Tramberend, Henrik: Avocado : a Distributed Virtual Environment framework. 2003
Inhalt
- Introduction
- Related work
- Non-distributed VE systems and standards
- Distributed VE systems
- Research prototypes for scalable DVE systems
- Clustered rendering systems
- Summary and discussion of related work
- DVE Systems - A conceptual approach
- DVE system requirements
- Distributed object and event model
- Display device abstraction
- Direct-manipulation user interaction
- Rapid prototyping and extensibility
- Platform and performance considerations
- From concept to architecture
- Object model
- Event model
- Distributed object model
- Distributed application layout
- Distributed event model
- Network transport layer
- Data input and device interface
- Scripting language selection and binding
- Target platform considerations
- Execution model: Single vs. multi-threading
- A base system for low-level tasks
- System API structure
- Extension mechanism: API and link strategy
- Avocado Architecture Summary
- Avocado - implementation of the framework foundation
- The Avocado object model
- Mapping the field concept to Performer
- Field containers for object state encapsulation
- Smart pointers and reference counting
- Adaption of Performer classes through subclassing
- Field connections for event dissemination
- Scene graph node classes
- Sensor objects
- An external device daemon process
- Integration of the ELK Scheme language
- A component interface for framework extension
- An interface for state object persistence
- Display device abstraction
- Tool-based interaction support
- An interaction metaphor for the Responsive Workbench
- Implementation of the interaction abstraction
- Extension through specialization
- Summary
- Avocado - implementation of the distribution architecture
- Distributed object model
- Distributed event model
- Distributed state change notification and event handling
- Field connections in the distributed context
- Guaranteeing application state consistency
- The Ensemble/Maestro group communication system
- Consistency through total message ordering
- Synchrony through view atomic message delivery
- Dynamic membership and atomic state transfer
- Distributed locking through total message ordering
- A simple distributed application example
- Pacman: A complex distributed application example
- Summary
- Scalability in distributed virtual environments
- Introduction
- Scalability analysis
- An environment model for scalability analysis
- Analysis of existing systems
- Comparison of tiling vs. no scalability mechanism
- Exploiting visibility for rendering
- View frustum culling in visual rendering
- Level-of-detail evaluation
- Hierarchical level-of-detail rendering
- Exploiting visibility for distribution
- Hierarchical environment partitioning
- Mapping distribution groups to scene partitions
- HLOD evaluation and the working set
- HLOD evaluation and view frustum culling combined
- Caching of inactive nodes
- Implementation of the HLOD node
- Scalability analysis of the hierarchical distribution approach
- Summary
- Results, applications and future work
