Contents

• A DIET platform
  • DIET components
  • Communications layer
  • DIET initialization
  • Solving a problem
  • DIET Extensions
    • Multi-MA
    • FAST
    • CoRI
  
  
• DIET installation
  • Dependencies
    • General remarks on DIET platform dependencies
    • Hardware dependencies
    • Supported compilers
    • Operating system dependencies
    • Software dependencies
  • Compiling the platform
    • Obtaining and installing cmake per se
    • Configuring DIET's compilation: cmake quick introduction
    • A ccmake walk-through for the impatients
    • DIET's main configuration flags
    • DIET's extensions configuration flags
    • DIET's advanced configuration flags
    • Compiling and installing
  • Diet client/server examples
    • Compiling the examples
  
  
• DIET data
  • Data types
    • Base types
    • Composite types
    • Persistence mode
  • Data description
  • Data management
    • Data identifier
    • Data file
  • Manipulating DIET structures
    • Set functions
    • Access functions
  • Data Management functions
    • Free functions
  • Problem description
  • Examples
    • Example 1: without persistency
    • Example 2: using persistency
  
  
• Building a client program
  • Structure of a client program
  • Client API
  • Examples
    • Synchronous call
    • Asynchronous call
  • Compilation
    • Compilation using cmake
  
  
• Building a server application
  • Structure of the program
  • Server API
  • Example
  • Compilation
  
  
• Batch and parallel submissions
  • Introduction
  • Terminology
  • Configuration for compilation
  • Parallel systems
  • Batch system
  • Client extended API
  • Batch server extended API and configuration file
  • Server API
    • Registering the service
    • Server configuration file
    • Server API for writing services
    • Example of the client/server 'concatenation' problem
  
  
• Cloud submissions
  • Introduction
  • Compatibility and requirements
  • Configuration for compilation
  • Server configuration file
  • Registering the service
  • Service API
  • Example of client/server
  
  
• Scheduling in DIET
  • Introduction
  • Default Scheduling Strategy
  • Plugin Scheduler Interface
    • Estimation Metric Vector
    • Standard Estimation Tags
    • Estimation Function
    • Aggregation Methods
  • Example
  • Scheduler at agents level
    • Scheduling from the agents side.
    • Aggregation methods overloading
    • The UserScheduler class
    • Easy definition of a new scheduler class
    • Creation and usage of a scheduler module
    • SeD plugin schedulers and agent schedulers interactions
    • A complete example of scheduler
  • Future Work
  
  
• Performance prediction
  • Introduction
  • FAST: Fast Agent's System Timer
    • Building FAST
    • Using FAST in the plug-in scheduler
    • Building a server application with FAST
    • Example with convertors
  • CoRI: Collectors of Ressource Information
    • Functions and tags
    • FAST
    • CoRI-Easy
    • CoRI batch
  • Future Work
  
  
• Deploying a DIET platform
  • Deployment basics
    • Using CORBA
    • DIET configuration file
    • Example
  • GODIET
  • Shape of the hierarchy
  
  
• DIET dashboard
  • LogService
  • VizDIET
  
  
• Multi-MA extension
  • Function of the Multi-MA extension
  • Deployment example
  • Search examples
  
  
• P2P DIET extension: DIET
  • P2P and JXTA
  • Description of the current architecture developed with JXTA
    • The JXTA components
    • Interfacing JXTA and DIET with JNI
  • The future of DIET
    • Remaining problems
  • Working with a DIET platform
    • Installation and configuration
    • Deploying a DIET platform
  
  
• JuxMem extension
  • Introduction
  • Overview of JuxMem
  • How to configure DIET to use JuxMem?
  • Example
  • Troubleshooting
  
  
• Workflow management in DIET
  • Overview
  • Quick start
  • Software architecture
  • Workflow description languages
    • MaDag language
    • Gwendia language
  • Client API
    • Structure of client program
    • The simplest example
  • Scheduling
    • Available schedulers
    • SeD requirements for workflow scheduling
  
  
• DAGDA extension
  • Overview
  • The DAGDA configuration options
  • Cache replacement algorithm
  • The DAGDA API
    • Note on the memory management
    • Synchronous data transfers
    • Asynchronous data transfers.
    • Data checkpointing with DAGDA
    • Create data ID aliases.
    • Data replication
  • On correct usage of DAGDA
  • Future works
  
  
• Dynamic management
  • Dynamically modifying the hierarchy
    • Motivations
    • ``And thus it began to evolve''
    • Example
  • Changing offered services
    • Presentation
    • Example
    • Going further
  
  
• DIET forwarders
  • Easy CORBA objects connections through ssh
  • The DIETForwarder executable
    • Command line options
    • Configuration file
  • Configuration examples
    • Simple configuration
    • Complex network topology
  
  
• Appendix
  • Configuration files
  
  
• Bibliography