A cloud platform expansion method based on intelligent monitoring

Through an intelligent monitoring system, the load status of cloud platform resource machines is evaluated and resource allocation is optimized. The problem of unbalanced resource allocation of cloud platform is solved, the resource pool expansion and load optimization are achieved, and the dynamic adjustment and prediction accuracy of cloud platform resources are supported.

CN114327848BActive Publication Date: 2024-11-15SHANGHAI SHENGXIAO CLOUD COMPUTING TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202011074615.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-10-09
Publication Date
2024-11-15
Estimated Expiration
2040-10-09

AI Technical Summary

Technical Problem

The CPU load of cloud platform resource machines is too high, resulting in high pressure, cloud host upgrade failure, and the timing of resource pool expansion is difficult to define, resulting in unbalanced resource allocation.

Method used

Using a capacity expansion method based on intelligent monitoring, the resource machine monitoring data is initialized, and the cloud host CPU growth trend and configuration are calculated, and whether the remaining resources of the resource machine meet the growth needs. If it is insufficient, the migration or expansion will be triggered. Optimize resource allocation.

Benefits of technology

It realizes horizontal and vertical expansion of cloud platform resource pool, optimizes resource allocation, reduces server load, and supports dynamic adjustment and prediction accuracy of cloud platform resources.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114327848B_ABST
    Figure CN114327848B_ABST
Patent Text Reader

Abstract

The present invention provides a cloud platform expansion method based on intelligent monitoring, comprising the following steps: initializing model data, obtaining resource machine monitoring data and related label data from an intelligent monitoring system, filtering the data to remove interference data, obtaining overall load data of each resource machine, obtaining cloud host monitoring data and related label data on the resource machine, substituting the cloud host data into a policy template, calculating the CPU growth trend and upgrade situation estimate of each cloud host, summarizing the calculation results of all cloud hosts on the resource machine, and evaluating whether the remaining resources of the resource machine can meet the growth demand, etc. The present invention optimizes resource allocation, reduces server load, and continuously adds optimization policy templates based on the prediction results to make the prediction results more accurate.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to a cloud platform capacity expansion method, and in particular to a cloud platform capacity expansion method based on intelligent monitoring. Background Art

[0002] For the cloud platform resource machines, the cloud host mainly allocates CPU, memory, and disk space from the resource machines. As the overall utilization of the resource pool increases, the utilization of a single resource machine also increases, which will lead to the following problems: 1. The CPU load is too high, resulting in excessive pressure on the resource machine, which affects the operation of some cloud hosts; 2. The cloud host upgrade fails. When the load of a single resource machine is too high and the cloud host already allocated to this resource machine needs to be upgraded and expanded, if the remaining resources are not enough to meet the expansion, the upgrade fails; 3. When the resource pool expands, the platform allocates resources through a balancing algorithm combined with the actual load. As the number of cloud host applications continues to increase, it is difficult to define when the resource pool will be expanded. Summary of the invention

[0003] The purpose of the present invention is to provide a cloud platform expansion method based on intelligent monitoring, comprising the following steps:

[0004] Step 1: Initialize the model data, using labels such as cloud host configuration, cloud host running time, cloud host high load ratio, user level, etc., and determine the possibility of cloud host expansion based on the labels;

[0005] Step 2: Obtain resource machine monitoring data and related label data from the intelligent monitoring system, filter the data to remove interference data, and obtain the overall load data of each resource machine;

[0006] Step 3: Obtain the cloud host monitoring data and related label data on the resource machine;

[0007] Step 4: Compare the cloud host data with the model data, calculate the CPU growth trend and upgrade estimate of each cloud host, and determine whether it is growing. If not, proceed to step 3.

[0008] Step 5: Summarize the calculation results of all cloud hosts on the resource machine and evaluate whether the remaining resources of the resource machine can meet the growth demand; if not, trigger a resource warning; if yes, execute step 3;

[0009] Step 6: Select a cloud host for migration through resource warning. If a resource host is found in the resource pool that meets the migration requirements, call the migration module for hot migration and then execute step 7. Otherwise, a resource pool warning is triggered to remind you to expand the resource pool and then execute step 7.

[0010] Step 7: Continuously optimize model data and decision models based on the operation status of the cloud host.

[0011] The monitoring data in step 2 include CPU utilization rate, memory utilization rate, disk utilization rate, number of disk read requests, number of disk write requests, disk read rate, disk write rate, network received data, network sent data, etc.

[0012] The second step of eliminating interfering data is to eliminate values ​​that have large deviations from the previous and next values.

[0013] The fifth step of summarizing the calculation results of all cloud hosts on the resource machine refers to statistically analyzing the growth of all cloud hosts on the resource machine, summarizing and accumulating them as the resource machine's expected resource growth, including CPU, memory, and disk requirements, and combining the current remaining resources of the resource machine to determine whether the remaining resources are greater than the growth needs, otherwise they cannot meet the growth needs.

[0014] Beneficial effects achieved by the technical solution of the present invention:

[0015] The design of the present invention solves the problem of cloud platform resource allocation. The cloud platform introduces an expansion method based on intelligent monitoring to support the horizontal and vertical expansion of the cloud platform resource pool. Based on the monitoring data of the intelligent monitoring platform, resource allocation is optimized to reduce server load. At the same time, based on the prediction results, optimization strategy templates are continuously added to make the prediction results more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic diagram of a cloud platform expansion method based on intelligent monitoring of the present invention. DETAILED DESCRIPTION

[0017] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments. The embodiments of the present invention are only used to illustrate the technical solution of the present invention but not to limit the protection scope of the present invention.

[0018] like Figure 1 As shown, the present invention provides a cloud platform expansion method based on intelligent monitoring, comprising the following steps:

[0019] Step 1: Initialize model data: use cloud host configuration, cloud host running time, cloud host high load ratio, user level and other labels to determine the possibility of cloud host expansion based on the labels;

[0020] In the specific implementation, the policy template can be set as follows:

[0021] The cloud host's cpu, memory, or disk space utilization remains above 80% for more than 3 consecutive days; the disk utilization growth rate continues to be greater than 0.1, and the disk utilization is greater than 50%; the disk Iops data is above the threshold for more than 1 / 4 of the time in a day; the cloud host's daily peak value, i.e., s growth rate, is greater than 10%, and the cloud host's cpu and memory utilization exceed 60% during the corresponding time; the cloud host's cpu utilization is more than twice the memory utilization, and the cpu utilization is greater than 60%.

[0022] Step 2: Obtain resource machine monitoring data and related label data from the intelligent monitoring system, filter the data to remove interference data, and obtain the overall load data of each resource machine;

[0023] Step 3: Obtain the cloud host monitoring data and related label data on the resource machine;

[0024] Step 4: Compare the cloud host data with the model data, calculate the CPU growth trend and upgrade estimate of each cloud host, and determine whether it is growing. If not, proceed to step 3.

[0025] Step 5: Summarize the calculation results of all cloud hosts on the resource machine and evaluate whether the remaining resources of the resource machine can meet the growth demand; if not, trigger a resource warning; if yes, execute step 3;

[0026] Step 6: Select a cloud host for migration through resource warning. If a resource host is found in the resource pool that meets the migration requirements, call the migration module for hot migration and then execute step 7. Otherwise, a resource pool warning is triggered to remind you to expand the resource pool and then execute step 7.

[0027] Step 7: Continuously optimize model data and decision models based on the operation status of the cloud host.

[0028] The monitoring data in step 2 include CPU utilization rate, memory utilization rate, disk utilization rate, number of disk read requests, number of disk write requests, disk read rate, disk write rate, network received data, network sent data, etc.

[0029] The second step is to eliminate interfering data, which mainly eliminates values ​​with large deviations from the previous and next values. It mainly refers to: eliminating data mainly eliminates values ​​with large deviations from the previous and next values, which are usually abnormal data caused by acquisition or calculation deviations, to avoid interference with normal data.

[0030] The fifth step of summarizing the calculation results of all cloud hosts on the resource machine refers to counting the growth of all cloud hosts on the resource machine, summarizing and accumulating them as the resource machine's expected resource growth, including CPU, memory, and disk requirements. Combined with the current remaining resources of the resource machine, if it is judged that the remaining resources are greater than the growth needs, the growth needs can be met, otherwise the growth needs cannot be met.

[0031] The present invention is based on the monitoring data of the intelligent monitoring platform. Through the calculation of the policy template based on the monitoring data, it predicts the growth trend of cloud host resource usage and possible expansion needs, counts the results of all virtual machines on the entire resource machine, and predicts whether the remaining resources of the resource machine can meet the expansion needs. If they cannot be met, the resource allocation is dynamically adjusted through the migration scheduling module, thereby optimizing resource allocation and reducing server load. At the same time, based on the prediction results, optimization policy templates are continuously added to make the prediction results more accurate.

Claims

1. A cloud platform expansion method based on intelligent monitoring, Features The following steps are involved: Step 1: Initialize the model data, using cloud host configuration, cloud host running time, cloud host high load ratio, and user level as labels, and determine the possibility of cloud host expansion based on the labels; Step 2: Obtain resource machine monitoring data and the label data from the intelligent monitoring system, filter the data to remove interference data, and obtain overall load data of each resource machine; Step 3: Obtain the cloud host monitoring data and the label data on the resource machine; Step 4: Compare the cloud host data with the model data, calculate the CPU growth trend and upgrade estimate of each cloud host, and determine whether it is growing. If it is not, execute step 3, and if it is, execute step 5; Step 5: Summarize the calculation results of all the cloud hosts on the resource machine and evaluate whether the remaining resources of the resource machine can meet the growth demand; if not, execute step 6; if yes, execute step 3; Step 6: trigger resource warning, select the cloud host for migration, if the resource machine queried in the resource pool can meet the migration requirements, call the migration module for hot migration, and execute step 7; otherwise, trigger resource pool warning, require the resource pool to expand, and then execute step 7; Step 7: Continuously optimize the model data and decision model through the operation of the cloud host.

2. According to claim 1, a cloud platform expansion method based on intelligent monitoring, Features The monitoring data in step 2 include CPU utilization rate, memory utilization rate, disk utilization rate, number of disk read requests, number of disk write requests, disk read rate, disk write rate, network received data, and network sent data.

3. According to the cloud platform expansion method based on intelligent monitoring according to claim 1, Features The second step of eliminating interference data refers to eliminating values ​​that have a large deviation compared with the previous and next values.

4. According to claim 1, a cloud platform expansion method based on intelligent monitoring, Features The step 5 of summarizing the calculation results of all cloud hosts on the resource machine refers to counting the growth of all cloud hosts on the resource machine, summarizing and accumulating them as the expected resource growth of the resource machine, including CPU, memory, and disk requirements, and combining the current remaining resources of the resource machine to determine whether the remaining resources are greater than the growth needs, otherwise they cannot meet the growth needs.