Efficient workspaces through semantic reliability

Transport level protocols typically provide best-effort data delivery where no attempts are made to recover lost packets, or reliable data delivery where techniques are used to ensure that all data sent out eventually reaches the receivers. However, it has been suggested that some applications will perform better when using dynamically configurable reliability based on application level semantics. A general framework has been created with the intent of delivering this type of semantic reliability to a wide variety of applications. To date, however, little research has been done to demonstrate how any applications can benefit from the use of such a framework. This paper addresses this problem in that it discusses semantic reliability in the context of collaborative workspaces.