CN108170254B - A Policy Scheduling Method Based on Sporadic Requests - Google Patents

Abstract

The invention discloses a strategy scheduling method based on sporadic requests, which comprises the steps of firstly obtaining historical comprehensive records of a cloud storage system, and establishing a scheduling strategy cost model based on sporadic requests; when a first sporadic request in a disk I/O port is detected, judging whether to execute the first sporadic request according to a scheduling strategy cost model; if yes, executing a first sporadic request; otherwise, the first sporadic request is dispatched to other disks according to a preset dispatching strategy. The technical scheme of the invention can be used for measuring whether to execute sporadic requests through the cost model, and the energy consumption of the disk is reduced to the maximum extent on the premise of ensuring the performance of the disk.

Description

A Policy Scheduling Method Based on Sporadic Requests

technical field

The invention relates to the field of computer technology, in particular to a strategy scheduling method based on sporadic requests.

Background technique

A disk is a carrier used to store computer data, and the data stored on the disk can be used repeatedly. During disk usage, there will be sporadic requests during both bursty and non-bursty periods. Traditional disks have insufficient perception in the face of sporadic requests, and cannot respond to sporadic requests in a timely manner.

Contents of the invention

The embodiment of the present invention proposes a strategy scheduling method based on sporadic requests, which can minimize the energy consumption of the disk on the premise of ensuring the performance of the disk.

An embodiment of the present invention provides a policy scheduling method based on sporadic requests, including:

Obtain historical comprehensive records of the cloud storage system, and establish a scheduling policy cost model based on sporadic requests; the historical comprehensive records include: storage resource utilization records, system load records and disk energy consumption status records;

When the first sporadic request in the disk I/O port is detected, according to the scheduling policy cost model, it is judged whether to execute the first sporadic request;

If so, execute the first sporadic request; otherwise, schedule the first sporadic request to other disks according to a preset scheduling policy.

Further, the acquisition of the historical comprehensive records of the cloud storage system, and the establishment of a scheduling policy cost model based on sporadic requests are specifically:

According to the historical comprehensive records of the cloud storage system obtained, determine the utilization rate of storage resources, the read/write ratio of I/O load, the historical number of sporadic requests, and the energy consumption status of disks, and according to the utilization rate of storage resources, I/O Based on the read-write ratio of O load, the historical number of sporadic requests, and the energy consumption status of disks, a scheduling strategy cost model based on sporadic requests is established.

Further, according to the scheduling policy cost model, it is judged whether to execute the first sporadic request, specifically:

According to the scheduling policy cost model, calculate the cost cost of executing the first sporadic request, if the cost cost is lower than a preset threshold, determine to execute the first sporadic request; otherwise, determine not to execute the First sporadic request.

Further, the detection of the first sporadic request in the disk I/O port is specifically:

Obtain the historical system behavior of the block data, and construct the behavior analysis mathematical model of the block data; the behavior analysis mathematical model records the description information and evaluation mechanism of several unexpected access behaviors; the behavior analysis mathematical model is used to track and predict sudden access behavior in the system;

The first sporadic request in the disk I/O port is detected through the behavior analysis mathematical model.

Further, the historical system behavior of acquiring block data is specifically:

Obtain the historical system behavior of block data through the Trace log file.

Implementing the embodiment of the present invention has the following beneficial effects:

The sporadic request-based policy scheduling method provided by the embodiment of the present invention first obtains the historical comprehensive records of the cloud storage system, and establishes a sporadic request-based scheduling policy cost model; when the first sporadic request in the disk I/O port is detected , according to the scheduling strategy cost model, judge whether to execute the first sporadic request; if yes, execute the first sporadic request; otherwise, schedule the first sporadic request to other disks according to a preset scheduling strategy. Compared with the lack of perception of sporadic requests in the prior art, the technical solution of the present invention measures whether to execute sporadic requests through a cost model, and minimizes the energy consumption of the disk on the premise of ensuring disk performance.

Description of drawings

Fig. 1 is a schematic flow chart of an embodiment of a policy scheduling method based on sporadic requests provided by the present invention.

Detailed ways

The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

Referring to FIG. 1 , it is a schematic flowchart of an embodiment of a sporadic request-based policy scheduling method provided by the present invention. The method includes steps 101 to 104, and each step is specifically as follows:

Step 101: Obtain historical comprehensive records of the cloud storage system, and establish a scheduling strategy cost model based on sporadic requests; historical comprehensive records include: storage resource utilization records, system load records, and disk energy consumption status records.

In this embodiment, step 101 is specifically: according to the obtained historical comprehensive record of the cloud storage system, determine storage resource utilization, I/O load read/write ratio, historical number of sporadic requests, and disk energy consumption status, And according to the utilization rate of storage resources, the ratio of read and write of I/O load, the historical number of sporadic requests and the energy consumption state of the disk, a scheduling policy cost model based on sporadic requests is established. The scheduling policy cost model is used to comprehensively measure the influence of multiple factors on the scheduling policy, such as the ratio of I/O load reads and writes, the number of sporadic requests, and different energy consumption states of disks, in the system based on multi-objective optimization theory.

Step 102: When the first sporadic request in the disk I/O port is detected, judge whether to execute the first sporadic request according to the scheduling policy cost model; if yes, execute step 103; otherwise, execute step 104.

In this embodiment, the detection of the first sporadic request in the disk I/O port in step 102 is specifically: obtaining the historical system behavior of the block data, constructing a behavior analysis mathematical model of the block data; and recording the behavior analysis mathematical model The description information and evaluation mechanism of several sudden access behaviors are provided; the behavior analysis mathematical model is used to track and predict the sudden access behavior in the system; through the behavior analysis mathematical model, the first disk I/O port is detected Sporadic requests.

In this embodiment, acquiring the historical system behavior of the block data is specifically: acquiring the historical system behavior of the block data through a Trace log file.

In this embodiment, in step 102, according to the scheduling policy cost model, it is judged whether to execute the first sporadic request, specifically: according to the scheduling policy cost model, the cost cost of executing the first sporadic request is calculated, if the cost cost is low If it is less than the preset threshold, it is determined to execute the first sporadic request; otherwise, it is determined not to execute the first sporadic request. The size of the preset threshold can be set according to the actual situation of the system.

Step 103: Execute the first sporadic request.

Step 104: According to a preset scheduling strategy, schedule the first sporadic request to other disks.

In this embodiment, the preset scheduling strategy is to schedule the first sporadic request to the disk with the lowest load or the lowest storage resource utilization according to the load of each disk and the utilization rate of storage resources. It can also be considered comprehensively and weighted After calculation, schedule to the disk with the highest weight.

As can be seen from the above, the strategy scheduling method based on sporadic requests provided by the embodiment of the present invention first obtains the historical comprehensive records of the cloud storage system, and establishes a scheduling policy cost model based on sporadic requests; For sporadic requests, judge whether to execute the first sporadic request according to the scheduling strategy cost model; if yes, execute the first sporadic request; otherwise, schedule the first sporadic request to other disks. Compared with the lack of perception of sporadic requests in the prior art, the technical solution of the present invention measures whether to execute sporadic requests through a cost model, and minimizes the energy consumption of the disk on the premise of ensuring disk performance.

Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be implemented through computer programs to instruct related hardware, and the programs can be stored in a computer-readable storage medium. During execution, it may include the processes of the embodiments of the above-mentioned methods. Wherein, the storage medium may be a magnetic disk, an optical disk, a read-only memory (Read-Only Memory, ROM) or a random access memory (Random Access Memory, RAM) and the like.

The above description is a preferred embodiment of the present invention, it should be pointed out that for those skilled in the art, without departing from the principle of the present invention, some improvements and modifications can also be made, and these improvements and modifications are also considered Be the protection scope of the present invention.

Claims (3)

1. A strategy scheduling method based on sporadic request is characterized by comprising the following steps:

Determining the utilization rate of storage resources, the read-write proportion of I/O (input/output) loads, the historical number of sporadic requests and the energy consumption state of a disk according to the acquired historical comprehensive record of the cloud storage system, and establishing a scheduling strategy cost model based on the sporadic requests according to the utilization rate of the storage resources, the read-write proportion of the I/O loads, the historical number of the sporadic requests and the energy consumption state of the disk; the history comprehensive record comprises: storing resource utilization rate records, system load records and disk energy consumption state records;

Acquiring historical system behaviors of block data, and constructing a behavior analysis mathematical model of the block data; the behavior analysis mathematical model records description information and an evaluation mechanism of a plurality of sudden access behaviors; the behavior analysis mathematical model is used for tracking and predicting sudden access behaviors in the system; detecting a first sporadic request in the disk I/O port through the behavior analysis mathematical model, and judging whether to execute the first sporadic request according to the scheduling strategy cost model when the first sporadic request in the disk I/O port is detected;

if so, executing the first sporadic request; otherwise, according to a preset scheduling strategy, the first sporadic request is scheduled to the disk with the lowest load or the lowest utilization rate of the storage resources or the highest weight.

2. the method for policy scheduling based on sporadic request according to claim 1, wherein said determining whether to execute the first sporadic request according to the scheduling policy cost model specifically comprises:

Calculating cost for executing the first sporadic request according to the scheduling strategy cost model, and determining to execute the first sporadic request if the cost is lower than a preset threshold; otherwise, determining not to execute the first sporadic request.

3. the method for policy scheduling based on sporadic request according to claim 1, wherein the obtaining of the historical system behavior of block data specifically comprises:

and acquiring the historical system behavior of the block data through the Trace log file.