Performance Modeling of Asynchronous Real-time Communication Within Bluetooth Networks

• Main Author: Case, Steven V.
• Published: NSUWorks 2003
• Subjects: Computer Sciences
• Online Access: http://nsuworks.nova.edu/gscis_etd/446

This research provides an advance in the application of wireless, ad hoc networks to the domain of distributed, real-time applications. Traditionally, wireless communications are not deployed within real-time systems, the attributes of wireless protocols tend to run counter to the temporal requirements of real-time systems. When wireless protocols have been used in real-time systems, the application tends to be limited to systems for which there exists a priori knowledge of the network structure or the network communication. This research provides a model (or methodology) for evaluating the extent by which Bluetooth supports deterministic communication, thus allowing system engineers the ability to validate Bluetooth's ability to support real-time deadlines within software applications based on asynchronous communication. This research consisted primarily of an evaluation of the applicability of Bluetooth protocols to asynchronous real-time communication. The research methodology consisted of three distinct stages of research and development. The first stage of the study comprised the development of an analytical model describing the expected behavior of Bluetooth's ACL transmissions and the ability of ACL data packets to meet real-time deadlines. During the second stage of the study, the focus turned to the implementation of the Bluetooth HCI and L2CAP protocol layers. This implementation served as a test harness to gather actual performance data using commercial Bluetooth radios. The final stage of the study consisted of a comparative analysis of the predicted behavior established during the first stage of the study and the actual behavior experienced using the Bluetooth implementation from the second stage of the study. The analysis demonstrates the effectiveness of the model (from the first stage) by measuring the model's ability to accurately predict a piconet's ability to meet real-time deadlines for asynchronous communication when measured at the HCI-L2CAP protocol layer boundary.