Modeling Performance of Smart Devices in Closed Systems
Smartphones and tablets are becoming dominant devices in the present market. Apart from offering the same features as traditional mobile phones, smart devices provide several other features as well. Smart device users try to optimize for mobility and lightness rather than for maximum computing performance and throughput. Therefore, designing smart devices to increase performance against these constraints becomes a challenge. In an effort to study the performance of smart devices, we modeled the components of smart devices using an M / M / 1 / N closed queue system. In this model, the number of requests in the system is fixed. The performance of requests under the open and closed system models are compared. We observe that average response time generally increases with increase in arrival rate, and increase in load values. The average response time however, decreases with increase in the service rate. It is further observed that requests experience lower average response time under the open system model at low arrival rates, low service times and low load. However, at high arrival rates, high service rates and high load values, requests experience lower response time under the closed system model. A similar trend is observed with the average queue length which generally increases with increase in arrival rate, and increase with increase in load values. The average queue length however decreases with increase in the service rate. It is further observed that requests experience lower average queue length under the open system model at low arrival rate, low service rate and low load. However, at high arrival rate, high service rate and high load values, requests experience lower average queue length. The analytical results indicate that closed system models offer better performance at high arrival rates and high load values, whereas open system models offer better performance at low arrival rates and low load values.