The Common Component Architecture and XCAT Indiana University Extreme! Lab.

Slides:



Advertisements
Similar presentations
LEAD Portal: a TeraGrid Gateway and Application Service Architecture Marcus Christie and Suresh Marru Indiana University LEAD Project (
Advertisements

Siebel Web Services Siebel Web Services March, From
TSpaces Services Suite: Automating the Development and Management of Web Services Presenter: Kevin McCurley IBM Almaden Research Center Contact: Marcus.
COM vs. CORBA.
This product includes material developed by the Globus Project ( Introduction to Grid Services and GT3.
GridRPC Sources / Credits: IRISA/IFSIC IRISA/INRIA Thierry Priol et. al papers.
Common Object Request Broker Architecture (CORBA) By: Sunil Gopinath David Watkins.
G O B E Y O N D C O N V E N T I O N WORF: Developing DB2 UDB based Web Services on a Websphere Application Server Kris Van Thillo, ABIS Training & Consulting.
Presentation 7 part 2: SOAP & WSDL. Ingeniørhøjskolen i Århus Slide 2 Outline Building blocks in Web Services SOA SOAP WSDL (UDDI)
6/11/2015Page 1 Web Services-based Distributed System B. Ramamurthy.
Distributed Systems Architectures
CS 501: Software Engineering Fall 2000 Lecture 16 System Architecture III Distributed Objects.
Ch 12 Distributed Systems Architectures
Systems Architecture, Fourth Edition1 Internet and Distributed Application Services Chapter 13.
Course Instructor: Aisha Azeem
Center for Component Technology for Terascale Simulation Software (aka Common Component Architecture) (aka CCA) Rob Armstrong & the CCA Working Group Sandia.
Center for Component Technology for Terascale Simulation Software 122 June 2002Workshop on Performance Optimization via High Level Languages and Libraries.
GRAPPA Part of Active Notebook Science Portal project A “notebook” like GRAPPA consists of –Set of ordinary web pages, viewable from any browser –Editable.
Middleware-Based OS Distributed OS Networked OS 1MEIT Application Distributed Operating System Services Application Network OS.
Checkpoint & Restart for Distributed Components in XCAT3 Sriram Krishnan* Indiana University, San Diego Supercomputer Center & Dennis Gannon Indiana University.
A Hybrid Decomposition Scheme for Building Scientific Workflows Wei Lu Indiana University.
COM vs. CORBA Computer Science at Azusa Pacific University September 19, 2015 Azusa Pacific University, Azusa, CA 91702, Tel: (800) Department.
NeSC Grid Apps Workshop Exposing Legacy Applications as OGSI Components using pyGlobus Keith R. Jackson Distributed Systems Department Lawrence Berkeley.
J2EE Structure & Definitions Catie Welsh CSE 432
GT Components. Globus Toolkit A “toolkit” of services and packages for creating the basic grid computing infrastructure Higher level tools added to this.
Architecting Web Services Unit – II – PART - III.
Grids and Portals for VLAB Marlon Pierce Community Grids Lab Indiana University.
1 Overview of the Application Hosting Environment Stefan Zasada University College London.
Grid Computing Research Lab SUNY Binghamton 1 XCAT-C++: A High Performance Distributed CCA Framework Madhu Govindaraju.
Components for Beam Dynamics Douglas R. Dechow, Tech-X Lois Curfman McInnes, ANL Boyana Norris, ANL With thanks to the Common Component Architecture (CCA)
Web Services BOF This is a proposed new working group coming out of the Grid Computing Environments Research Group, as an outgrowth of their investigations.
Center for Component Technology for Terascale Simulation Software CCA is about: Enhancing Programmer Productivity without sacrificing performance. Supporting.
SCIRun and SPA integration status Steven G. Parker Ayla Khan Oscar Barney.
Shannon Hastings Multiscale Computing Laboratory Department of Biomedical Informatics.
1 Geospatial and Business Intelligence Jean-Sébastien Turcotte Executive VP San Francisco - April 2007 Streamlining web mapping applications.
Extreme! Computing Lab, Dept. of Computer Science, Indiana University 1 Programming the Grid with Components Madhu Govindaraju Aleksander Slominski Dennis.
Grid Architecture William E. Johnston Lawrence Berkeley National Lab and NASA Ames Research Center (These slides are available at grid.lbl.gov/~wej/Grids)
Databases JDBC (Java Database Connectivity) –Thin clients – servlet,JavaServer Pages (JSP) –Thick clients – RMI to remote databases –most recommended way.
Presented by An Overview of the Common Component Architecture (CCA) The CCA Forum and the Center for Technology for Advanced Scientific Component Software.
1 UNIT –II Architecting Web Service. 2 Why SOA? – business point of view  Information Technology (IT) workers face many challenges, including: Limited.
NA-MIC National Alliance for Medical Image Computing UCSD: Engineering Core 2 Portal and Grid Infrastructure.
Grid Services I - Concepts
Distribution and components. 2 What is the problem? Enterprise computing is Large scale & complex: It supports large scale and complex organisations Spanning.
Hwajung Lee.  Interprocess Communication (IPC) is at the heart of distributed computing.  Processes and Threads  Process is the execution of a program.
CS 501: Software Engineering Fall 1999 Lecture 12 System Architecture III Distributed Objects.
GRID Overview Internet2 Member Meeting Spring 2003 Sandra Redman Information Technology and Systems Center and Information Technology Research Center National.
 Common Object Request Broker Architecture  An industry standard developed by OMG to help in distributed programming.
S imple O bject A ccess P rotocol Karthikeyan Chandrasekaran & Nandakumar Padmanabhan.
Enabling Self-management of Component-based High-performance Scientific Applications Hua (Maria) Liu and Manish Parashar The Applied Software Systems Laboratory.
CCA Common Component Architecture CCA Forum Tutorial Working Group CCA Status and Plans.
Kemal Baykal Rasim Ismayilov
On Using BPEL Extensibility to Implement OGSI and WSRF Grid Workflows Aleksander Slomiski Presented by Onyeka Ezenwoye CIS Advanced Topics in Software.
Overview of Grid Webservices in Distributed Scientific Applications Dennis Gannon Aleksander Slominski Indiana University Extreme! Lab.
Intro to Web Services Dr. John P. Abraham UTPA. What are Web Services? Applications execute across multiple computers on a network.  The machine on which.
CCA Common Component Architecture CCA Forum Tutorial Working Group Common Component Architecture.
WP3 Implementing R-GMA grid services in GT3 Abdeslem Djaoui & WP3 Grid Services Task Force 7 th EU Datagrid meeting 26/09/2003
Copyright 2007, Information Builders. Slide 1 iWay Web Services and WebFOCUS Consumption Michael Florkowski Information Builders.
Center for Component Technology for Terascale Simulation Software (CCTTSS) 110 April 2002CCA Forum, Townsend, TN CCA Status, Code Walkthroughs, and Demonstrations.
Center for Component Technology for Terascale Simulation Software (CCTTSS) 110 April 2002CCA Forum, Townsend, TN This work has been sponsored by the Mathematics,
CCA Distributed Framework Interoperability. Goals Assume you have two (or more) framework instances. –Assume it contains a network of component instances.
CCA Common Component Architecture CCA Forum Tutorial Working Group Common Component Architecture.
Grid Resource Allocation Agreement Protocol Working Group
Distribution and components
Inventory of Distributed Computing Concepts and Web services
Introduction to Web Services
CS590L Distributed Component Architecture References: - Objects, components, and framenworks with UML, Ch A Component Based Services Architecture.
Distributed System using Web Services
Distributed System using Web Services
Common Component Architecture (CCA)
Presentation transcript:

The Common Component Architecture and XCAT Indiana University Extreme! Lab

What is a Component Architecture? A systematic way of encapsulating special functionality and behaviors of a piece of software into reusable units. More than an object model. It defines the rules for the way objects can be instantiated and composed. It defines the environment of services that a component can use. It defines the way in which we discover how to interact with them.

Isn’t this what a subroutine library does? –Subroutine libraries or class libraries have well defined interfaces encapsulate functionality –But are often hard to reuse because they often have complex resource requirements make conflicting assumptions about their operating environments. –Subroutine libraries that follow very strict sets of design and use rules get reused. Above all else, a component architecture is framework of standard rules of behavior for objects.

Standard Component Archtectures Examples: –Microsoft COM/DCOM, COM++ the foundation of all Microsoft office products –Java Beans Java standard for building user interfaces –Enterprise Java Beans a standard for client-server applications in Java –CORBA 3.0 The new component spec for CORBA.

Look at One Design in Detail The U.S. Dept of Energy DOE2000 project –The Common Component Architecture Lawrence Livermore National Lab Sandia Labs Argonne National Labs Los Alamos National Labs Universities: Utah, NCSA, Indiana –A specification for component design for parallel and distributed applications

Two parts to the CCA Architecture Components –An encapsulated “object” defined by its public interfaces. –Can be Java, C, Python, C++ or C++ and Fortran. Frameworks –The software systems that provides basic services to components –Used to compose components to make apps. –Many implementations: Parallel and Distributed.

Some Concepts in Detail Ports: the public interfaces of a component –defines the different ways we can interact with a component and the ways the component uses services and other components. Image Processing Component setImage(Image I) Image getImage() adjustColor() setFilter(Filter) calls doFFT(…) Provides Ports - interfaces functions provided by component Uses Ports - interface of a service used by component

How do you “compose” two components? There are three basic approaches: –Events: A component can generate an “event” of a particular type. Other components are added as “listeners” for such events from that component. –Services: A source component “uses” services “provided” by a target components. “use” = calls a function provided by the target. –Dataflow: A typed data “output” stream is connected to a similarly typed “input” port of another. In XCAT all three are supported. –The services model is the basic mechanism. –Dataflow is emulated in CCA by “pushing” data from a user to a provider, or by having a user “pull” data from a provider. –Events are managed by a separate mechanism.

Acme FFT component Building Applications by Composition Connect uses Ports to Provides Ports. Image Processing Component getImage() adjustColor() Image tool graphical interface component Image database component setImage(…) doFFT(…)

Ports and Interfaces (CCA) Component composition: –Connect (possibly at runtime) a “provides” port of one component to a “uses” port of another. –A provides port is simply an implementation in the component of the interface defined by the port type (interface defn.). –A Uses port is a “proxy” that a component “uses” to make a request of another component. Source Component Target Component A Provides port A uses port

How do you describe the interfaces a port has? Standard Methods: –Use an Interface Definition Language - a formal description of the interface objects. CORBA IDL is one. The LLNL Scientific IDL by Scott Kohn, Andrew Cleary, Steven Smith, and Brent Smolinski is better for CCA. –A simple IDL compiler can be used to automatically generate the Uses and Provides Port code for a specific IDL type. –For the XCAT java version the interface of a component port are defined as Java interfaces.

Component Communication Use a simple Remote Procedure Call Mechanism – XSOAP Implementation of Simple Object Access Protocol (SOAP) Performance Issues – in-process calls – Events/Messages Objects encoded as XML documents – Proteus Multiprotocol Messaging and RMI

X Creation Service Creates a running instance of another component Encapsulates authentication issues Creation Service Component Launch an instance of component X on resource Y Returns: remote reference to new component instance Globus resource Y

Connection Service A component that can be used to connect a “uses” port of one component to the “provides” port of another Can export ports of another component Y Connection Service Component X Connect port A of component X to port B of component Y A B

Builder Service combination of Creation and Connection service standardized as part of recently updated CCA specification we are now in process of implementing it in XCAT

Simple Example in Java public class SimpleEchoImpl implements SimpleEcho { public String echoHello(String s) throws RemoteException { return "SimpleEchoImpl says: Hello " + s; }

Simple Example in Java public class EchoPrinterComponent implements Component { public void setServices(Services cc){ PortInfo portType = new PortInfoImpl( "simpleEchoProvidesPort", " simpleEchoImpl = new SimpleEchoImpl(); cc.addProvidesPort(simpleEchoImpl,portType); }

Simple Example in Java public class EchoGeneratorComponent implements Component { public void setServices(Services cc){ PortInfo portType = new PortInfoImpl( "simpleEchoUsesPort", " cc.registerUsesPort(portType); }

Simple Example in Java SimpleEcho usesSimpleEcho = (SimpleEcho) cc.getPort("simpleEchoUsesPort"); usesSimpleEcho.echoHello("Extreme Lab"); cc.releasePort(“simpleEchoUsesPort"); Echo Generator Component Echo Printer Component A Provides port simpleEchoProvidesPort A uses port simpleEchoUsesPort

Scripting XCAT Applications import xcat generator = xcat.createComponent(‘GeneratorComponet’) printer = xcat.createComponent(‘Printer’) xcat.setCreationMechanism(generator, ‘gram’) xcat.setCreationMechanism(printer, ‘ssh’) xcat.createInstance(generator) xcat.createInstance(printer) xcat.connectPorts(generator, ‘simpleEchoUsesPort’, printer, ‘simpleEchoProvidesPort’)

Scripting XCAT Applications generatorComponent = EnvObj() generatorComponent.put("exec-name", "simpleGeneratorComponent") generatorComponent.put("exec-fqn", "samples.simpleEchoGenerator.SimpleEchoGe neratorComponent") printerComponent = EnvObj() printerComponent.put("exec-name", "simplePrinterComponent") printerComponent.put("exec-fqn", "samples.simpleEchoPrinter.SimpleEchoPrinte rComponent") # create component wrappers generator = cca.createComponent(generatorComponent) printer = cca.createComponent(printerComponent) # assign a machine name printermc = "exodus.extreme.indiana.edu" generatormc = "exodus.extreme.indiana.edu" cca.setMachineName(generator,generatormc) cca.setMachineName(printer, printermc) # create live instances cca.createInstanceWithTimeOut(printer, ) cca.createInstanceWithTimeOut(generator, ) # connect their ports cca.connectPorts(generator, "simpleEchoUsesPort", printer, "simpleEchoProvidesPort") # start the components – invoke go() on GoPort usesGoPortClassName = "samples.idl.goPort.UsesGoPort" usesPortType = " providesPortName = "providesGoPort" methodName = "go" methodParams = zeros(0, Object) cca.invokeMethodOnComponent(generator, usesGoPortClassName, usesPortType, providesPortName, methodName, methodParams) print "Done"

Scientific Components and Applications Recognized as one of the “Top Ten Science Achievements in 2002” by the DOE Office of Science (see Many demonstrated at SC2001 and SC2002 Combine application-specific components with more general-purpose components that can be reused across a range of applications –More than 40 components, many reused in apps such as PDEs on unstructured and adaptive structured meshes Unconstrained minimization problems Leverage and extend parallel software developed at different institutions –Including CUMULVS, CVODES, Global Arrays, GrACE, MPICH, PETSc, PVM, and TAO

Current Interface Development Activities CCA Scientific Data Components Working Group Basic Scientific Data Objects –Lead: D. Bernholdt (ORNL) Unstructured Meshes –Lead: L.F. Diachin (SNL, formerly ANL) –Collaboration with TSTT ISIC –TSTT = Terascale Simulation Tools and Technologies, PIs: J. Glimm, D. Brown, L.F. Diachin, scidac.orghttp:// scidac.org Structured Adaptive Mesh Refinement –Lead: P. Colella (LBNL) –Collaboration with APDEC ISIC –APDEC = Algorithmic and Software Framework for Applied PDEs, PI: P. Colella, Other Groups Linear and nonlinear solvers, eigensolvers, and optimizers –Coordinator: L.C. McInnes (ANL) –Collaboration with TOPS ISIC –TOPS = Terascale Optimal PDE Simulations, PI: D. Keyes, MxN Parallel Data Redistribution –Lead: J. Kohl (ORNL) –Part of CCTTSS MxN Thrust Quantum Chemistry –Leads: C. Janssen (SNL) and T. Windus (PNNL) –Part of CCTTSS Applications Integration Thrust

Motivation for Common Interfaces Many-to-Many couplings require Many 2 interfaces –Often a heroic effort to understand details of both codes –Not a scalable solution Common Interfaces: Reduce the Many-to-Many problem to a Many-to-One problem –Allow plug-and-play interchangeability & interoperability –Require domain specific experts –Typically difficult & time- consuming –A success story: MPI –Challenges Interface agreement Functionality limitations Maintaining performance NWGrid Overture MDB/CUBIT AOMD Hypre PETSc SuperLU mesh libraries linear solver libraries NWGrid Overture MDB/CUBIT AOMD SuperLU PETSc Hypre DataData Others … SolversSolvers TSTT Data Interfaces TOPS Solver Interfaces

Framework Stays “Out of the Way” of Component Parallelism Single component multiple data (SCMD) model is component analog of widely used SPMD model Each process loaded with the same set of components wired the same way P0P1P2P3 Components: Blue, Green, Red Framework: Gray MCMD/MPMD also supported Different components in same process “talk to each” other via ports and the framework Same component in different processes talk to each other through their favorite communications layer (i.e., MPI, PVM, GA)

There are 3 CCA Frameworks There’s 3 parallelism models in scientific computing Ccaffeine SPMD Distributed Threaded Uintah

Ccaffeine

SPMD GUI and Scripted Interface Interactive or Batch Serial or Parallel Components written in C++ Separates CCA Pattern from Implementation –3 Bindings Classic Components SIDL Components Chasm Components Demonstrated at SC2001 & SC2002 Used in CCA Tutorials Characteristics Achievements

SCIRun/BioPSE/Uintah multithreaded & distributed C++ only Used in “real science” (gov’t, academic, commercial) 1K procs Testbed for CCA Concepts Novel work in IDL-based MxN Parallelism and concurrency research for CCA done on Uintah SCIRun2 will integrate SCIRun, BioPSE and Uintah CharacteristicsAchievements

XCAT Distributed/Grid model Web Services Developed Proteus multi- protocol communication package Novel MxN work at MPI-I/O level Java and C++ components CharacteristicsAchievements Material Archive Go Data GSCntl Go GS Go Data Provider Comp Simulation Component exported Application Coordinator sendParameters, start, kill Application Specific component Application Factory Service Directory/ Discovery Service Authorization Service Resource Broker Service wsdl Ensemble Application 8 Application control

A Science Portal View of “programming” the Grid A Science Portal is a Web server that –Uses a “prepackaged” set of scripts to Get the users proxy cert from a cert repository Configure a specific application to users needs Contact the appropriate application factories –Look for event histories of application execution –Allow the user to contact and control the app. User Portals/ Science Portals Launch, configure And control App factory App Instance

Application Factory Service A service that understands a how to instantiate a running instance of an app component. –You provide it with appropriate requirements initial conditions, etc. via an XML file –The service checks you credentials and authorization May consult resource broker launches the app or runs the appropriate grid script. App factory App Instance Provide me with an instance of application X

Application Component Execution Environment Specification WebsterComponent Indiana University Extreme! computing Dennis Gannon olympus.cs.indiana.edu linbox2.extreme.indiana.edu rainier.extreme.indiana.edu gram ssh exec-dir /u/gannon/xcat_tutorial/scripts …. The information needed to generate input to the factory service

Steps in creating a app instance from the portal to the factory User Configures Application from Web Browser Sets application parameters Launches job User Portals/ Science Portals Web Server Web Server Web Server contacts appropriate application factory service (FS) Supplies FS with task parameters FS contacts Resource Broker and secures job launches. Returns App WSDL to server to browser Job begins execution Publishes its contact point (WSDL) to discovery service Begins publishing status events to event channel. Web server discovers application Allows user to interrogate it or retrieve event traces App factory Resource Broker App Instance Event channel Disc serv

An Application Instance Typical Instance –Publishes event stream to “well known” channel. –Has a Control Interface to allow portal level interrogation or remote steering –May be linked to other components/services Links To Other Apps/ services Control Interface check status control messages Application Instance Event stream

Encapsulating Legacy Apps Common Case –Legacy App that reads files and writes files. –Use a “scripted component” called an app manager. Component runs a python script loaded at startup or through a control command –The App Script Stages files Launches and monitors application Writes Output files publishes event streams application App Mgr input files output files App Script Event Stream Control

XCAT And OGSI A component based model to the Grid services –Component Assembly: Composition in space –Workflow: Composition in time Ability to use widely available Web services tooling to interact with components A richer messaging and notification system that extends the model proposed by OGSI An application factory service that extends the standard OGSI factory service

XCAT-OGSI Big Picture XCAT Component GridService Port (for lifecycle and metadata) Other XCAT Provides Ports ComponentID (GSH) ( uniquely identifies Component ) ServiceData (includes list of Port references)

Incorporating OGSI into XCAT Converting an XCAT component to a Grid service –Addition of an OGSI GridService Port –Addition of Service Data Elements (SDEs) containing references to each of the other ports –Uniquely identifying the component using a ComponentID (GSH) –Using a WSDL representation for the GSR Adding a set of helper services that are OGSI compliant –A HandleResolver service for mapping a GSH to a GSR –Modifications to the XCAT Creation service to make appropriate calls to the Factory and GridService ports for lifetime management

XCAT Vs OGSI Messaging OGSI messaging is simple, point-to-point, and non- reliable, whereas XCAT uses XMessages that provides a reliable, persistent network of message channels suited for application level messaging and events OGSI messaging is push based, and only current moment state can be pulled by accessing SDEs whereas XMessages can be both push and pull based In XCAT, clients can have more mobility and interoperate more easily with firewalls there is no assumption that client has IP reachable address

Factories in XCAT and OGSI OGSI provides a standard Factory Port type for instantiating services XCAT generalizes this Factory Port type to instantiate a set of XCAT components that constitute an application –Accepts a description of a connected network of components –Launches an Application Coordinator for each application that creates, and links together instances of the described network of components –Exports a subset of the functionality of the components, which are useful for the application, to the end-user –Enables creating and managing complex distributed applications