XSEDE14 SciGaP-Apache Airavata Tutorial

SciGaP Tutorial: Developing Science
Gateways using Apache Airavata
XSEDE14, July 14th 2014, Atlanta GA

https://cwiki.apache.org/confluence
/display/AIRAVATA/XSEDE+2014+Tut
orial
Tutorial Material and Links

More Information
• Contact Us:
– marpierc@iu.edu, smarru@iu.edu
– Join dev@airavata.apache.org,
users@airavata.apache.org,
architecture@airavata.apache.org
• Websites:
– SciGaP Project: http://scigap.org
– Apache Airavata: http://airavata.apache.org
– Science Gateway Institute: http://sciencegateways.org

Audience Introductions
Tell us about your work or studies and
your reasons for taking the tutorial
Sign in and take survey

Role Description
Gateway
Developers and
Providers
Use the Airavata API through and
SDK in their favorite programming
language.
Airavata
Developers
Want to change Airavata
components, experiment with
different implementations.
Middleware
Developers
Want to extend Airavata to talk to
their middleware clients.
Resource
Providers
Want to configure Airavata to work
with their middleware.

What You Will Learn
• What a Science Gateway is and does.
• How to build a simple Science Gateway.
• How to integrate existing gateways with
XSEDE and other resources
• What it takes to run a production
gateway.

Gateway
Example
Gateway
Abstractions
APIs and
SDKs
(Hands On)
Gateway
Services
Gateway
Operations
(CIPRES
Lessons)
Tutorial Agenda

What You Can Take Away
• A “test-drive” PHP gateway
– Open source code in GitHub
– We’ll run SciGaP services for you
• OR: Gateway client SDKs for your
programming language of choice
– Integrate with your gateway
– We’ll run SciGaP services for you
• OR: Airavata source code
– Play, use, or develop and contribute back

Example Science Gateways
UltraScan, CIPRES, and Neuroscience
Gateways

A Science Gateway for Biophysical Analysis
Usage statistics (in SUs) for the UltraScan Gateway Jan-Dec 2013
There are over 70 institutions in 18 countries actively using UltraScan/LIMS
Implemented currently on 7 HPC platforms, including one commercial
installation (non-public). In 2013 over 1.5 Million service units provided by
XSEDE, Jülich and UTHSCSA.
Ongoing Projects:
Integration of SAXS/SANS modeling
Integration of Molecular Dynamics (DMD)
Integration of SDSC Gordon Supercomputer with GRAM5 and UNICORE
Integration of Multi-wavelength optics (500-1000 fold higher data density)
Development of Performance Prediction algorithms through datamining
Development of a meta scheduler for mass submissions (of MW data)

UltraScan Goals:
Provide highest possible resolution in the analysis – requires HPC
Offer a flexible approach for multiple optimization methods
Integrate a variety of HPC environments into a uniform submission framework
(Campus resources, XSEDE, International resources)
Must be easy to learn and use – users should not have to be HPC experts
Support a large number of users and analysis instances simultaneously
Support data sharing and collaborations
Easy installation, easy maintenance
Robust and secure multi-user/multi-role framework
Provide check-pointing and easy to understand error messages
Fast turnaround to support serial workflows (model refinement)
A Science Gateway for Biophysical Analysis

Browser
UltraScan Gateway (UTHSCSA)
Airavata Services (Quarry VM Hosting)
Apache Thrift API Server
Apache Thrift Client SDK
Campus
Resources
Middleware Plugins
Trestles
(XSEDE)
Stampede
(XSEDE)
Juelich
HTTPS
TCP/IP
GRAM, GSI-SSH UNICORE, JSDL, BES

UltraScan and Gateway Patterns
• Users create computational packages (experiments)
– Data to be analyzed, analysis parameters and control files.
– Computing resources to use and computing parameters (#
of processors, wall time, memory, etc)
• Experiments are used to launch jobs.
• Running jobs are monitored
• User may access intermediate details
• Results from completed jobs (including failures) need
to be moved to permanent data management system

UltraScan and Gateway Patterns,
Continued
• UltraScan users are not just XSEDE users.
• UltraScan provides access to campus and
international resources as well as XSEDE.
• Gateways are user-centric federating layers
over grids, clouds, and no-middleware
solutions

PHP Gateway for Airavata (PGA)
Demo of and hands on with the Test
Drive Portal
http://test-drive.airavata.org/
PGA is a GTA

Application Field Scenario
AMBER Molecular
Dynamics
Production constant pressure and temperature
(NPT) MD run of alanine dipeptide.
Trinity Bioinformatics RNA-Seq De novo Assembly
GROMACS Molecular
Dynamics
Minimize the energy of a protein from the PDB.
LAMMPS Molecular
Dynamics
Atomistic simulation of friction.
NWChem Ab Initio
Quantum
Chemistry
Optimized geometry and vibrational frequencies of
a water molecule.
Quantum
Espresso
Ab Initio
Quantum
Chemistry
Self-consistent field calculations (that is,
approximate many-body wave function) for
aluminum.
WRF Mesoscale
Weather
Forecasting
US East Coast storm from 2000.
See https://github.com/SciGaP/XSEDE-Job-Scripts/ for more information

http://test-drive.airavata.org/

Browser
PHP Gateway (gw120)
PHP Command Line Tools
(gw104)
Airavata API Server (gw111)
Airavata Components
Resource 1
(Trestles)
Resource 2
(Stampede)
Resource 3
(Big Red II)
(…)
HTTPS
TCP/IP (Thrift)
TCP/IP (Thrift)
Moving parts in the test-drive.
Identity Server (gw26)
HTTPS
SSH, GSI-SSH

Portal Hands On Time
Run all the applications

Getting the Demo Portal Code
• Demo Portal is available from GitHub
– Note: this may be moved to the Airavata Git Repo
– Look for announcements on the Airavata mailing lists
• Go to https://github.com/SciGaP/PHP-Reference-
Gateway
– Create a clone
• Contribute back with pull requests and patches!

A Tour of the API
Examining PHP Clients

Client Objects for Job Submission
Experiment Inputs
(DataObject)
User
Configuration
Computational
Resource
Experiment
Advanced Output
Handling

How to Create an Experiment
• Go to GitHub
– http://s.apache.org/php-clients
• Download the Airavata PHP client zip.
– https://dist.apache.org/repos/dist/dev/airavata/0.13/
RC0/apache-airavata-php-sdk-0.13-SNAPSHOT-
bin.tar.gz
– Look at createExperiment.php
– We also have C++
• If you prefer, we have training accounts on gw104
– Need SSH to access

More API Usage
PHP Sample What it shows
launchExperiment.php How to run an experiment.
getExperimentStatus.php How to check the status.
getExperimentOutput.php How to get the outputs.

Hands On with PHP Command-
Line Clients
Exercises with the clients.
Use the tar on your local machine or
else log into gw104

Behind the API
• The Airavata API server deposits Experiment
information in the Registry.
• API calls Orchestrator to validate and launch jobs.
• Orchestrator schedules the job, uses GFAC to
connect with a specific resource.
• Multiple resources across administrative domains
– Campus, XSEDE, etc
– This requires different access mechanisms

A Few Observations on Successful
Gateways
• Support familiar community applications.
• Make HPC systems easy for new user
communities who need HPC.
• Have champions who build and support the
community.
• Have a lot of common features.

SciGaP Goals: Improve sustainability by converging on a
single set of hosted infrastructure services

Understanding the Airavata API,
Part 1
Creating, executing, and managing
experiments.

Stable,
slowly
changing
Many,
Growing API Calls
Data
Models

Go to
http://airavata.apache.org/docume
ntation/api-docs/0.13/
http://airavata.apache.org/document
ation/ and then navigate

Airavata API and Apache Thrift
• We use Apache Thrift to define the API.
• Advantages of Thrift
– Supports well-defined, typed messages.
– Custom defined errors, exceptions
– Generators for many different programming
languages.
– Some shielding from API versioning problems.
• Downsides of Thrift
– No message headers in TCP/IP, so everything must be
explicitly defined in the API.
– Limits on object-oriented features (polymorphism)

More Information
• ApacheCon 2014 Presentation: “RESTLess
Design with Apache Thrift: Experiences from
Apache Airavata”
• http://www.slideshare.net/smarru/restless-
design-with-apache-thrift-experiances

Understanding the Airavata API,
Part 2
Managing applications, computing
resources, and gateway profiles

Application Catalog Summary
• The Application Catalog includes descriptions
of computing resources and applications.
– It is part of the Registry.
• Call the API to add, delete, or modify a
resource or an application.
• The App Catalog API and data models are in
transition to Thrift.
– You can also still use legacy methods, which fully
featured.

App Catalog
Model
Compute
Resource Model
Application
Interface Model
Application
Deployment
Model
Gateway Profile
Moedel

Data Model Description
Compute
Resource
Everything you need to know about a
particular computing resource.
Application
Interface
Resource independent information
about an application.
Application
Deployment
Compute Resource-specific information
about an application.
Gateway
Resource Profile
Gateway-specific preferences for a
given computing resource.

Using the Application Catalog API
Walk through how to create, modify
applications and computing resources
Live examples with {lammps, gromacs,
espresso,…} on {stampede, trestles, br2}
http://s.apache.org/register-apps

Application Catalog Details
Back to
http://airavata.apache.org/document
ation/api-docs/0.13/

Airavata Under the Hood
More detail about how Airavata
works

Airavata’s Philosophy
• There a lots of ways to build Web interfaces for
Science Gateways.
– By Hand: PHP, Twitter Bootstrap, AngularJS, …
– Turnkey Frameworks: Drupal, Plone, Joomla, …
– Science Gateway Frameworks: the SDSC Workbench,
HUBzero, WS-PGrade
• Gateway developers should concentrate on
building interfaces that serve their community.
• And outsource the general purpose services to
Airavata.

Apache Airavata Components
Component Description
Airavata API Server Thrift-generated server skeletons of the API and data
models; directs traffic to appropriate components
Registry Insert and access application, host machine,
workflow, and provenance data.
Orchestrator Handles experiment execution request validation,
scheduling, and decision making; selects GFAC
instances that can fulfill a given request
Application Factory
Service (GFAC)
Manages the execution and monitoring of an
individual application.
Workflow Interpreter
Service
Execute the workflow on one or more resources.
Messaging System WS-Notification and WS-Eventing compliant
publish/subscribe messaging system for workflow
events

Getting Help
Help Description
Airavata
Developer List
Get in touch with Airavata developers. Good place for asking
general questions. Lots of automated traffic noise. See
apache.airavata.org/community/mailing-lists.html. Mailing list
is public and archived online.
Airavata Jira Official place for posting bugs, tasks, etc. Posts here also go to
the developer mailing list. See
https://issues.apache.org/jira/browse/airavata
Airavata users,
architects lists
Can also post here. Use “users” list for questions. The
architects’ list is for higher level design discussions.
Airavata Wiki
and Website
The Wiki is frequently updated but also may have obsolete
material. The Website should have time-independent
documentation.
What else would you like to see? Google
Hangouts, IRCs, YouTube, social networks?
Improvements to the Website?

Airavata and SciGaP
• SciGaP: Airavata as part of a multi-tenanted
Gateway Platform as a Service
• Goal: We run Airavata so you don’t have to.
• Challenges:
– Centralize system state
– Make Airavata more cloud friendly, elastic

“Apache” Means “Open”
Join the Airavata developer mailing list, get involved,
submit patches, contribute.
Use Give Back

Getting Involved, Contributing Back
• Airavata is open source,
open community software.
• Open Community: you can
contribute back
– Patches, suggestions, wiki
documentation, etc
• We reward contributors
– Committers: write access
to master Git repo
– Project Management
Committee members: full,
binding voting rights

Some Contribution Opportunities
Component Research Opportunities
Registry Better support for Thrift-generated objects; NoSQL and
other backend data stores; fault tolerance
Orchestrator Pluggable scheduling; load balancing and elasticity
GFAC ZooKeeper-like strategies for configuring and managing.
Messenger Investigate AMQP, Kafka, and other newer messaging
systems
Workflow
Interpreter
Alternative workflow processing engines.
Overall Message-based rather than direct CPI calls.
Airavata components (should) expose Component Programming Interfaces
(CPIs) that allow you to switch out implementations. GFAC is also designed to
be pluggable.

Advice on running a gateway
Experiences with CIPRES

Building on the Tutorial Material
• Test-drive gateway will remain up.
• Go over the PHP samples.
• Get other code examples
– JAVA, C++, Python, …
• Contact the Airavata developer list for help.

Get Involved in Apache Airavata
• Join the architecture mailing
list.
• Join the developer list.
• All contributions welcome.
– Apache Software Foundation
owns the IP.
– Submit patches and earn
committership, PMC
membership.

Where Is Airavata 1.0?
• Airavata 1.0 will be
the stable version of
the API.
• Developer
community vote.
– Get involved
– September 2014
target
• Semantic versioning

Get Connected, Get Help
• Get Connected: Join the XSEDE Science
Gateway program community mailing list.
– Email majordoma@xsede.org
– “subscribe gateways”
• Get Help: Get XSEDE extended collaborative
support for your gateway.
• Get Info: https://www.xsede.org/gateways-
overview

More Information
• Contact Us:
– Join
– architecture@airavata.apache.org
– dev@airavata.apache.org,
• Websites:

Using Airavata Directly
Slides on how to run your own
Airavata services

Downloading Airavata
Method Advantages Disadvantages
Clone the trunk through
GitHub or Apache’s Git
repo (mirrored)
Latest check-ins; for
developers.
May break; we (usually)
have frequent releases
that are more stable.
Download official binary
release from Apache
Well tested; should
work out of the box.
Won’t have latest
features
Download official source
from Apache and build
Well tested; you can
make local
modifications.
You still need to build it.
Current Airavata release is 0.11. Version 0.12 will be the first Thrift-based
release, will be made in mid-June. We will return to a regular 6-8 week
release stride.
Airavata as a Service is under development through the NSF-funded SciGaP
Project. No download required.

Building Airavata from Source
• You need Apache Maven.
• Use “mvn clean install”
• Would you like to see other packaging?
–Pre-configured VMs?
–PaaS tools like OpenShift, Apache Stratos?
–Docker?

Running Airavata
• Modify airavata-server.properties (optional)
– airavata/modules/distribution/server/target/apac
he-airavata-server-{version}/bin/ if building from
source.
– Typically required to specify the XSEDE community
credential.
• Start the server using airavata-server.sh.
– Same location as airavata-server.properties

Generating Client SDKs with
Thrift

Generating New Thrift Clients
• airavata/airavata-api/generate-thrift-
files.sh
• Shows how to do this for Java, C++, and
PHP
• Caveat: requires Thrift installation
• See airavata/airavata-api/thrift-interface-
descriptions/ for our Thrift definitions

Writing a PHP Interface to the
Airavata API
This is a set of slides that summarizes
what you need to do to call the API
from PHP.

Anatomy of the PHP Command Lines
• Imports of Thrift Libraries
– require_once $GLOBALS['THRIFT_ROOT'] .
'Transport/TTransport.php';
– An so on
• Make sure the path to the libraries
(‘THRIFT_ROOT’) is correct.
• The libs are generated from the IDL.

Importing Airavata Libraries
• Import of Airavata API
– require_once $GLOBALS['AIRAVATA_ROOT'] .
'API/Airavata.php';
• Import data models
'Model/Workspace/Types.php';
'Model/Workspace/Experiment/Types.php’;
• These are 1-1 mappings from the our Thrift API
definition files.

Initialize the AiravataClient
• AiravataClient has all the API functions.
• You can configure the client manually or
through the AiravataClientFactory object.
– The examples do this manually
• We use Binary protocol and TSocket for wire
protocol and transport
• We recommend giving the client a long
timeout (5 seconds or more).

Make the API Calls
• $airavataclient = new
AiravataClient($protocol);
• $project=$airavataclient->getProject(‘user1’)
• And so on.
• Put these in try/catch blocks and catch
Airavata exceptions.
• Thrift exceptions will also be thrown.
• $transport->close() when done

Plan: Ask the Experts
• We are collaborating with Von Welch’s Center
for Trustworthy Cyberinfrastructure (CTSC)
– Von, Jim Basney, Randy Heiland
• This is an ongoing review
• Open issues are
– Thrift over TCP/IP or Thrift over HTTPS?
– Verifying user identity, gateway identity, and user
roles.
• Airavata architecture mailing lists

AuthN, AuthZ, and the Client API
• Airavata assumes a trust relationship with
gateways.
– Best practice: use SSL, mutual authentication.
– Firewall rules: clients come from known IPs.
• The gateway makes AuthN and AuthZ decisions
about its users.
• Airavata trusts the gateway’s assertions
• For command line examples, you are the gateway,
not the user

Auth{Z,N} Problems and Solutions
• Client-Airavata security is turned off for demo
• The approach does not scale.
– Clients may not have well-known IP addresses
(Desktop clients for example).
– Airavata service operators serving multiple
gateways have to manually configure
• Not self-service
• We are evaluating a developer key scheme.
– Modeled after Evernote client SDKs

More on Airavata Security
• “A Credential Store for Multi-tenant Science
Gateways”, Kanewala, Thejaka Amilia; Marru,
Suresh; Basney, Jim; Pierce, Marlon
– Proceedings of CCGrid 2014 (this conference!)
• http://hdl.handle.net/2022/17379

Extending Airavata through
Plugins
See
https://cwiki.apache.org/confluence/di
splay/AIRAVATA/Developing+GFAC+Plug
ins

Airavata and Middleware
• Airavata’s philosophy is that gateways need ad-hoc access
to all kinds of resources that are not centrally managed.
– Grids, clouds, campus clusters
• These resources are accessed through many different
mechanisms.
– SSH, JSDL/BES, HTCondor, Globus, …
• These resources may also run job management systems
– PBS, SLURM
– Workflow engines
– Parameter sweep tools
– Hadoop and related tools
• So we’ll describe how to extend Airavata’s GFAC

GFAC Plugins: Providers and Handlers
• Providers: clients to common middleware.
– Airavata comes with SSH, Globus GRAM, BES/JSDL,
“localhost”, EC2 and other providers.
• Handlers: customize providers for specific resources.
– Set-up, stage-in, stage-out, and tear-down operations.
• A given GFAC invocation involves 1 Provider and 0 or
more Handlers.
• Gfac-core invokes Handlers in sequence before (“in-
handlers”) and after (“out-handlers”) the provider.

A Simple Example
• Let’s write two handlers and a localhost
provider.
• Prerequisite: you have Airavata installed
• InHandler: emails user
• Provider: runs “echo” via local execution
• OutHandler: emails user

Configure the InHandler
• Place this in gfac-config.xml.
• Note gfac-core lets you provide arbitrary properties.
<Handler
class="org.apache.airavata.gfac.local.handler.InputEmailHandler
">
<property name="username" value="gmail-
username@gmail.com"/>
<property name="password" value="gmail-password-xxx"/>
<property name="mail.smtp.auth” value="true"/>
<property name="mail.smtp.starttls.enable" value="true"/>
<property name="mail.smtp.host” value="smtp.gmail.com"/>
<property name="mail.smtp.port" value="587"/>
</Handler>

Configure the OutHandler
These are read by the following code. initProperties is a
GFacHandler interface method.
<Handler class="org.apache.airavata.gfac.local.handler.OutputEmailHandler">
<property name="username" value="gmail-username@gmail.com"/>
<property name="password" value="gmail-password-xxx"/>
<property name="mail.smtp.auth” value="true"/>
<property name="mail.smtp.starttls.enable" value="true"/>
<property name="mail.smtp.host” value="smtp.gmail.com"/>
<property name="mail.smtp.port" value="587"/>
</Handler>
private Properties props;
public void initProperties(Properties properties) throws
GFacHandlerException {
props = properties;
}

Write InputEmailHandler
public void invoke(JobExecutionContext jec)
throws GFacHandlerException {
Map<String, Object> parameters =
jec.getInMessageContext().getParameters();
StringBuffer buffer = new StringBuffer();
for (String input : parameters.keySet()) {
ActualParameter inParam =
(ActualParameter) parameters.get(input);
String paramDataType =
inParam.getType().getType().toString();
if ("String".equals(paramDataType)) {
String stringPrm = ((StringParameterType)
inParam.getType()).getValue();
buffer.append("Input Name:” + input + ”Input Value: “
+stringPrm + "n");
}
}
message.setText(buffer.toString());
}
Bold font indicates Airavata-specific code.

Write the Output Handler
• In our example, it is nearly identical.

Write a “Localhost” Provider
• This runs a command on the same machine as
the GFAC instance.
• We already have one in Airavata’s code base.

gfac-config.xml
<Provider class="org.apache.airavata.gfac.local.provider.impl.LocalProvider"
host="org.apache.airavata.schemas.gfac.impl.HostDescriptionTypeImpl”>
<InHandlers>
<Handler
class="org.apache.airavata.gfac.local.handler.LocalDirectorySetupHandler"/>
<Handler
class="org.apache.airavata.gfac.local.handler.InputEmailHandler">
<!– Parameters shown in previous example -->
</Handler>
</InHandlers>
<OutHandlers>
<Handler class="org.apache.airavata.gfac.local.handler.OutputEmailHandler">
<!– Parameters shown in previous example 
</Handler>
</OutHandlers>
</Provider>

More Information
• Contact Us:
– Join dev@airavata.apache.org,
architecture@airavata.apache.org
• Websites:

Quick Thanks to Devs and Testers
• Eroma Abeysinghe
• Lahiru Gunathilake
• Yu Ma
• Raminder Singh
• Saminda Wijeratne
• Chathuri Wimalasena
• Sachith Withana
• Some may join via Google Hangout

XSEDE Science Gateways
Program Overview
Quick overview, plus cookbook
[I think this must go into
supplemental material]

Why Use Airavata for Gateways
• Open community, not just open source.
– Become a stakeholder, have a say.
• Gateway-centric overlay federation.
– Not tied to specific infrastructure.
• No additional middleware required.
– Campus clusters
– Let your sysadmins know
• Stable funding
• Lots of room for distributed computing research
– Pluggable services

Why Use SciGaP?
• Outsource common services, concentrate on
community building.
• Use same services as UltraScan, CIPRES, NSG
• Open deployments.
– No secrets about how we run services
• Lots of eyes on the services.

Upcoming Stuff
Things on the way for Airavata,
SciGaP

What’s On the Way?
• Airavata 1.0 targeted for September
– 0.13, 0.14 will follow in the next 2 months.
• OpenStack service hosting
– Rackspace has donated $2000/month usage.
• API production usage by UltraScan
– Have been using Airavata for last 9 months
• CIPRES integration
– NSG has some advanced requirements
• Continued improvement to PGA Portal
– Java, Python, and other GTAs welcome
• More fun with Identity Server and API
• Desktop SDKs

What is Analytical Ultracentrifugation (AUC)?
A technique to separate macromolecules dissolved in solution in a centrifugal force
field. AUC can “watch” the molecules while they are separating and identify and
differentiate them by their hydrodynamic properties and partial concentration. AUC
is a “first Principle” method, which does NOT require standards.
How does AUC work?
By applying a very large centrifugal force, molecules are separated based on their
molar mass and their shape. The molecules are observed using different optical
systems that detect different properties of the molecules, such as refractive index,
UV or visible absorbance, or fluorescent light emission.
What molecules can be studied?
Virtually any molecule, colloid or particle that can be dissolved in a liquid can be
measured by AUC, as long as it does not sediment by gravity alone. The molecule
or particle can be as small as a salt ion, or as large as an entire virus particle.
The ultracentrifuge can spin at 60,000 rpm, generating forces close to 300,000 x g.
Even small molecules like salt ions will sediment in such a force.
AUC Background

Simplified access to
phylogenetics codes
on powerful XSEDE
resources
The CIPRES Gateway has
allowed 8600+ users to
access 45.7 million core
hours of XSEDE time for
260,000 jobs, resulting in
700+ publications.

Easy user interface – providing easy model
upload, running of codes
Complete set of neuronal simulation tools –
NEURON, GENESIS, Brian, NEST, PyNN –
widely used by computational neuroscientists
Ability to easily get to the results, download
results
Democratize computational
neuroscience
The NSG is a simple and secure online science portal that provides access to computational
neuroscience codes on XSEDE HPC resources
http://www.nsgportal.org
Amit Majumdar (PI), Maryann Martone (Co-PI), Subha Sivagnanm (Sr. Personnel)
Kenneth Yoshimoto (Sr. Personnel), Anita Bandrowski (Sr. Personnel), Vadim Astakhov UCSD
Ted Carnevale (PI), Yale School of Medicine
NSF Awards: ABI #1146949; ABI #1146830

The UltraScan science gateway supports
high performance computing analysis of
biophysics experiments using XSEDE,
Juelich, and campus clusters.
Desktop analysis tools
are integrated with the
Web portal.
Launch analysis
and monitor
through a
browser
We can help you build
gateways for your lab or
facility.

XSEDE14 SciGaP-Apache Airavata Tutorial

Recommended

More Related Content

What's hot (20)

Viewers also liked (6)

Similar to XSEDE14 SciGaP-Apache Airavata Tutorial (20)

More from marpierc (14)

Recently uploaded (20)

XSEDE14 SciGaP-Apache Airavata Tutorial