Performance Evaluation of Data Sensor Platform
Oral Defence Date:
Wednesday, May 14, 2014 - 09:30
Profs. Puder and Dujmovic
Wireless sensor networks consist of spatially distributed autonomous devices using sensors to cooperatively monitor physical or environmental conditions at different locations. Data collected by those sensors must be transferred to processing centers. During transfer, the data may travel through intermediate nodes -- which may manipulate the data, encounter poor quality communication channels or overloaded components which affect the network’s scalability and performance. Networked Bay Environmental Assessment and monitoring Stations (NetBEAMS) is an environmental sensor networks developed as collaborative effort between the Computer Science Department at San Francisco and the Romberg Tiburon Center (RTC). Data Sensor Platform (DSP) is a proposed implementation for NetBEAMs. We aim at developing a simulator to evaluate performance and scalability of NetBEAMs implemented over DSP. The main idea is to model DSP components as Extended Queues which are extensions to Queuing Systems. These extensions support a more detailed modeling of DSP components. The Simulator models the behavior of individual Extended Queues and the interaction with other Extended Queues when transferring sensor data through the network. We collect metrics on individual Extended Queues which allows the identification of performance degradation and bottlenecks in the NetBEAMs. This approach has a couple of advantages: the Simulator only models Extended Queues which make the code much simpler to implement and easier to validate. Additionally, Extended Queues are extension to Queuing Systems and same key performance metrics can be used to analyze the DSP behavior. This work describes the DSP and how components are modeled as Extended Queues. The Simulator design and validation are discussed. We then simulate the behavior of DSP under different configurations and workload analyzing the resulting performance.
Sensor Network, Performance, Simulation, Queue Systems