What Are the Different Types of Data Processing Jobs?
In a certain time interval, several jobs are performed in the information processing system at the same time, and the processing method of sharing the central processing unit resources in a time-sharing manner. Multi-channel job processing is proposed to overcome the shortcomings of low utilization rate of equipment that often exist in single-channel job processing. Single-channel job processing means that only one job is executed in the information processing system, and all the resources of the system are occupied by this job. Other jobs must be executed after this job is completed.
- Chinese name
- Multi-job processing
- Foreign name
- Multichannel job processing
- Information processing system
- deal with
- Scheduling method with short average turnaround time
- In a certain time interval, several jobs are performed in the information processing system at the same time, and the processing method of sharing the central processing unit resources in a time-sharing manner. Multi-channel job processing is proposed to overcome the shortcomings of low utilization rate of equipment that often exist in single-channel job processing. Single-channel job processing means that only one job is executed in the information processing system, and all the resources of the system are occupied by this job. Other jobs must be executed after this job is completed.
- Single-track job processing selects a scheduling method that shortens the average turnaround time of the job, enables the job to continue automatically, and improves resource utilization. Commonly used scheduling algorithms are: First come first served scheduling algorithm. Scheduling is based on the order in which the jobs arrive, giving priority to the job with the longest waiting time in the system, regardless of the length of time it requires. Shortest job priority scheduling method. Select the job that requires the shortest run time as the target for the next execution, regardless of how long it waits in the system. Because most jobs are short jobs, this scheduling algorithm can satisfy most users, but it is possible that after a long job enters the system, it will not be able to run due to the continuous entry of shorter jobs. The scheduling algorithm with the highest response priority. The response ratio RP is defined as: where the response time is the waiting time after the job enters the system plus the estimated running time. The response ratio of a job increases with the increase in waiting time, as long as it waits long enough, there is always the possibility that it will become the highest response ratio, and get the opportunity to run. This algorithm is a compromise between the above two scheduling algorithms, but the algorithm is more complicated.
Multiple job processing can load several jobs at the same time in the memory. When one job cannot work because it is waiting for external transmission, the central processing unit can immediately execute another job. If the second job cannot continue for some reason During execution, the central processing unit executes the third job, and so continues until the first job is externally transmitted before executing the first job. Adopting a reasonable combination of multiple operations (for example, combining scientific computing operations with a large amount of computation and data processing operations with a large amount of input and output) can improve the utilization of resources and enhance the input and output capabilities of the system. Multi-job processing is performed in a pseudo-parallel manner. From a time interval, each job has started to execute, but has not completed. From a certain moment, only one job is executed in the central processing unit, and each job occupies a time slice, and the central processing unit is used alternately and serially. The jobs do not end in the order in which they were started.
Multiple jobs share various resources of the processing system, but the system's resources are limited, and the number and type of resources required for each job are different. Therefore, the scheduling of multiple jobs uses the following strategies according to the different resource requirements of each job: First-come-first-served, a backup queue is established according to the job entry order, and the scheduler scans the backup queue from the beginning to find the first job that the resource can be satisfied, and inserts it into the current queue for execution. Scheduling by priority, the system selects the job with the highest priority for execution. The priority of the job can be specified by the user (the system charges a higher fee for the job with a higher priority), or it can be determined by the system (the system specifies according to the waiting time of the job, the length of the running time, and the amount of system resources required). This priority number can be determined when the job enters the system, or it can be calculated each time the job is scheduled. Jobs with the same priority are still scheduled on a first come, first served basis. Balanced scheduling, classifying jobs according to their own characteristics. For example, Class A is an input-output time-consuming job, Class B is an input-output and computing time-balanced job, and Class C is a computing-time-consuming job. The job scheduler selects running jobs from these different types of jobs in turn, so that the resources are used in a balanced manner, the system efficiency is exerted, and the users are satisfied.
In addition to the number of multi-task operations, which are directly related to the configuration of the system's external equipment and the size of the memory capacity, they are also related to the type of operation, which is usually 2-10. When multiple jobs are processed, there will be two or more jobs waiting forever because of competing system resources, which is called deadlock. Deadlock can bring serious consequences to the system, so precautions must be taken when designing the system.