A Job Scheduling Optimization Method Based on Burst Buffer Resource Reservation

Abstract

The present invention belongs to the technical field of supercomputers and discloses an optimization method for job scheduling based on the reservation of Burst Buffer resources. A reservation mechanism for Burst Buffer resources is introduced, which specifically includes the following: job submission and resource requirement reservation. The reservation mechanism for Burst Buffer resources specifically includes that the system allocates corresponding Burst Buffer resources for each job in advance according to its requirements for computing resources and Burst Buffer; a time window mechanism for job scheduling, and the scheduling process adopts a time window mode; the job execution order adopts the traditional method, and the execution order is further selected by comparing the optimization index jobs; job backfilling; the optimization target selects the job execution order with the best effect for any one or comprehensive optimization target. The present invention improves the job scheduling efficiency and the utilization rate of Burst Buffer resources during the scheduling process.

Description

Technical Field

[0001] The present invention belongs to the technical field of supercomputers, and particularly relates to a job scheduling optimization method based on Burst Buffer resource reservation. Background Art

[0002] In high-performance computing (HPC) systems, with the rapid development of computing power, the processing capacity of computing tasks has been continuously improved. However, I / O operations (such as data reading and writing) often fail to keep up with the improvement of computing power synchronously, resulting in a significant performance gap between computing power and I / O performance. This gap not only affects the execution efficiency of tasks but also causes excessive waiting time for jobs, thereby leading to inefficient use of resources. To alleviate this problem, Burst buffer technology has emerged. A Burst buffer is a high-speed cache storage device between computing nodes and storage systems, aiming to solve the bottleneck problem of traditional I / O subsystems. By providing a high-throughput cache area, computing tasks can perform data exchange more efficiently during execution. Although Burst buffer technology can effectively improve I / O performance, its resource management and scheduling issues still pose challenges.

[0003] Traditional job scheduling methods do not fully consider the resource reservation of the Burst Buffer, resulting in unbalanced allocation of computing resources and I / O resources, and thus causing a large number of resource gaps. These gaps not only waste precious computing resources but also increase the latency of job backfilling, further affecting the execution efficiency of jobs and the overall system performance.

[0004] When dealing with task scheduling, existing scheduling methods often rely on fixed sequences and simple queuing mechanisms, and fail to flexibly adapt to resource reservation and dynamic scheduling requirements. Especially in high-performance computing systems, due to the complexity of jobs and the uncertainty of computing requirements, traditional scheduling methods cannot efficiently manage Burst buffer resources in a dynamically changing environment. In addition, these methods usually cannot break through the limitations of the scheduling order. For example, backfilled jobs cannot change the execution order and start time of originally scheduled jobs, which leaves much room for optimization in resource usage and job scheduling of the system.

[0005] Currently, there is a small amount of research on the scheduling of burst buffers. It mainly directly adds a burst buffer resource reservation mechanism to traditional scheduling methods, takes this resource into account in resource requirements, and then executes jobs based on traditional scheduling methods. However, it still cannot effectively solve the problem of resource utilization. In addition, existing scheduling algorithms also fail to consider multiple optimization goals, such as the overall system throughput, the execution time of jobs, the efficient utilization of resources, etc. This makes existing research unable to take into account multiple factors when facing complex job and resource scheduling, thus affecting the comprehensiveness and accuracy of scheduling. Summary of the Invention

[0006] The object of the present invention is to provide an optimized job scheduling method based on burst buffer resource reservation to address the above-mentioned problems. It is expected to improve the sequential limitation problem of existing scheduling methods in processing task scheduling. By focusing on multiple optimization goals, it ensures that in various job queue environments, the execution performance of jobs and the resource utilization rate of burst buffers are improved.

[0007] The technical solution adopted by the present invention is as follows: An optimized job scheduling method based on burst buffer resource reservation. The optimization method introduces a burst buffer resource reservation mechanism, which specifically includes the following:

[0008] Job submission and resource requirement reservation: When a user submits a job, they need to provide the requirement parameters for computing resources and burst buffer resources. The burst buffer resource reservation mechanism specifically includes that the system allocates corresponding burst buffer resources for the job in advance according to the requirements of each job for computing resources and burst buffers, and ensures that these resources can be guaranteed before the job execution.

[0009] Time window mechanism for job scheduling: The scheduling process adopts a time window mode.

[0010] Job scheduling method: The execution order of jobs is determined by comparing optimization metrics, and further by the order in which jobs are selected for execution; job backfilling.

[0011] Optimization goal selection: According to the system load composed of the resource requirements of the job queue, select the job execution order with the best effect for any one or comprehensive optimization goal.

[0012] Furthermore, the time window mode means that within a fixed time interval, the jobs in the waiting queue will be fixed and scheduling optimization will be performed; the size of the time window is set by the administrator.

[0013] It should be noted that the size of the time window is set by the administrator or operator according to specific needs.

[0014] Furthermore, the optimization metrics include:

[0015] The total waiting time of the queue, which represents the sum of the waiting times of all jobs in the waiting queue;

[0016] The job slowdown ratio of the queue, which represents the sum of the slowdown ratios of all jobs in the waiting queue;

[0017] The completion time of the queue, which represents the difference between the end time of the last job and the start time of the first job in the waiting queue.

[0018] Furthermore, the formula for calculating the slowdown ratio is:

[0019]

[0020] Where is the slowdown ratio, is the job waiting time, is the job running time.

[0021] Furthermore, during the job backfilling process, appropriate jobs will be selected from each subsequent job in the determined order for early backfilling scheduling;

[0022] When a certain running job ends and releases resources, if the resource requirements of the first job in the waiting queue cannot be executed, the first job whose resource requirements are suitable for backfilling will be selected from the subsequent waiting queue according to the determined order.

[0023] It should be noted that even if the backfilled job will delay the execution of the first job (the running time of the backfilled job is too long), backfilling can still be performed.

[0024] Furthermore, for the three optimization metrics, first standardize and unify the dimensions of each target, and then set the triple weight parameters , and , which correspond to the weight parameters of each optimization metric respectively, and find the set of optimization metrics with the smallest weighted sum. The corresponding job execution order is the optimal order;

[0025] Where corresponds to the total waiting time of the queue, corresponds to the job slowdown ratio of the queue, corresponds to the completion time of the queue.

[0026] Furthermore, the weight parameters are selected by the administrator according to specific needs to suit the current load and resource requirements.

[0027] Furthermore, the size of the time window is set according to the actual load of the system and the requirements of the jobs.

[0028] In summary, due to the adoption of the above technical solutions, the beneficial effects of the present invention are as follows:

[0029] 1. By introducing a reservation mechanism for Burst Buffer resources, a scheduling method for time windows, and a backfilling mechanism, the job scheduling efficiency and the utilization rate of Burst Buffer resources are improved during the scheduling process;

[0030] 2. Break the latency limitation in traditional backfilling scheduling and provide multiple queue optimization objectives for administrators to select and evaluate; ensure that the execution performance of jobs is improved in various job queue environments.

[0031] 3. It can select the most suitable job for scheduling according to the current system resource status and the specific requirements of the job when each job event occurs, so as to achieve efficient utilization of resources;

[0032] 4. It can greatly reduce the gaps in computing resources and I / O resources, and can also significantly reduce the latency of job backfilling, achieving the high efficiency of job execution and the maximization of system resource utilization. This flexible and efficient scheduling method gives the present invention important technical advantages in the field of supercomputer job scheduling and can effectively improve the overall performance of supercomputing systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 is a flowchart of the method of the present invention;

[0034] Figure 2 is a schematic diagram of the architecture of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0035] The present invention will be described in detail below with reference to the accompanying drawings.

[0036] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0037] In high-performance computing (HPC) systems, with the rapid development of computing power, the processing ability of computing tasks has been continuously improved. However, I / O operations (such as data reading and writing) often fail to keep up with the improvement of computing power synchronously, resulting in a significant performance gap between computing power and I / O performance. This gap not only affects the execution efficiency of tasks but also causes excessive waiting time for jobs, leading to inefficient use of resources. To alleviate this problem, Burst buffer technology has emerged. A Burst buffer is a high-speed cache storage device between computing nodes and storage systems, aiming to solve the bottleneck problem of traditional I / O subsystems. By providing a high-throughput cache area, it enables computing tasks to perform data exchange more efficiently during execution. Although Burst buffer technology can effectively improve I / O performance, its resource management and scheduling issues still pose challenges.

[0038] Currently, most job scheduling methods in supercomputer environments rely on traditional scheduling methods, such as First-Come-First-Served (FCFS), Shortest Job First (SJF), etc. These methods usually determine the scheduling priority according to the submission order of jobs or the expected execution time of jobs. However, traditional job scheduling methods fail to consider the reservation and dynamic scheduling of Burst buffer resources, resulting in low job scheduling efficiency and even possible problems such as a large number of resource gaps and job delays, affecting the overall system performance.

[0039] Currently, the implementation scheme closest to the present invention is the optimization method based on backfill scheduling used in the field of high-performance computing. In these systems, backfill scheduling can combine the scheduling methods of FCFS or SJF to reduce the waiting time of jobs. By inserting smaller jobs into the idle time periods of long-running jobs, the resource utilization efficiency is improved. Such scheduling methods rely on the management of job queues and determine the execution order according to the resource requirements of jobs and the submission time of jobs.

[0040] When many traditional scheduling algorithms execute backfill scheduling, they are often subject to sequential constraints. For example, jobs in backfill scheduling must strictly abide by the execution order and start time of the original jobs and cannot be flexibly adjusted. Due to the high coupling between the job scheduling order and resource allocation, the limitations of traditional scheduling methods have significantly restricted the system performance.

[0041] However, backfill scheduling also has obvious shortcomings. First, traditional backfill scheduling cannot handle the reservation of burst buffer resources, resulting in poor coordination between I / O resources and computing resources, and cannot fully utilize the efficient data transmission capabilities provided by the burst buffer. In addition, the job scheduling process only considers the demand for job computing resources. Jobs with high burst buffer requirements cannot be backfilled, resulting in serious job starvation. Secondly, traditional backfill scheduling usually only executes scheduling based on the predetermined order of jobs, and cannot flexibly respond to dynamic changes in the job execution process. For example, in backfill scheduling, once the execution order of jobs is determined, it often cannot be adjusted. This makes it impossible for the system to effectively optimize resource allocation when facing resource shortages or job changes.

[0042] There are a few studies on Burst buffer scheduling, which mainly add the Burst buffer resource reservation mechanism directly to the traditional scheduling method, take this resource into account in the resource demand, and then execute the job based on the traditional scheduling method, which still cannot effectively solve the problem of resource utilization. In addition, the existing scheduling algorithms also fail to focus on multiple optimization goals, such as the overall system throughput, job execution time, and efficient resource utilization. This makes it impossible for existing research to take into account multiple factors when facing complex job and resource scheduling, thus affecting the comprehensiveness and accuracy of scheduling.

[0043] like Figure 1 As shown, the present invention proposes a method for optimizing job scheduling based on Burst Buffer resource reservation. The core idea of ​​the method is to realize the reservation of Burst buffer resources and break the order limitation in traditional scheduling. Specifically, the present invention can select the most suitable job for scheduling according to the current system resource status and the specific needs of the job when each job event occurs, thereby realizing efficient utilization of resources. At the same time, the present invention focuses on multiple optimization goals to ensure that the execution performance of the job is improved in a variety of job queue environments.

[0044] Through this optimization method, the present invention can not only significantly reduce the gaps in computing resources and I / O resources, but also significantly reduce the delay of job backfilling, thereby achieving high efficiency of job execution and maximum utilization of system resources. This flexible and efficient scheduling method enables the present invention to have important technical advantages in the field of supercomputer job scheduling and can effectively improve the overall performance of supercomputer systems.

[0045] like Figure 1 and Figure 2As shown in the figure, a job scheduling optimization method based on Burst Buffer resource reservation introduces a reservation mechanism for Burst Buffer resources and is based on the shared Burst buffer architecture of a supercomputer, such as Figure 2 shown in the figure, to optimize the scheduling in the job waiting queue and improve the job scheduling efficiency and the resource utilization rate of the Burst buffer.

[0046] The optimization method introduces a reservation mechanism for Burst Buffer resources, which specifically includes the following:

[0047] Job submission and resource requirement reservation: When a user submits a job, they need to provide the requirement parameters for computing resources and Burst Buffer resources; the reservation mechanism for Burst Buffer resources specifically includes that the system allocates corresponding Burst Buffer resources for the job in advance according to the requirements of each job for computing resources and Burst Buffer.

[0048] Time window mechanism for job scheduling: The scheduling process adopts a time window mode.

[0049] Job scheduling method: The execution order of jobs is determined by comparing optimization metrics, and further by the order in which jobs are selected for execution; job backfilling.

[0050] Optimization goal selection: According to the system load composed of the resource requirements of the job queue, select the job execution order with the best effect for any or comprehensive optimization goal.

[0051] The specific technical solution is as follows:

[0052] During the scheduling process, considering the limited nature of Burst Buffer resources, a reservation mechanism for Burst Buffer resources is introduced. When each job is submitted, the system allocates corresponding Burst Buffer resources for it in advance according to its requirements for computing resources and Burst Buffer, and ensures that these resources are guaranteed before the job execution. The job scheduling process adopts a time window mode, that is, fixing the waiting jobs within a fixed time interval and then performing scheduling optimization, so as to select a better execution method among the limited jobs. The size of the time window is generally set by the administrator according to the system status.

[0053] The scheduling method includes two parts: execution order and job backfilling.

[0054] 1. When a user submits a job, they need to provide the resource requirement parameters of the burst buffer, so that during the job scheduling process, resource reservation can be performed simultaneously for computing resources and the burst buffer.

[0055] 2. The job execution order adopts traditional methods, such as first-come-first-served, shortest job first, or a custom job execution order. After comparing the optimization metrics, the job selects the execution order. Once the execution order of the job is determined, the job execution priority is set according to the execution order.

[0056] 3. During the backfilling process, suitable jobs will be selected from subsequent jobs in a determined order for early backfilling scheduling. Specifically, when a running job finishes and releases resources, if the resource requirements of the first job in the waiting queue cannot be executed, the first job with suitable resource requirements for backfilling will be selected from the subsequent waiting queue according to the determined order. Even if the backfilled job will delay the execution of the first job (the running time of the backfilled job is too long), it can still be backfilled, while traditional job scheduling cannot achieve this step because it forcibly requires that the backfilled job cannot delay the execution of the first job.

[0057] Among them, the optimization metrics include three items:

[0058] The total waiting time of the queue, which represents the sum of the waiting times of all jobs in the waiting queue;

[0059] The job slowdown ratio of the queue, which represents the sum of the slowdown ratios of all jobs in the waiting queue; by optimizing the job slowdown ratio, the system can ensure the execution efficiency of the job and reduce the "delay" time of the job.

[0060] The calculation formula for the slowdown ratio is:

[0061]

[0062] Among them, is the slowdown ratio, is the job waiting time, is the job running time.

[0063] The completion time of the queue, which represents the difference between the end time of the last job and the start time of the first job in the waiting queue.

[0064] The administrator can select the job execution order with the best optimization target effect for any one or comprehensive according to the resource requirements composition of the job queue and the system load. For the three optimization metrics, each target is first standardized to unify the dimension, and then the triple weight parameters ( , , ), which correspond to the weight parameters of each optimization metric respectively, are used to find the set of optimization metrics with the smallest weighted sum. The corresponding job execution order is the optimal one. The weight parameters are determined by the administrator.

[0065] Example 1

[0066] As Figure 1 shown, an embodiment of the present invention is a method for optimizing job scheduling based on Burst Buffer resource reservation, and its technical solution specifically includes:

[0067] 1. Job submission and resource reservation:

[0068] (1) Resource requirements when users submit jobs: When users submit jobs, they must provide the required computing resources and the demand parameters of Burst Buffer resources. Specifically, the job not only needs to specify how much computing resources are required, but also needs to clarify the required Burst Buffer capacity (such as storage space size and access speed requirements). The core purpose of this step is to enable the system to know in advance the resource requirements of each job before job execution and perform resource reservation during the scheduling process.

[0069] (2) Resource reservation mechanism for Burst Buffer: According to the resource requirements submitted by users, the system will allocate Burst Buffer resources in advance for each job. In this way, before the job starts to execute, the system can ensure that the required resources have been allocated in place, avoiding resource shortages or resource waiting situations. Resource reservation is an optimization for the finiteness of Burst Buffer resources to ensure that there are no resource conflicts during scheduling.

[0070] 2. Job scheduling within the time window:

[0071] Time window mechanism: During the job scheduling process, the system adopts the "time window" mode to optimize job scheduling. Specifically, the system will "freeze" the jobs in the current waiting queue at fixed time intervals (for example, every certain period), and then schedule according to the job resource requirements during this period.

[0072] Setting of the time window size: The size of the time window is set by the administrator according to the actual load of the system and the requirements of the jobs. The setting of the time window size will directly affect the scheduling effect. For example, a smaller window can schedule more frequently, but may bring higher scheduling overhead; a larger window may miss some optimization opportunities.

[0073] 3. Job scheduling order and execution:

[0074] Selection of the job scheduling order: Within the time window, the system determines the job execution order according to the pre-set scheduling method. Traditional scheduling methods such as "First Come First Served" (FCFS) or "Shortest Job First" (SJF) can be selected, or the optimal scheduling order can be selected based on optimization metrics (such as waiting time, slowdown ratio, etc.).

[0075] Job execution priority: Once the job execution order is determined, the system assigns priorities to each job according to this order. Jobs with higher priorities will be scheduled and executed first, while jobs with lower priorities will be executed later.

[0076] 4. Backfill mechanism:

[0077] The backfill mechanism is during the job execution process. When a job is completed and releases resources, if the jobs queued in front in the waiting queue cannot be executed immediately, the system will select a job with appropriate resource requirements from the subsequent jobs in the waiting queue for "early backfill". Even if the execution of the backfill job may delay the execution of the first job, the system will still perform backfill. This can avoid the idle of job resources, thereby improving the system resource utilization rate.

[0078] Traditional job scheduling usually does not allow the backfill job to delay the execution of the first job. However, in the method of this embodiment, even if the running time of the backfill job is long, backfill is still allowed, which can effectively fill the resource idle period. When selecting a backfill job, the system will select the first suitable job from the subsequent jobs according to the current resource requirements and the order of the waiting queue to ensure the efficient use of resources.

[0079] 5. Optimization metrics:

[0080] The total waiting time of the queue, which represents the sum of the waiting times of all jobs in the waiting queue. The waiting time of each job refers to the time when the job is waiting to be executed. Optimizing this metric can reduce the average waiting time of jobs;

[0081] The job slowdown ratio of the queue, which represents the sum of the slowdown ratios of all jobs in the waiting queue. The slowdown ratio is a measure of the slowdown effect of each job;

[0082] The formula for calculating the slowdown ratio is:

[0083]

[0084] Where, is the slowdown ratio, is the job waiting time, is the job running time.

[0085] By optimizing the slowdown ratio of jobs, the system can ensure the execution efficiency of jobs and reduce the "delay" time of jobs.

[0086] The completion time of the queue represents the difference between the end time of the last job and the start time of the first job in the waiting queue. Optimizing this metric helps to shorten the processing time of the entire queue, thereby improving the overall efficiency.

[0087] 6. Comprehensive optimization and weight parameters:

[0088] Target standardization: When there are multiple optimization metrics, each metric needs to be standardized first to convert it into a unified dimension, so as to effectively perform weighted calculation.

[0089] Setting of weight parameters: The system will assign a weight parameter to each optimization metric according to the administrator's settings. 、 and , and these weights represent the relative importance of each metric in the optimization goal. The administrator can adjust according to the specific situation and select the optimization goal that best suits the current system load and resource requirements.

[0090] Weighted sum minimization: By performing weighted summation on each optimization metric, the system calculates the optimal job execution order. The goal is to minimize the weighted sum, so as to select the optimal job scheduling order and ensure that the system maximizes the job scheduling efficiency and resource utilization on the premise of meeting resource reservation.

[0091] Through the above technical solutions, it is possible to effectively improve the efficiency of job scheduling, reduce waiting time, and improve the resource utilization rate of the BurstBuffer.

[0092] Embodiment 2

[0093] As Figure 2 shown, another embodiment of the present invention is that a job scheduling optimization system based on Burst Buffer resource reservation adopts the Lustre parallel file system. Lustre adopts a distributed lock management mechanism and object storage technology, slices large files and dispersedly stores them on multiple OSTs (object storage targets), thereby improving the concurrency of multi-user access and the aggregated I / O bandwidth, and is suitable for large-scale file read and write operations. The computing nodes are separated from the storage nodes, and high-speed network interconnection is used to achieve efficient data transmission. The job scheduling system not only schedules computing tasks but also is responsible for reasonably allocating I / O resources to ensure that the backend storage will not become a performance bottleneck when encountering a large number of sudden I / O requests. The shared Burst Buffer node, as a high-speed cache device, is located between the computing node and the backend storage and is specifically used to buffer and accelerate a large number of instantaneous I / O operations, can quickly respond to burst I / O requests, reduce the waiting time of the computing node, and thus improve the overall job execution efficiency.

[0094] Figure 2CN is the I / O (input / output) interaction channel between the computing nodes that execute computing tasks and the external storage system. It includes two working modes: when cache optimization is required for a job, the computing nodes read the original data from the parallel file system through the I / O interface, and temporarily store the frequently accessed intermediate results in the Burst Buffer nodes with cache characteristics to relieve the direct access pressure on the parallel file system. Finally, the final results are written to the parallel file system through the I / O interface for persistent storage; if the job does not require Burst Buffer acceleration, the computing nodes directly complete the data reading and writing operations with the parallel file system through the I / O.

[0095] Due to its limited resource capacity and high cost, the scheduling system must evaluate the I / O requirements of tasks during job allocation, preferentially allocate the limited cache resources to tasks that are sensitive to I / O performance or have obvious I / O burst characteristics, and dynamically adjust the resource allocation according to the real-time load during task execution to make full use of the limited high-speed storage resources, ensure that each job obtains sufficient bandwidth at critical moments, and achieve efficient coordinated operation of computing and data storage.

[0096] The above are only the preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for optimizing job scheduling based on Burst Buffer resource reservation, characterized in that: The optimization method introduces a Burst Buffer resource reservation mechanism, which specifically includes the following: Job submission and resource demand reservation: When submitting a job, the user needs to provide the demand parameters for computing resources and Burst Buffer resources. The Burst Buffer resource reservation mechanism specifically includes the system allocating corresponding Burst Buffer resources to the job in advance based on the computing resources and Burst Buffer requirements of each job. The time window mechanism of job scheduling, the scheduling process adopts the time window mode; Job scheduling method, the order of job execution is further selected by the job by comparing the optimization indicators; job backfilling; Optimization target selection, based on the system load formed by the job queue resource requirements, select the job execution order that best achieves any or a combination of optimization targets.

2. The method for optimizing job scheduling based on Burst Buffer resource reservation according to claim 1, characterized in that: The time window mode means that within a fixed time interval, the jobs in the waiting queue will be fixed and the scheduling optimization will be performed; the size of the time window is set by the administrator.

3. The method for optimizing job scheduling based on Burst Buffer resource reservation according to claim 1, characterized in that: The optimization indicators include: The total waiting time of the queue, which represents the sum of the waiting times of all jobs in the queue; The job deceleration ratio of the queue, which represents the sum of the deceleration ratios of all jobs in the waiting queue; The completion time of the queue, which is the difference between the end time of the last job in the waiting queue and the start time of the first job.

4. The method for optimizing job scheduling based on Burst Buffer resource reservation according to claim 3, characterized in that: The reduction ratio calculation formula is: in, is the reduction ratio, is the job waiting time, The time the job runs.

5. The method for optimizing job scheduling based on Burst Buffer resource reservation according to claim 1, characterized in that: During the job backfilling process, a suitable job will be selected from each subsequent job in a determined order for early backfill scheduling; When a running job is finished and resources are released, if the resource requirement of the first job in the waiting queue cannot be executed, the first job with resource requirement suitable for backfilling will be selected from the subsequent waiting queue according to the determined order.

6. The method for optimizing job scheduling based on Burst Buffer resource reservation according to claim 3, characterized in that: For the three optimization indicators, each target is first standardized and unified, and then the triple weight parameters are set , and , respectively corresponding to the weight parameters of each optimization index, find the minimum set of weighted optimization indexes, and the corresponding job execution order is the optimal order; in, The total waiting time of the corresponding queue, The job reduction ratio of the corresponding queue, The completion time of the corresponding queue.

7. The method for optimizing job scheduling based on Burst Buffer resource reservation according to claim 6, characterized in that: The weight parameters are selected by the administrator according to specific needs to suit the current load and resource requirements.

8. The method for optimizing job scheduling based on Burst Buffer resource reservation according to claim 2, characterized in that: The size of the time window is set according to the actual load of the system and the requirements of the operation.

Images