1. Grid Rapid Application Virtualization Interface (gRAVI) - Service Oriented Science
Ravi K Madduri, Argonne National Laboratory/ University of Chicago
Joshua Boverhof, Lawrence Berkeley Laboratory
Kyle Chard, University of Wellington, NZ
Dinanath Sulakhe, University of Chicago
Cem Onyukusel, CMU
Ian Foster, University of Chicago/ANL

2. Agenda What is Service Oriented Science ? Software as a service
Examples from industry (google, salesforce.com etc)
Examples of SoS from Scientific communities
Why are we excited about this ?
How this helps Science
Sum of parts greater
Sustainable infrastructure
Virtualization
Different aspects of SoS
Authoring
Deploying and Securing
Discovery
Provisioning on demand
Composition
How does it help you ?
End users perspective – Brian’s talk
Bio workflow
Conclusions and Further resources
Q & A

3. Examples of SaaS from Industry
Salesforce.com
Google Calendar, Google docs, Google spell check etc..
Amazon S3, EC2
Flickr
Facebook
Microsoft Live
And on and on….

4. Service-Oriented Science
People create services (data or functions) …
which I discover (& decide whether to use) …
& compose to create a new function ...
& then publish as a new service.
 I find “someone else” to host services, so I don’t have to become an expert in operating services & computers!
 I hope that this “someone else” can manage security, reliability, scalability, … ! !
“Service-Oriented Science”, Science, 2005

5. “Service-Oriented Science”, Science, 2005
Creating Services
People create services (data, code, instr.) …
which I discover (& decide whether to use) …
& compose to create a new function ...
& then publish as a new service.
 I find “someone else” to host services, so I don’t have to become an expert in operating services & computers!
 I hope that this “someone else” can manage security, reliability, scalability, … ! !
“Service-Oriented Science”, Science, 2005

6. Creating Services Introduce + gRAVI
Shannon Hastings
Scott Oster
David Ervin
Stephen Langella
Joshua Boverhof
Kyle Chard
Ravi Madduri

7. Introduce Overview
A framework which enables fast and easy creation of Globus based grid services
Provide easy to use graphical service authoring tool.
Hide all “grid-ness” from the developer
Utilize best practice layered grid service architecture
Integration with other core grid services and architecture components
GAARDS Security Infrastructure (Dorian, GridGrouper, CSM)
Globus Index Service
Global Model Exchange (GME)
Cancer Data Standards Repository
Extension Framework for integrating with other architecture components

8. Inside the Introduce created service
Services have many moving and configurable parts which support features such as:
Advertisement
Discovery
Invocation
Security (Authentication/Authorization)
Stateful Resources
The Introduce Toolkit can keep all these features in sync as the developer creates and modifies the grid service

9. Introduce Features Supports modification of operations
Adding operations
Removing Operations
Updating Operations
Importing Operations
Graphical Configuration
Advertisement
Security
Service Metadata Specification
Service Metadata Editing
Service Configuration Properties
Auto Generates Code for Service
Auto generates a client API for service.
Graphical Deployment of Service
Globus
Tomcat
JBoss

10. gRAVI Grid Remote Application Virtualization Interface Appln Service
Create
Grid Remote Application Virtualization Interface
Builds on Introduce
Define service
Create skeleton
Discover types
Add operations
Configure security
Wrap arbitrary executables
Introduce
Store
Repository Service
Index service
Advertize
Discover
Transfer
GAR
Invoke;
get results
Container
Deploy

11. “Service-Oriented Science”, Science, 2005
Discovering Services
People create services (data or functions) …
which I discover (& decide whether to use) …
& compose to create a new function ...
& then publish as a new service.
 I find “someone else” to host services, so I don’t have to become an expert in operating services & computers!
 I hope that this “someone else” can manage security, reliability, scalability, … ! !
“Service-Oriented Science”, Science, 2005

12. Discovering Services WS-MDS+ Taverna
Laura Pearlman
Mike Darcy
myGrid Team

13. Discovering Services Assume success  Billions of services Semantics
Permissions
Reputation
 Billions of services
 The ultimate arbiter?
 Types, ontologies
 Can I use it?
Use funnel animation? A B

14. Discovery (1): Registries
Globus

15. “Service-Oriented Science”, Science, 2005
Composing Services
People create services (data or functions) …
which I discover (& decide whether to use) …
& compose to create a new function ...
& then publish as a new service.
 I find “someone else” to host services, so I don’t have to become an expert in operating services & computers!
 I hope that this “someone else” can manage security, reliability, scalability, … ! !
“Service-Oriented Science”, Science, 2005

16. Taverna A sample caGrid workflow
The GT4 plug-in support semantic based service query. Users can input multiple (up to three, in current scenario three is enough. We can add more in our program upon request) service query criteria and input the corresponding value.
We can combine multiple criteria. The initial GUI only shows one query criteria, but more can be added by clicking “Add Service Query” button. For example, we can query the caGrid services whose “Research Center” name is “Ohio State University”, with Service Name “DICOMDataService”, and has operation “PullOp”.
caGrid Scavenger with semantic/metadata
based caGrid service query

17. “Service-Oriented Science”, Science, 2005
Hosting Services
People create services (data or functions) …
which I discover (& decide whether to use) …
& compose to create a new function ...
& then publish as a new service.
 I find “someone else” to host services, so I don’t have to become an expert in operating services & computers!
 I hope that this “someone else” can manage security, reliability, scalability, … ! !
“Service-Oriented Science”, Science, 2005

18. Provisioning Services WS-GRAM + VWS
Martin Feller
Stuart Martin
Kate Keahey
Tim Freeman
Joshua Boverhof

19. Provisioning using WS-GRAM
gRAVI uses JSDL for Application Description
Generates a method on the generated service called <appName>GRAM
Generates implementation that creates a GramJob from Application Description
Uses the bootstrapped community credential to run the application as a grid job
Used widely in realizing the usecases from caBIG community

20. Using VWS/Cloud Computing
gRAVI service
Wrap the application as Service
Create the GAR and put in repository
Provision resources
Use clouds (VMM)
start service
Transfer gar, deploy
Index Service
Grid service registers itself
stratus
nimbus
provision
Create VM
Register
service
Transfer,
deploy
Repository Service
Index service
portal
GAR
discovery 20

21. Cloud Computing 21 21

22. SoS Deployments

23. cancer Biomedical Informatics Grid (caBIG™)
A national, NCI-funded program with participants from cancer centers and research institutions
Improve cancer research and accelerate efforts to find a cure for cancer
Make cancer research data and tools (maintained at different sites) more efficiently accessible to researchers
Create scalable, federated, actively managed biomedical informatics network that will connect members of the cancer research enterprise
Informatics support for basic, clinical, and translational research
“National Standards, Local Management”*
caBIG community
50+ Cancer Centers, 30+ Organizations, over 900 participants
Cancer researchers are working in isolated environments--not leveraging findings.
Researchers speak different ‘dialects’ (e.g., molecular biology and anatomical pathology).
Researchers can use one term for different concepts or different terms for the same concept.
Data are typically kept locally and are not easily shared.
The assumption is that combined efforts will move research progress along more rapidly.
caBIG™ provides the unifying infrastructure for sharing and integrating data; using common standards.

24. Example Scenario caGrid Environment Microarray and protein
Registered Object Definitions
Microarray and protein
databases at other institutions
caGrid Service
Interfaces
Location A
Microarray, Protein,
Image data
Advertisement
Location B
Microarray, Protein,
Image data
Discovery
caGrid Environment
Location C
Microarray, Protein,
Image data
Query and Analysis Workflow
Log on, Grid credentials
Location C
Image Analysis
Location D
Image Analysis

25. Users Experience Early Adopters End users Love it !
45 services from 50+ cancer centers (and growing)
Some services are lot more popular than others
Bio-informatics group at Argonne – Details in subsequent slides
APS plans to pursue this model – Brian Tieman will talk more about this
End users
caBIG hack-a-thon participants were happy about workflows
Cancer Centers
Not too happy right now
Would like to see some RoI

26. Integration of caBIG-geWorkbench-Teragrid

27. caGrid Security (GTS, Grid Grouper, Dorian, CDS)

28. Transposon Workflow Details
Transposons are small mutant sequences of DNA that move between positions within the genome of a cell
Gene functions can be identified by inserting mutant sequences into a genome and analysing where the sequence resides (between genes or on a gene) and how it affects its phenotype
BLAST is used to compare the genome after insertion against the original genome
This identifies where the transposon resides

29. Realizing the workflow
Genome Sequences
Find Transposon
Find sequences that have the given transposon
Format DB
Create a fasta file representing the
Genome sequences
Compare these sequences against the original genome
BLAST
Loop Until there are no misses or
all genomes have been searched
Compile a report summarising where the transposon was inserted and results of the BLAST search
Create Report
BLAST
Misses
BLAST
Hits
Neighbouring
Genes

31. Workflow Automation at DOE Facilities
Advanced Photon Source
Storage
Metadata
Analysis
Visualization
Automation
Reproducibility
Security
Reusability
Developer benefits:
Developers spend time developing scientific applications instead of rewriting code to turn into services.
Developer time spared through code reuse—legacy apps are wrappered into services instead of rewritten as services
Learning gRavi easier than learning service development—especially marshaling of security resources!
Traditional applications are not secure and can not provide an audit trail
Scientific benefits:
Scientists spend time developing workflows rather than moving data from application to application
Scientists interact with GUI to focus on scientific process not how to run applications.
Services provide easy access to all scientific tools in one integrated GUI.
Access to grid resources promotes cross disciplinary collaboration as it becomes easier to share data and applications across institutional boundaries
Center for Enabling Distributed Petascale Science

32. Lessons Learned Nice higher level abstraction
Suitable for a subset of scientific computing usecases
Sustainable infrastructure
Work well with hospital/cancer center/APS IT infrastructure
Workflows
Scalability
Provenance
No Vendor lock-in (if you are careful)

33. Further Resources Further Details on gRAVI
Further Details on Introduce
Details on Taverna + gRAVI
gRAVI users mailing list