user proxies
for authenticating to Grid sites services.
Information Queries
Tools for defining queries to determine resource
availability status, dataset location,
monitoring jobs, etc.
Grid Resource Management
Tools for configuring user resources such as
specifying appropriate location for mass
secondary data storage on remote file systems.
Grid Software Management
Tools for checking out, compiling, deploying and
configuring software on remote sites, running and
monitoring software during execution.
Grid Data Management
Tools for common file management on remote sites
tools for locating and accessing large shared
datasets, etc.
15
Cactus Portal Architecture
Grid Portal Site boundary
Stronghold
Client code downloaded onto http client
DHTML
16
Application Components
Cactus Grid Components
17
Cactus Portal Grid
Each user selects one site on the Grid to be her
personal repository from which software and data
is deployed onto other sites on the Grid. This
allows a user to interact with the repository
outside of the the Cactus Portal and potentially
to use this same repository in conjunction with
other Grid Portals
Comp
Comp
Data
GIS
www.ascportal.org
Comp
Comp
Portal
Soft
GIS
Data
GIS
Comp
Data
Personal repository
Data
Data
myproxy.ascportal.org
giis.ascportal.org
cvs.cactuscode.org
GIS
Comp
GIS
Proxy
Soft
The Cactus Portal uses a publicly shared Cactus
Software Repository in conjunction with allowing
users to import custom libraries for distribution
onto Grid sites.
MyProxy Service
Portal GIIS
Cactus CVS
18
Future Goals
Cactus Portal Goals

• Develop generic and robust source code
  management tools.
• Develop generic and robust grid data management
  tools.
• Allow users to share access to data and
  software.
• Develop robust online administration tools.
• Develop reusable Grid Portal software libraries.
  

19
Research Ideas
Integrated environments

• Build an online Cactus integrated development
  environment (IDE).

Edit
Compile
Debug
Run
Version

• Develop online scripting environment for
  customizing tasks.

for (ii 0 ii appl.setParameterFile(parfii.getUrl())
appl.run()