national collaboratories middleware projects
DESCRIPTION
National Collaboratories Middleware Projects. Ray Bair NC Program Coordinator January 15, 2002. Topics. What is middleware? Goals of NC Middleware research What’s coming from the NC program Interacting with NC projects Tools you can use now Appendix: More about middleware projects. - PowerPoint PPT PresentationTRANSCRIPT
National CollaboratoriesMiddleware Projects
Ray BairNC Program Coordinator
January 15, 2002
Topics
What is middleware? Goals of NC Middleware research What’s coming from the NC program Interacting with NC projects Tools you can use now Appendix: More about middleware projects
Middleware: software that connects or mediates between two otherwise separate programs
NC Middleware Program is focused on… Technology to enable ubiquitous access to remote resources –
computation, information and expertise Capabilities that make it easier for distributed teams to work
together, over the short and long term Standard services and protocols for access to networked
resources, that aid software development/interoperability Middleware advances that enable scientific computing, e.g.,
high performance for scientific applications
Middleware RolesPortals and PSEsPortals and PSEs
ApplicationsApplications
The GridThe Grid
ComponentArchitectures
Collaboratory
DistributedSecurity
Architectures
ScientificAnnotationMiddleware
Reliable/Secure
Communication
PervasiveCollaborativeEnvironment
Groupto Group
Collaboration
DataGridMiddleware
StorageResource
Management
CoGKits
Middlewarefor Science
Portals
DOE NC Middleware Projects
Realizing the Science Grid
Tools for Distributed Applications
Realizing and Enhancing the Science Grid
Consistent access to resources and information High performance data transport Reliable and secure communication among peers Tools to manage data Tools to utilize metadata
Metadata: information about data or other information
Distributed Security Architectures Mary Thompson, LBNL Concept
Secure/flexible way to authorize access to distributed resources• Based on signed Policy, Use-condition and Attribute certificates
Stakeholders flexibly set access policy on a per resource granularity Existing Technologies
Akenti 1.0beta, Secure Akenti-enhanced web server beta Sample secure ORBIX CORBA ORB with Akenti integration
Future Capabilities Enhanced, secure Akenti-enabled Apache web server Web interface to generate UseCondition/Attribute certificates Proxy credential delegation capabilities Integration with GSI – Grid Security Infrastructure
SciDAC DataGrid MiddlewareI. Foster, ANL; C. Kesselman, ISI; Miron Livny, UW
Concept Innovative techniques for co-reservation of Grid compute, network, and storage
resources, and market brokering services High performance I/O, with intelligent, adaptive recovery Efficient, distributed replica management to improve access efficiency
Existing Technologies Globus Toolkit, including GridFTP, GRAM, … Condor
Future Capabilities New Data Mover Technology supporting very high speed, reliable data movement, via
new protocols Data Transfer Management for orchestrating data transfers Collective Data Management to support replication, mirroring, wide area hierarchical
storage management
Reliable and Secure Group CommunicationDeb Agarwal, LBNL
Concept Many-to-many group communication that scales to the Internet Flexible message delivery in terms of reliability and ordering Peer-to-peer, secure, reliable, ordered multicast
Existing Technologies InterGroup Protocol Design, and prototype implementation
Future Capabilities Group Security Layer, for secure group communications
• Creates an SSL equivalent for group communication Software tools that support these protocols
Storage Resource Management for Data Grid App’sArie Shoshani, LBNL; Don Petravick, FNAL
Concept Grid Middleware to support dynamic storage management for long lasting
simulation and analysis tasks in a distributed environment Coordinate distributed disk caches by “pinning” of files
Existing Technology Existing HPSS-HRM (Hierarchical Resource Manager) used by several HEP
applications, based on logical file requests Future Capabilities
Disk Resource Manager – manages a single shared disk cache Storage Resource Manager – manages what data is on each storage device Sophisticated pinning capabilities – manages requests to cache files
Scientific Annotation MiddlewareJim Myers, Elena Mendoza, PNNL; Al Geist, Jens Schwidder, ORNL
Concept Unified, lightweight metadata infrastructure to support the creation and
use of metadata and annotations Annotations shared among portals and problem solving environments,
software agents, scientific applications, electronic notebooks, … Existing Technology
DOE 2000 Electronic Notebooks Future Capabilities
Search Tool, Metadata Viewer, Graphical Relationship Browser, Data Viewer, Data Signature Widget
Notebook Explorer & Viewer
Building Distributed Applications Collaboratories, Portals, and Problem-Solving Environments
Use standard application components in portals Use Grid services from high level frameworks Tools for constructing scientific workflow Create virtual venues for collaborative work
Portal: a science-oriented PSE, typically with a web browser interface, that allows scientists to compose and run distributed applications, or to access and analyze distributed data.
Middleware Technology to Support Science PortalsDennis Gannon, Randall Bramley, Indiana U.
Concept A Science Portal that makes it easy to build and access Grid-based
scientific applications from desktop web tools Organized as a set of “Active Notebooks” with web forms to launch and
control the application, as well as histories Existing Technology
GCE-WG Portal Software Repository, Indiana Active Notebook Future Capabilities
Active Documents – for science-centric input and analysis Active Notebooks – contain grid-scripts and run histories, that can be
modified and shared Component composition of application services
SciDAC Commodity Grid (CoG) KitsGregor von Laszewski, ANL; Keith Jackson, LBNL
Concept Reusable “Web Services” that access underlying Grid services. Support rapid development of Science Portals, Problem Solving
Environments, and applications that access Grid resources Existing Technology
CoG Kit prototype Future Capabilities
CoG access to basic Grid services• GRAM, MDS, Security, co-scheduling, …
Support to portal development teams Components that are composable with a visual tool
Pervasive Collaborative Computing EnvironmentDeb Agarwal, LBNL; Miron Livny, U Wisconsin Concept
Collaboration tools that support the continuum of collaborative interaction in a persistent environment
Workflow tools that enable coordination of Grid computing processes and human tasks
Existing Technology Range of DOE 2000 tools, university and commercial software Collaborative Virtual Workspace – building/room metaphor
Future Capabilities Building blocks for typical problems in scientific collaborations Workflow Framework Integrate messaging, Grid security, async. collaboration, Condor-G
Middleware to Support Group to Group CollaborationRick Stevens, ANL Concept
Collaborative work sharing beyond simple application sharing High end visualization env’ts integrated into collaborative spaces Extending asynchronous collaboration capabilities to embrace all types of
data streams exchanged in a collaboration Existing Technology
Access Grid Future Capabilities
Scalable virtual venue server Improved AG security model Workspace Docking (app sharing) Easier node management Tiled display interfaces Multimedia record/playback
How do we get some of these things?Can we influence the direction of projects? Interaction and feedback
are fundamental elements of the NC program.
All projects have dissemination plans.
Learn project details at tomorrow’s Poster Session.
Keep the dialogue going.http://DOEcollaboratory.pnl.gov
NC Project Interactions
Supplemental Information
More about individual middleware projects
Middleware Project Contacts and Web Sites Distributed Security Architectures
Mary Thompson, [email protected] High-Performance Data Grid Toolkit
Ian Foster, [email protected] Reliable and Secure Group Communication
Deb Agarwal, [email protected], www-itg.lbl.gov/CIF/GroupComm/ Storage Resource Management for Data Grid App’s
Arie Soshani, [email protected], sdm.lbl.gov/srm Scientific Annotation Middleware
Jim Myers, [email protected], www.emsl.pnl.gov:2080/docs/collab/sam/ Middleware Technology to Support Science Portals
Dennis Gannon, [email protected] CoG Kits
Gregor von Laszewski, [email protected], www.cogkits.org Pervasive Collaborative Computing Environment
Deb Agarwal, [email protected], www-itg.lbl.gov/Collaboratories/pcce.html Middleware to Support Group-to-Group Collaboration
Rick Stevens, [email protected], www-fp.mcs.anl.gov/fl/g2g/
Impact and Connections IMPACT.
Provide an authorization service based on X.509 identity certificates, compatible with GSI/SSL connections, that can be easily used by distributed applications.
Resource stakeholders can flexibly set access policy on a per resource granularity
CONNECTIONS: Used by the National Fusion Collaboratory to provide
authorization for their distributed applications Run as a server on the DOE Science Grid to allow
applications running there to use it. Used in an SBIR project, CoDeveloper which facilitates
multi-domain code development.
Milestones/Dates/Status Release Akenti 1.1 DONE
libraries and standalone server more flexible certificate formats
Recognize new style proxy X.509 certificates Yr -1 add code to verify certificate chains
Integration with GSI callouts Yr-2 allow Akenti authorization to replace grid-mapfile
Support for additional dynamic policies as required by Grid and collaboratory applications Yr 2-3
The Novel Ideas Authorization to access distributed resources based on X.509 identity certificates and other signed certificates: Policy, Use-condition and Attribute certificates. Multiple stakeholders remotely controlling access to distributed resources. Authorization server easily called from a resource gatekeepers
Principal Investigators: Mary Thompson, LBNLFunding Level: 3 FTEs Oct 23, 2001
MICS Program Manager: Mary Anne Scott
WebServer
LDAP
Certificate Servers
Akenti
Internet
FileServers
Fetch Certificate
CacheManager
LogServer
DN
Akenti Server Architecture
Client
DN
DN
ResourceServer
Distributed Security Architectures
Impact and Connections IMPACT.
Wide spread acceptance of the protocols and services developed will insure interoperability of data grids.
High quality APIs and SDKs implementing these protocols will be provided to allow easier access to data grid tech.
Efficient, distributed replica management will improve data access efficiency
CONNECTIONS: To be used by numerous SciDAC collaboratories, including DOE
Science Grid, Particle Physics Data Grid, Earth Systems Grid, and Fusion Collaboratory. Uses Security Middleware components
Also to be used by many non-DOE projects worldwide, including NSF PACI DTF, NASA IPG, and EU DataGrid
Milestones/Dates/StatusMon Yr DONE
Deliver GridFTP clients and servers (non-striped) 10/01Deliver Replica Management Service 04/02Replica Management Service used in real data grid 08/02Distributed Replica Catalog 01/03Deliver Extended Resources Mgr & Info Svc. 06/03New data channel technologies demonstrated: 10/03 Non-TCP, dynamic rate limiting, FEC, etc.Replica Management Svc used w/ SRM for reliable, 05/04 high performance, scheduled transferDemonstrate alternative data mgmt approaches 10/05Deliver Adv. DataGrid Mgmt systems w/ common 10/06SRM services
The Novel Ideas Develop new protocols taking advantage of unique properties of data gridsDevelop innovative techniques for co-reservation of compute, network, and storage resources with guarantees for each stage.Develop market brokering services to make efficient decisions in the face of constraints, when multiple resources are available to fulfill a request.Investigate variants of two-phase I/O strategies used in parallel I/O systems for data transfer optimizationDevelop intelligent, adaptive recovery and performance strategies based on knowledge of end to end routes and guarantees.
Principal Investigators: I. Foster, ANL; C. Kesselman, ISI; Miron Livny, UWSeptember 14, 2001
National Collaboratories Program
MICS Program Manager: Mary Ann Scott
ControlInterface
Scheduling Modules
Data Channel
Bulk Transfer Protocol
TCP Transfer Protocol
Rate Limiting Interface
Enquiry
GRIP GridFTP TCP, BTP… TCP, BTP…
Data Channel
Bulk Transfer Protocol
TCP Transfer Protocol
Rate Limiting Interface
Resource Mgmt.(disk, NIC)
GRAM
SciDAC DataGrid Middleware
Impact and Connections IMPACT:
Improved communication infrastructure for collaborative applications enabling truly peer-to-peer applications
Many-to-many group communication that scales to the Internet
A secure group layer that creates an SSL equivalent for group communication
Flexibility to implement a broad range of application requirements.
CONNECTIONS: Pervasive Collaborative Computing Environment
Milestones/Dates/Status The primary goal of this project is the development and implementation of group communication capabilities that are reliable and secure
Reliable Multicast Year - Development of InterGroup 1-2 - Beta release of the InterGroup protocol 2 - Testing and implementation of additional feature 2-4 Secure Group Layer: - Proofs of security for the cryptographic algorithms 1-2 - Implementation of protocols 2-4 Improvements - Enhancements to scalability and features 5
The Novel Ideas Developing the infrastructure needed to support true peer-to-peer communication Secure group communication that is peer-to-peer and based on crypto algorithms that are provably secure Reliable multicast capabilities that are scalable to the Internet Flexible message delivery options in terms of reliability and ordering
Principal Investigators: Deb Agarwal - LBNL
Picture/DiagramRelated to Project
Date Prepared
MICS/SciDAC Program Name
PI Org’sLogo
MICS Program Manager: Mary Anne Scott
GroupCommunication
Reliable and Secure Group Communication
Impact and Connections IMPACT
Provides essential component of Grid Middleware Will enable the dynamic coordination of compute and
storage resources Support storage management for long lasting
simulation and analysis tasks in a distributed environment
Manage job recovery from storage system and network failures, facilitating uninterrupted operation
CONNECTIONS Particle Physics Data Grid Earth Science Data Grid Globus Grid projects
Milestones/Dates/StatusMon Yr DONE
Design and develop SRM prototypes - Design functionality and interfaces Oct 2001 done - Implement prototype Disk RM Jan 2002 - Implement prototype Tape RM Feb 2002 Deploy in PPDG STAR project - Initial installation at LBNL Mar 2002 - Initial installation at BNL Jun 2002 Robustness design and development - design Jan 2003 - development of robust SRMs Jun 2003 - deployment Jan 2004
The Novel Ideas SRMs provide Grid Middleware to manage storage resources
- Complement management of compute and network resources- Large files transfer can form the bottleneck
Coordinate distributed disk caches by “pinning” of files- Use “smart” replacement policies
Manage seamless access to tape storage
- automate staging and archive requests in background
- insulate client from hardware & network failures
Principal Investigator: Arie Shoshani, LBNL Co-principal Investigator: Don Petravick, FNAL 9/7/2001
Collaboratory Middleware
MICS Program Manager: Mary Ann Scott
SRM(site A)
RequestManager
SRM(site B)
metadataindex
Replicacatalog
NetworkWeatherService
Grid
LocalSRM
DiskCache
clientclient ...
tape system
SRM(site C) ...
GridMiddleware
Services
DiskCache
DiskCache
tape system
SRM(site A)
RequestManager
SRM(site B)
metadataindex
Replicacatalog
NetworkWeatherService
Grid
LocalSRM
DiskCache
clientclient ...
tape systemtape system
SRM(site C) ...
GridMiddleware
Services
DiskCache
DiskCache
tape systemtape system
Storage Resource Managementfor Data Grid Applications
Impact and Connections IMPACT:
Single unified metadata infrastructure A service for next-generation scientific computing
environments with significant reduction of integration barriers
An advanced notebook view of annotation data CONNECTIONS: The project will work closely with interested
Collaboratory Pilot projects including the Collaboratory for Multiscale Chemical Science, and will investigate integration/connection opportunities with other infrastructure efforts including those developed through the Scientific Data Management Center and Portal Middleware projects. External partnerships are also being investigated.
A lightweight, flexible middleware to support the creation and use of metadata and annotations
Layered architecture Use of standard protocols
Sharing of annotations among portals and problem solving environments, software agents, scientific applications, and electronic notebooks
Support for arbitrary schema Configurable schema translation and metadata extraction
Improved completeness, accuracy, and availability of the scientific record Integrating annotation and records functionality with primary data stores
Principal Investigators: Jim Myers, Elena Mendoza – PNNL Al Geist, Jens Schwidder – ORNL 11/02/01
National Collaboratories Program
MICS Program Manager: Mary Anne Scott
Proposed Timetable
Specification Alpha Release 1.0 Release 1.5 Release
Metadata Services 9/01 3/02 9/02 9/03
Semantic Services 12/01 7/02 12/02 7/04
Notebook Services3/02 12/02 7/03 7/05
Interface Components 3/02 12/02 7/03 7/05
Pedigree Schema 9/02 7/03
Notebook Interface3/02* 12/03 7/04
Novel Ideas
Scientific Annotation Middleware
Impact and Connections IMPACT.
Encapsulating Grid application into science portals will make grid-based solutions available to many more scientists than are currently using grid technology
Using component composition of application services provides the first simple model for Grid application programming.
This will eliminate many of the obstacles to both building and using grid-based applications.
CONNECTIONS: This work will use the common component technology from CCTTSS as well as SciDac collaboration and Grid technology. Numerous SciDac application communities will be approached to test the ideas.
Milestones/Dates/StatusMilestone Category: System Design Mon Yr DONE - Notebook Archive 11 / 01 11 / 01 - Secure Access to grid services 11 / 01 11 / 01 - Advanced Scripting (LBNL interaction) 6 / 02 - Full multi-user, secure, server + archive 8 / 02 - DOE Science Grid integration 11 / 02Milestone Category: Collaboration technology - Integration with DOE Notebook interface 8 / 02 - Integration with Access Grid Technology 1 / 03 Milestone Category: Applications - PPDG+ (Atlas/Griphyn) prototype 3 / 02 - ESG + NCSA environmental hydrology 5 / 02 - Other collaboratories 5/02 thru 3 / 03
The Novel Ideas A Science Portal is tool that make it easy to access Grid-based scientific applications from simple desktop web tools. The portal is organized as a set of “Active Notebooks” which contain the web forms needed to launch and control the application as well as histories of the user’s prior experiments with that application. This histories contain parameter used in the run and links to output files. The execution of a notebook application is governed by a “grid script” for that notebook. Users may edit the scripts to create new application notebooks which can be shared with others.
- Notebook applications may, and often do, link multiple, scientific grid “components” and services.
Principal Investigators: Dennis Gannon, Randal Bramley, Department of Computer Science, Indiana University
Date Prepared 1/02
MICS/SciDAC National Collaboratories and High Performance Networks: Middleware
MICS Program Manager: Mary Ann Scott
Middleware Technology to Support Science Portals: a Gateway to the Grid
Impact and ConnectionsIMPACT. Allow application developers to make use of Grid services from higher-
level frameworks such as Java and Python. Easier development of advanced Grid services. Easier and more rapid application development. Encourage code reuse, and avoid duplication of effort amongst the
collaboratory projects. Encourage the reuse of Web Services as part of the Grids.
CONNECTIONS: We are working closely or as part of with the Globus research project, we work with a variety of major funded applications through SciDAC, NSF, en EU grants, E.g. DOE Science Grid, Earth Systems Grid, Supernova Factory, NASA IPG.
The Novel Ideas• Develop a common set of reusable components for accessing Grid services. • Focus on supporting the rapid development of Science Portals, Problem Solving Environments, and science applications that access Grid resources.• Develop and deploy a set of “Web Services” that access underlying Grid services.
• Integrate the Grid Security Infrastructure (GSI) into the “Web Services” model. Provide access to higher level Grid services that are language independent and are described via commodity Web technologies such as WSDL..
Principal Investigators: Gregor von Laszewski, ANL Keith. Jackson, LBL 09/07/2001
MICS/SciDAC Program Name
MICS Program Manager: Marry Ann Scott
Milestones/Dates/Status The main goal of this project is to create Software Development Kits in both Java and Python that allow easy access to Grid services.
Provide access to basic Grid services: Year - GRAM, MDS, Security, GridFTP 1 - Replica Catalog, co-scheduling 1&2 Composable Components: - Develop guidelines for component development 1 - Design and implement component hierarchies 1&2 - Develop a component repository 2&3 Web Services: - Integrate GSI 1 - Develop an initial set of useful web services 1&2
GlobusToolkit
Java based Grid Portals and Applications
Java CoG ToolkitPythonCoG
Toolkit
CommodityPython
Tools andServices
CommodityJava
Tools and Services
PortalHigh
Energy Physics
Biology PSEChemistryPython
IDEEarthScience
JavaIDE
Java Distributed Programming
Framework
Java CoG Globus Service
…
…
Composable CoG Components
SciDAC CoG Kits
Impact and Connections IMPACT:
A persistent operating environment that facilitates day-to-day operations within collaborations
Natural collaboration capabilities for solving computation-based problem
Readily available building blocks that focus on typical problems found in scientific collaborations
CONNECTIONS: DOE Science Grid and Supernova Factory
Milestones/Dates/Status The primary goal of this project is the development of the Pervasive Collaborative Computing Environment
Workflow development: Year - Installation of Condor-G 1 - Develop underlying services 1&2 Collaboration capabilities: - Integration of Grid services 1 - Secure human interaction environment 2 - Support for asynchronous collaboration 3 Integration of components: 3 -
The Novel Ideas Focus on providing collaboration tools that enable connectivity and collaboration on a day-by-day basis Develop workflow tools that will enable coordination of Grid computing processes and human tasks in a workflow framework Leverage off of the Grid computing environment (e.g. security and directory services) Support the continuum of collaborative interaction
Principal Investigators: Deb Agarwal – LBNL, Miron Livny - UWDate Prepared
MICS/SciDAC Program Name
PI Org’sLogo
MICS Program Manager: Mary Anne Scott
Pervasive Collaborative Computing Environment
Impact and Connections IMPACT.
Wide-spread deployment and use of high-end collaboration technologies to further scientific inquiry
Advances in our understanding of the effects of distance based collaboration environments on group dynamics and communication quality
Extending asynchronous collaboration capabilities to embrace all types of data streams exchanged in a collaboration, with synchronized capture and playback
CONNECTIONS: SciDAC collaboratory pilot projects. SciDAC Software Centers, Grid middleware for discovery, security and information services. Other scientific collaborations
Milestones/Dates/StatusMilestones for End of year 1
Venues Services - V1.0 Architecture document, Prototype - Access Control Architecture - Docking Architecture, API Display: - Node Management Architecture, Software V1.0 release - Xplit Prototype, Architecture white paper Asynchronous Collaboration tools: - Software Architecture Definition documents - New media type plugins for existing tools - Generalized Voyager server V2.0 release
The Novel Ideas Peer-to-peer Virtual Venues servers to enable worldwide, secure virtual communities through the use of high-end collaboration environments Collaborative work sharing beyond simple application sharing Integration of High end visualization environments into collaborative spaces Methods of asynchronous collaboration: capture, synchronization, record, playback and annotation of collaborative experiences
Principal Investigators: Rick Stevens, Argonne National Laboratory
Picture/DiagramRelated to Project
9/13/2001
MICS/SciDAC Middleware
MICS Program Manager: Mary Anne Scott
Middleware to Support Group to Group Collaboration
Out Takes
Distributed Security Architectures Mary Thompson, LBNL
Novel Ideas Secure and flexible way to authorize access to distributed resources
• Based on signed Policy, Use-condition and Attribute certificates. Multiple stakeholders remotely control access to resources. Authorization server easily called from a resource gatekeepers
Impact An authorization service, based on X.509 identity certificates and
compatible with GSI/SSL connections, that can be easily used by distributed applications.
Resource stakeholders can flexibly set access policy on a per resource granularity.
Distributed Security Architectures Mary Thompson, LBNL
Novel Ideas Secure and flexible way to authorize access to distributed resources
• Based on signed Policy, Use-condition and Attribute certificates. Multiple stakeholders remotely control access to resources. Authorization server easily called from a resource gatekeepers
Impact An authorization service, based on X.509 identity certificates and
compatible with GSI/SSL connections, that can be easily used by distributed applications.
Resource stakeholders can flexibly set access policy on a per resource granularity. Web
ServerLDAP
Certificate Servers
Akenti
Internet
FileServers
Fetch Certificate
CacheManager
LogServer
DN
Client
DN
DN
ResourceServer
SciDAC DataGrid MiddlewareI. Foster, ANL; C. Kesselman, ISI; Miron Livny, UW
Novel Ideas New protocols take advantage of unique properties of data grids Innovative techniques for co-reservation of compute, network, and storage
resources, and market brokering services Variants of two-phase I/O strategies Intelligent, adaptive recovery and performance strategies
Impact Common protocols & services will insure interoperability of data grids APIs and SDKs implementing these protocols will be provided to allow
easier access to data grid technology Efficient, distributed replica management will improve data access efficiency
Reliable and Secure Group CommunicationDeb Agarwal, LBNL
Novel Ideas Infrastructure to support true peer-to-peer communication Secure peer-to-peer group communication Reliable multicast capabilities that are scalable to the Internet Flexible message delivery in terms of reliability and ordering
Impact Flexible communication infrastructure for collaborative applications
that are truly peer-to-peer Many-to-many group communication that scales to the Internet Group layer that creates an SSL equivalent for group
communication
Storage Resource Management for Data Grid App’sArie Shoshani, LBNL; Don Petravick, FNAL
Novel Ideas Grid Middleware to manage storage resources Coordinate distributed disk caches by “pinning” of files Manage seamless access to tape storage
Impact Provides essential component of Grid Middleware Will enable dynamic coordination of compute and storage resources Support storage management for long lasting simulation and analysis
tasks in a distributed environment Manage job recovery from storage system and network failures,
facilitating uninterrupted operation
Storage Resource Management for Data Grid App’sArie Shoshani, LBNL; Don Petravick, FNAL
Novel Ideas Grid Middleware to manage storage resources Coordinate distributed disk caches by “pinning” of files Manage seamless access to tape storage
Impact Provides essential component of Grid Middleware Will enable dynamic coordination of compute and storage resources Support storage management for long lasting simulation and analysis
tasks in a distributed environment Manage job recovery from storage system and network failures,
facilitating uninterrupted operation
SRM(site A)
RequestManager
SRM(site B)
metadataindex
Replicacatalog
NetworkWeatherService
Grid
LocalSRM
DiskCache
clientclient ...
tape system
SRM(site C) ...
GridMiddleware
Services
DiskCache
DiskCache
tape system
SRM(site A)
RequestManager
SRM(site B)
metadataindex
Replicacatalog
NetworkWeatherService
Grid
LocalSRM
DiskCache
clientclient ...
tape systemtape system
SRM(site C) ...
GridMiddleware
Services
DiskCache
DiskCache
tape systemtape system
Scientific Annotation MiddlewareJim Myers, Elena Mendoza, PNNL; Al Geist, Jens Schwidder, ORNL
Novel Ideas Lightweight, flexible middleware to support the creation and use of
metadata and annotations Sharing of annotations among portals and problem solving
environments, software agents, scientific applications, and electronic notebooks
Single unified metadata infrastructure Impact
Improved completeness, accuracy, and availability of the scientific record
Significant reduction of integration barriers An advanced notebook view of annotation data
Scientific Annotation MiddlewareJim Myers, Elena Mendoza, PNNL; Al Geist, Jens Schwidder, ORNL
Novel Ideas Lightweight, flexible middleware to support the creation and use of
metadata and annotations Sharing of annotations among portals and problem solving
environments, software agents, scientific applications, and electronic notebooks
Single unified metadata infrastructure Impact
Improved completeness, accuracy, and availability of the scientific record
Significant reduction of integration barriers An advanced notebook view of annotation data
Electronic Notebook Interface
Applications
Agents
Problem SolvingEnvironments
Scientific Annotation Middleware
Notebook Services
records mgmt., annotationtimestamps, signaturesimport/export/archive
Search &Semantic Navigation
Services
Metadata Management
Services
DataArchives
Web
Com
pon
ents
an
d S
ervi
ce I
nte
rfac
es
Dat
a S
tore
In
terf
ace
Building Distributed Applications Collaboratories, Portals, and Problem-Solving Environments
Use standard application components in portals Use Grid services from high level frameworks Tools for constructing scientific workflow Create virtual venues for collaborative work
Portal: a science-oriented PSE, typically with a web browser interface, that allows scientists to compose and run distributed applications, or to access and analyze distributed data.
Middleware Technology to Support Science PortalsDennis Gannon, Randall Bramley, Indiana U.
Novel Ideas A Science Portal that makes it easy to access Grid-based
scientific applications from simple desktop web tools Organized as a set of “Active Notebooks” with web forms to
launch and control the application, as well as histories Execution is governed by a “grid script” for that notebook
Impacts Encapsulating Grid app’s into science portals will make grid-
based solutions easier to build and available to more scientists Using component composition of application services provides
the first simple model for Grid application programming
SciDAC Commodity Grid (CoG) KitsGregor von Laszewski, ANL; Keith Jackson, LBNL
Novel Ideas Common set of reusable components for accessing Grid services Support rapid development of Science Portals, Problem Solving
Environments, and applications that access Grid resources “Web Services” that access underlying Grid services.
Impact Allow use of Grid services from higher-level frameworks Easier development of advanced Grid services Easier and more rapid application development Encourage code reuse, and reuse of Web Services
SciDAC Commodity Grid (CoG) KitsGregor von Laszewski, ANL; Keith Jackson, LBNL
Novel Ideas Common set of reusable components for accessing Grid services Support rapid development of Science Portals, Problem Solving
Environments, and applications that access Grid resources “Web Services” that access underlying Grid services.
Impact Allow use of Grid services from higher-level frameworks Easier development of advanced Grid services Easier and more rapid application development Encourage code reuse, and reuse of Web Services
GlobusToolkit
Java based Grid Portals and Applications
Java CoG Toolkit PythonCoG
Toolkit
CommodityPython
Tools andServices
CommodityJava
Tools and Services
PortalHigh
Energy Physics
Biology PSEChemistry
PythonIDEEarth
Science
JavaIDE
Java Distributed Programming
Framework
Java CoG Globus Service
…
…
Composable CoG Components
Pervasive Collaborative Computing EnvironmentDeb Agarwal, LBNL; Miron Livny, U Wisconsin
Novel Ideas Collaboration tools that enable connectivity and collaboration on a
day-by-day basis Workflow tools that enable coordination of Grid computing
processes and human tasks Support the continuum of collaborative interaction
Impact A persistent operating environment that facilitates day-to-day
operations within collaborations Natural collaboration capabilities for computation-based problems Building blocks for typical problems in scientific collaborations
Middleware to Support Group to Group CollaborationRick Stevens, ANL Novel Ideas
Peer-to-peer Virtual Venues servers to enable worldwide, secure virtual communities via high-end collaboration env’ts
Collaborative work sharing beyond simple application sharing High end visualization env’ts integrated into collaborative spaces Methods of asynchronous collaboration
Impact Wide-spread deployment and use to further scientific inquiry Advances in our understanding of the effects of distance based
collaboration environments on group dynamics Extending asynchronous collaboration capabilities to embrace all types of
data streams exchanged in a collaboration
How do we get some of these things?Can we influence the direction of projects? Feedback is essential.
Middleware PIs want to interact with you !
All projects have dissemination plans.
Learn project details at tomorrow’s Poster Session.
Keep the dialogue going.
What collaboratory and middleware tools can we use today? To Work Together
Conferencing: Access Grid, H.323, ISDN, NetMeeting, MSN Messenger, vic/vat, ImmersaDesk
eMail Lists: majordomo, mailman Shared Documents: Web, NFS, ELN, EN, Notes Shared Display: NetMeeting, Access Grid, vnc, SameTime Code Repository: cvs
To Build Distributed Applications Grid Services: Globus, Condor, Legion, Harness, Cactus,
CoG Authentication Certificates: Netscape, Akenti, Globus