Scalable, adaptive streaming for nonlinear media

Abstract

Streaming linear media objects has become ubiquitous on today's Internet. Interactive, nonlinear media objects, such as large 3D models and visualization databases, have proven difficult to stream. The Channel Set Adaptation (CSA) framework lets clients request custom data flows for interactive applications using standard broadcast or multicast join and leave operations. CSA scales to support large user groups while providing interactive data access to clients. © 2007 IEEE.