摘要
Mobile grid is a branch of grid computing that incorporates mobile devices into the grid infrastructure. It poses new challenges because mobile devices are typically resource-constrained and exhibit unique characteristics such as instability in network connections. New scheduling strate- gies are imperative in mobile grid to efficiently utilize the devices. This paper presents a scheduling algorithm that con- siders dynamic properties of mobile devices such as avail- ability, reliability, maintainability, and usage pattern in mo- bile grid environments. In particular, usage patterns caused by voluntarily or involuntarily losing a connection, such as switching off the device or a network interruption could be important criteria for choosing the best resource to execute a job. The experimental results show that our scheduling al- gorithm provides superior performance in terms of execution time, as compared to the other methods that do not consider usage pattern. Throughout the experiments, we found it es- sential to consider usage pattern for improving performance in the mobile grid.
Mobile grid is a branch of grid computing that incorporates mobile devices into the grid infrastructure. It poses new challenges because mobile devices are typically resource-constrained and exhibit unique characteristics such as instability in network connections. New scheduling strate- gies are imperative in mobile grid to efficiently utilize the devices. This paper presents a scheduling algorithm that con- siders dynamic properties of mobile devices such as avail- ability, reliability, maintainability, and usage pattern in mo- bile grid environments. In particular, usage patterns caused by voluntarily or involuntarily losing a connection, such as switching off the device or a network interruption could be important criteria for choosing the best resource to execute a job. The experimental results show that our scheduling al- gorithm provides superior performance in terms of execution time, as compared to the other methods that do not consider usage pattern. Throughout the experiments, we found it es- sential to consider usage pattern for improving performance in the mobile grid.
