| 摘要 |
A method of scheduling periodic processes for execution in an electronic system, in particular in a network, in a data processor or in a communication device, wherein the electronic system includes a controller for performing the scheduling, wherein a number of N processes Pi are already scheduled with each process Pi having an individual service interval Sli, and wherein an additional process PN+1 is to be scheduled with a service interval SIN+1, is characterized in that the additional process PN+1 is scheduled in such a way that the minimum time lag between the execution times of the additional process PN+1 and the execution times of the number of N already scheduled processes Pi is maximized. Furthermore, a corresponding device for scheduling periodic processes for execution in an electronic system is disclosed. |
| 主权项 |
1. Method of scheduling periodic processes for execution in an electronic system, including in a network, in a data processor, or in a communication device, wherein said electronic system includes a controller for performing said scheduling, wherein a number of N processes Pi are already scheduled with each process Pi having an individual service interval SIi, and wherein an additional process PN+1 is to be scheduled with a service interval SIN+1, said method comprising the steps of
scheduling, at each of plural scheduling stages, said additional process PN+1 to maximize a minimum time lag between the execution times of said additional process PN+1 being currently scheduled and the execution times of said number of N already scheduled processes Pi by assigning an appropriate initial execution time tie to said additional process PN+1 being currently scheduled, wherein said controller calculates the minimum time lag between the execution times of said additional process PN+1 to be currently scheduled and the execution times of each of said number of N already scheduled processes Pi, and expresses each calculated minimum time lag as a function of said initial execution time tie to be assigned to said additional process PN+1, and wherein a minimum of all individual minimum time lags between said additional process PN+1 to be currently scheduled and each of said number of N already scheduled processes Pi defines an absolute minimum time lag which is maximized by computing a period T of a function that represents the absolute minimum time lag as a function of the initial execution time tie of the additional process PN+1 to be currently scheduled, and wherein said initial execution time tie of said additional process PN+1 to be scheduled is determined by way of breaking an N-flows problem into a number of N different 2-flow problems. |
