XCAT Science Portal Status & Future Work July 15, 2002 Shava Smallen Extreme! Computing Laboratory Indiana University.

XCAT Science Portal

Status & Future Work

July 15, 2002

Shava SmallenExtreme! Computing Laboratory

Indiana University

Outline1. Current Status:

a. Architectureb. Case Study: Grappa

i. US-ATLASii. GriPhyN

2. Future Work:a. Recent developments

i. Web portal frameworks & portletsii. Application Factory Web Services (AFWS)iii. Web Services for Remote Portals (WSRP)

b. xportlets

Motivation

• Lots of Grid tools & services

• Requires a good deal of expertise to– develop an

application– use an application

(more than what many users want to know)• Grid portal:

– web portal by which users can access Grid tools & services

– common approach

XCAT Science Portal screenshot

XCAT Science Portal

• Grid portal framework for building personal science portals

• Active notebook (execution management)– HTML pages to describe the features of the

notebook and how to use it– HTML forms which can be used to launch

parameterizable scripts – Results archived - parameters stored in a

sub-notebook

(previously known as Active Notebook project )

Scripts

• Very flexible• Jython - access to Java

classes– Globus Java CoG kit– XCAT – XMESSAGES

• Not every user has to write scripts

• Notebooks can be shared among users– Import/export capability

jythonscript

app1

portal

Grid

Invoke jython interpreter

Launch app1

parameters

Portal Web Server(tomcat server + java servlets)

JythonIntepreter

NotebookDatabase

GSI Authentication

XCAT Science Portal Architecture

User’s Web Browser

Grid

Case Study: Grappa

• Grappa: Grid access portal for physics applications– provide a Grid portal front-end by which

physicists can manage computation and data

• Joint work with Rob Gardner’s group (IU Physics & U. of Chicago Physics)– Part of the ATLAS collaboration (US-

ATLAS)

• Subproject of Grid Physics Network (GriPhyN)

ATLAS • Detector for the Large

Hadron Collider (LHC) at CERN

• 2000 physicists, 150+ universities & laboratories, 34 countries• Expected to be on-line in 2006

• Why the Grid?– Raw data: 2 PB/yr – Analysis: 8 PB/yr

• Participant in several Grid projects: LCG, GridPP, INFN Grid, DataGrid, PPDG, GriPhyN/IVDGL

ATLAS detector

XCAT Science Portal & ATLAS

• Athena = ATLAS software framework • Athena Notebook Prototype/Demo

– ATLAS Software Week (March)– Allow user to submit sequence of Athena job

options files to US-ATLAS testbed resources– File staging, remote job option file editing,

basic monitoring, ad-hoc viz– Packaged with Pacman – installed with single

command

• Distributed to US-ATLAS testbed this month– Interactive jobs– Production – scripting interface to launch and

portal to monitor

GriPhyN: Grid Physics Network

• Targets data-intensive applications – ATLAS, CMS, LIGO, and SDSS

• Provide a set of tools for collaborative data analysis– 90% of data is derived– Virtual data - transparency with respect

to location and materialization• Data replicated & distributed – transfer or

recompute

Virtual Data Browser

• Search virtual data• Publish virtual

data: create transformations and derivations

Application

Planner

Executor

Catalog Services

Info Services

Policy/Security

Monitoring

Repl. Mgmt.

Reliable TransferService

Compute Resource Storage Resource

DAG

DAG

• User Profile – bookmarks, credential mgmt, resource mgmt, etc.

Provide functionality analogous to that of a web browser

Outline1. Current Status:

a. Architectureb. Case Study: Grappa

i. US-ATLASii. GriPhyN

2. Future Work:a. Recent developments

i. Web portal frameworks & portletsii. Application Factory Web Services (AFWS)iii. Web Services for Remote Portals (WSRP)

b. xportlets

Web Portal Frameworks

• Separation of appearance & functionality– User customization of

appearance (multi-user support)

– Functionality packaged into portlets

• Examples: IBM Websphere, Apache Jetspeed, etc.

• Many Grid portal efforts • Encourages interoperability

Jetspeed screenshot

Application Factories

• Applications are difficult to deploy– Libraries, licensing, environment setup, etc.– E.g., Athena: 10 GB library, Redhat 6.1

• Web services approach– XCAT implementation– App. factory web service (AFWS) is a stateless,

persistent service and launches one instance of the application

– Distributed component applications• Both individual components and composite

application are web services

WSRP Specification

• Web Services for Remote Portals• Generic proxy portlet

– Talks to any remote WSRP-enabled web service

• Web service handles content• WSRP web service registry• Advantage?

– No installation required; just contact WSRP web service

Putting It Together

AFWSRegistry

app1

Application coordinator

Application instance

(1)

(2)

(3)(4)

(5)

(6)(7)

GridPortal

AFWSproxy portlet

Notebook(s)

AFWS

jython scripts

Archiving(sub-notebooks)

Portal

Grid

AFWS = Application Factory Web Service

xportlets

• Use Jetspeed as web portal framework– Open source from apache

• xportlets – bundle of portlets– AFWS proxy portlet– MyProxy portlet– Xmessages pull– GridFTP File Browser

• Scripting API to create Application Factories – Different kinds – XCAT, notebooks, etc.

Summary

• Current XCAT Science Portal architecture– Active notebooks, scripts

• Case Study: Grappa– US-ATLAS– GriPhyN

• Redesign to portlet & web services architecture– Jetspeed takes care of presentation– Interoperability – pluggable into any portlet-

enabled portal framework– Easy deployment – web services & generic

proxy

More Information

• Extreme! Computing Laboratory:http://www.extreme.indiana.edu

• XCAT Science Portal:http://www.extreme.indiana.edu/xcatsp

• Grappa:http://iuatlas.physics.indiana.edu/grappa

• xportlets: (coming soon…)http://www.extreme.indiana.edu/xportlets

• Email:[email protected]

Putting It Together

AFWSRegistry

AFWS

Notebook(s)

Componentinstance

Component instance

componentInstance

Application coordinator

Application instance

(1)

(2)

(3)(4)

(5)

(6)(7)

jython scripts

GridPortal

AFWSproxy portlet

Archiving(sub-notebooks)

Portal

Grid

AFWS = Application Factory Web Service