An optimization method, system, device and medium for dynamically adjusting QOS

Abstract

The application belongs to the field of computers, and particularly relates to an optimization method, system, device and readable storage medium for dynamically adjusting QOS. The method comprises the following steps: acquiring each performance parameter in real time or at a fixed time based on a predetermined mode; determining a basic performance parameter and giving a trend prediction value based on the basic performance parameter according to a predetermined algorithm; and issuing a control instruction or early warning information based on the trend prediction value according to a predetermined strategy. The application has the following beneficial effects: the trend prediction algorithm gives a trend prediction of imminent exhaustion, and the performance-related adjustment based on the trend prediction can effectively avoid the problem that the system-level set resource item intelligence can only be triggered when it is reached, and the problem that the performance cannot be predicted in advance, so that the stability and performance of system software, i.e. user virtual machine, can be effectively ensured.

Description

Technical Field

[0001] This invention belongs to the field of computer science, and specifically relates to an optimization method, system, device, and readable storage medium for dynamically adjusting QoS. Background Technology

[0002] Currently, basic cloud infrastructure services play a crucial role in the operation of data centers for various enterprises, government agencies, and the military. These data centers are primarily based on private or hybrid cloud architectures, relying mainly on mature virtualization software or hyper-converged infrastructure and other hardware and software products.

[0003] These software programs run virtual machines containing the client's business data, requiring stable operation to support the applications and data needed for their business. Clients typically experience peak and off-peak data usage patterns. For example, hospital users see peak patient arrival times during the day, especially in the morning, resulting in high system load. To ensure overall system stability, the product incorporates resource control technologies that limit virtual machine resources, such as CPU, memory, disk, and network card usage. These technologies restrict resource usage by the virtual machine to within set thresholds, thus ensuring overall system stability.

[0004] However, current mainstream QoS (Quality of Service) settings heavily rely on static threshold settings, making it difficult for the system to intelligently achieve optimal and stable performance. This means the system cannot effectively utilize various underlying technologies to provide better service for specified network communications, serving as a network security mechanism and a technology to address network latency and congestion issues. This concept also applies to limitations in other areas, such as disks, memory, and CPUs, which all have similar QoS concepts.

[0005] Therefore, an intelligent solution is urgently needed to address the above problems. Summary of the Invention

[0006] To address the above problems, this invention proposes a method for dynamically adjusting QoS, comprising:

[0007] Various performance parameters are acquired in real time or at scheduled intervals based on a predetermined method;

[0008] Determine the basic performance parameters and, based on these parameters, provide trend prediction values ​​according to a predetermined algorithm;

[0009] Based on the predicted trend values, control instructions or early warning information are issued according to a predetermined strategy.

[0010] In some embodiments of the present invention, acquiring various performance parameters in real time or at regular intervals based on a predetermined method includes:

[0011] Install a testing tool in the virtual machine to monitor various performance parameters of the virtual machine system in real time;

[0012] The management node receives various performance parameters sent by the detection tool or actively polls each virtual machine to obtain the corresponding performance parameters of the virtual machine.

[0013] In some embodiments of the present invention, determining basic performance parameters and providing trend prediction values ​​based on the basic performance parameters according to a predetermined algorithm includes:

[0014] The basic performance parameters are determined according to a predetermined method, and a dynamic early warning threshold is given based on the basic performance parameters using a quadratic exponential smoothing algorithm.

[0015] In some embodiments of the present invention, the dynamic early warning threshold is given based on basic performance parameters using a quadratic exponential smoothing algorithm, including:

[0016] The effective dataset is obtained by traversing back in reverse order to remove all performance parameters before the most recent steady-state point, and then the effective dataset is used to predict the dynamic early warning threshold based on the quadratic exponential smoothing algorithm.

[0017] In some embodiments of the present invention, the method further includes:

[0018] Determine whether the performance parameters at the current time point have changed abruptly compared to the performance parameters at the previous time point. If no abrupt change has occurred, the current time point is taken as the steady state point.

[0019] In some embodiments of the present invention, determining whether the performance parameters at the current time point have changed abruptly compared to the performance parameters at the previous time point includes:

[0020] Determine whether the changes in the performance parameters at the current time point compared to the previous time point exceed the predetermined values. If the changes exceed the predetermined values, then the current time point is designated as the point of sudden change.

[0021] In some embodiments of the present invention, the method of providing a dynamic early warning threshold based on basic performance parameters using a quadratic exponential smoothing algorithm further includes:

[0022] The weighting parameter values ​​for the quadratic exponential smoothing algorithm are determined by comparing the variance of the complete dataset with the variance of the tailed data.

[0023] Another aspect of the present invention proposes an optimization system for dynamically adjusting QoS, comprising:

[0024] The monitoring module is configured to acquire various performance parameters in real time or at regular intervals based on a predetermined method.

[0025] The prediction module is configured to determine basic performance parameters and provide trend prediction values ​​based on the basic performance parameters according to a predetermined algorithm.

[0026] An adjustment module is configured to issue control instructions or early warning information based on the trend prediction value according to a predetermined strategy.

[0027] Another aspect of the present invention provides a computer device comprising:

[0028] At least one processor; and

[0029] A memory storing computer instructions executable on the processor, which, when executed by the processor, implement the steps of any of the methods described in the above embodiments.

[0030] Another aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of any of the methods described in the above embodiments.

[0031] The present invention has the following beneficial effects: by providing a trend prediction algorithm to predict when resources will be exhausted, and making performance-related adjustments based on the trend prediction, it can effectively avoid the problem that system-level resource items can only be triggered when they are reached, and cannot be predicted in advance. This can effectively ensure the stability and performance of the system software, i.e., the user's virtual machine. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 A flowchart illustrating an optimization method for dynamically adjusting QoS provided in an embodiment of the present invention;

[0034] Figure 2 This is a schematic diagram of the structure of an optimization system for dynamically adjusting QoS provided in an embodiment of the present invention;

[0035] Figure 3 A schematic diagram of the structure of a computer device provided in an embodiment of the present invention;

[0036] Figure 4This is a schematic diagram of the structure of a computer-readable storage medium provided in an embodiment of the present invention. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to specific examples and the accompanying drawings.

[0038] This invention aims to solve the problem of virtual machine resource configuration in a cluster. Currently, most traditional IaaS (Infrastructure as a Service, the lowest layer of cloud services, primarily providing basic resources) layer software implements QoS restrictions on different resources of business virtual machines based on their different operating system kernels. However, this restriction only applies to a single resource of a single virtual machine, such as a network card. While a limit can be manually set for the receive / transmit rate through an interface or command line based on system conditions, ensuring that the virtual machine's rate does not exceed the manually entered threshold during operation, this ensures stable system operation for business purposes.

[0039] However, this approach also has significant drawbacks. It requires precise configuration of specific virtual machine resources based on real-time system load and the load of the business virtual machines. If multiple resources of a business virtual machine need to be restricted, multiple modifications or parameter specification steps are required. Business virtual machines only support static QoS settings for individual resources and cannot dynamically and intelligently adjust QoS, consuming significant operational manpower and being prone to inaccurate configuration. This is a common pain point for similar products in the industry.

[0040] like Figure 1 As shown, to solve the above problems, this invention proposes an optimization method for dynamically adjusting QoS, including:

[0041] Step S1: Acquire various performance parameters in real time or at scheduled intervals based on a predetermined method;

[0042] Step S2: Determine the basic performance parameters and, based on the basic performance parameters, provide trend prediction values ​​according to a predetermined algorithm;

[0043] Step S3: Based on the trend prediction value, issue control instructions or early warning information according to a predetermined strategy.

[0044] In this embodiment, the performance parameters in step S1 include parameters related to CPU, memory, storage, and network. Each performance parameter includes the current values ​​of the CPU, memory, storage, and network parameters for each virtual machine, as well as the current values ​​of the overall CPU, memory, storage, and network parameters for the server containing multiple virtual machines.

[0045] The performance parameters of servers and virtual machines can be obtained by either polling the management node to the servers and virtual machines or by actively pushing them to the management node from each server and virtual machine.

[0046] In step S2, the performance parameters obtained in step S1 are input into the corresponding prediction algorithm to obtain trend prediction values ​​based on the current performance parameters. The trend prediction values ​​represent the changing trends of the performance parameters over multiple previous time points. In other words, it indicates approximately how much predetermined time or sampling time is needed for each parameter to reach the system-set threshold.

[0047] Furthermore, in step S3, after obtaining the trend prediction value in step S2, the time to reach the alarm threshold is calculated based on the trend prediction value and compared with the predetermined time. Assuming that based on the current trend prediction value, the storage space will reach the set alarm limit within 3 days, a corresponding warning message is issued. For adjustable performance parameters such as CPU and memory, instructions can be appropriately sent to the cloud platform management system to release pre-reserved memory or trigger memory bubbles to release more available memory, thereby optimizing performance parameters.

[0048] In some embodiments of the present invention, acquiring various performance parameters in real time or at regular intervals based on a predetermined method includes:

[0049] Install a testing tool in the virtual machine to monitor various performance parameters of the virtual machine system in real time;

[0050] The management node receives various performance parameters sent by the detection tool or actively polls each virtual machine to obtain the corresponding performance parameters of the virtual machine.

[0051] In this embodiment, it is recommended to use the Ganglia detection tool to monitor various performance parameters of the server or virtual machine.

[0052] The management node can receive various performance parameters pushed by Ganglia testing tools, and at the same time...

[0053] In some embodiments of the present invention, determining basic performance parameters and providing trend prediction values ​​based on the basic performance parameters according to a predetermined algorithm includes:

[0054] The basic performance parameters are determined according to a predetermined method, and a dynamic early warning threshold is given based on the basic performance parameters using a quadratic exponential smoothing algorithm.

[0055] In this embodiment, the basic performance parameters refer to reference values ​​under good operating conditions for the server platform. That is, performance data from the initial period of good operation is used as a benchmark. Based on experience with similar products, a performance drop of less than 30% is expected, and specific resource-related alarms and analyses are provided. As the documentation states, performance is generally good in the initial stages of product deployment and operation. In other words, from initial deployment until overall resource utilization is between 50%, this can be considered a good performance benchmark.

[0056] In some embodiments of the present invention, the dynamic early warning threshold is given based on basic performance parameters using a quadratic exponential smoothing algorithm, including:

[0057] The effective dataset is obtained by traversing back in reverse order to remove all performance parameters before the most recent steady-state point, and then the effective dataset is used to predict the dynamic early warning threshold based on the quadratic exponential smoothing algorithm.

[0058] In some embodiments of the present invention, the method further includes:

[0059] Determine whether the performance parameters at the current time point have changed abruptly compared to the performance parameters at the previous time point. If no abrupt change has occurred, the current time point is taken as the steady state point.

[0060] In some embodiments of the present invention, determining whether the performance parameters at the current time point have changed abruptly compared to the performance parameters at the previous time point includes:

[0061] Determine whether the changes in the performance parameters at the current time point compared to the previous time point exceed the predetermined values. If the changes exceed the predetermined values, then the current time point is designated as the point of sudden change.

[0062] In some embodiments of the present invention, the method of providing a dynamic early warning threshold based on basic performance parameters using a quadratic exponential smoothing algorithm further includes:

[0063] The weighting parameter values ​​for the quadratic exponential smoothing algorithm are determined by comparing the variance of the complete dataset with the variance of the tailed data.

[0064] In this embodiment, quadratic exponential smoothing is used instead of the mainstream linear fitting algorithm, which solves the problem of inaccuracy of the algorithm after a sudden change from one steady state to another. However, it does not completely avoid interference from similar problems, so further optimization of the algorithm is needed: the data before the most recent steady state is removed and no longer used as the basis for predicting the trend of subsequent data, which better avoids the unreasonable data that may be caused by the sudden change of recent data.

[0065] Furthermore, the criteria for identifying abrupt change points are that the change relative to the previous time point does not exceed 1%, the criteria for steady-state points are that there is no change compared to the previous time point, and the elimination logic is a reverse traversal, where data before the steady-state point is eliminated and no longer included in the valid dataset. Generally, cloud products using trend prediction algorithms employ linear fitting algorithms, which cannot solve the problem of inaccurate budgets caused by abrupt changes in steady-state conditions. Therefore, quadratic exponential smoothing is introduced. However, directly introducing this method also has many drawbacks. The optimization points of this solution are:

[0066] 1. Criteria for identifying a sudden change point: The change in the current time point relative to the previous time point exceeds 1%;

[0067] 2. Criteria for determining a steady state: The current time point has not changed abruptly relative to the previous time point;

[0068] 3. Sudden change elimination logic: Perform reverse traversal, eliminate data before the most recent steady-state time point to obtain the effective dataset, and then use the quadratic exponential smoothing algorithm for prediction;

[0069] 4. Determination of weighting parameters in the quadratic exponential smoothing algorithm: For the effective dataset, let the variance of the complete dataset be a and the variance of the tail-removed dataset be b. When a÷b>0.05 (indicating that the last point of the effective dataset is relatively active, in order to better connect the starting point of the predicted data with the current true value, the weighting parameter is 0.8, otherwise the weighting parameter is 0.05).

[0070] Example:

[0071] Mature IaaS layer software requires a system-level monitoring system. This invention recommends using the Ganglia framework for system-level monitoring. It offers rich plugins and powerful functionality, employing a lightweight multicast listening / announcement protocol. It provides mature interfaces, allowing for simple configuration of monitoring sampling periods and other settings for all nodes in the cluster, making it suitable for large-scale systems. Ganglia separates polling from data storage and data direction, ensuring each host is an active participant in the cluster. This organically distributes the workload, allowing Ganglia to achieve more functionality with only minor configuration changes to each host compared to traditional monitoring systems.

[0072] Ganglia has its services running on the system cluster nodes and can send and receive monitoring data from within the host via the gmond service. Unlike other monitoring proxies, it doesn't wait for data monitoring requests from an external polling engine or upload directly to a centralized poller; instead, it flexibly polls according to its own locally configured settings.

[0073] Currently, the common practice for IaaS layer software to impose system-level QoS restrictions is to statically set thresholds for each individual resource, such as the overall CPU and memory utilization not exceeding 80%, and these thresholds are applied synchronously to each individual host. If the utilization of a single host's related resources exceeds the set threshold, a system-level alarm is triggered to warn the user that there is a risk to the system.

[0074] In this embodiment, more granular monitoring and analysis are performed on the collected or pushed monitoring data. The separation of data collection and monitoring service reception at the management node reduces the system pressure from periodic polling and allows for more timely updates of monitoring information. Traditional solutions, such as polling the monitoring system every 30 seconds to update the data, risk data delays within those 30 seconds. Furthermore, for cloud platforms with large amounts of software monitoring information, a single polling update places significant pressure on the overall system, potentially impacting system stability. This embodiment employs a separate design, with both management node active data collection and monitored node active push notifications. This separation effectively addresses the issues of untimely data updates and high system pressure.

[0075] In the initial stages of implementation, the performance and response speed of products such as servers are generally quite good. Based on this, and combined with trend prediction techniques and quadratic exponential smoothing, more dynamic early warnings can be provided.

[0076] The system no longer sets basic performance parameters for similar resources through system settings. After system implementation, performance data from the initial period of good operation is used as a benchmark. Based on experience with similar products, performance degradation is less than 30%, and specific resource-related alarms and analyses are provided. For example, in the initial stage of system deployment, performance is generally good; that is, from initial deployment until overall resource utilization is between 50%, this can be considered a good performance benchmark. Furthermore, to address the time-sensitive needs of industries such as hospitals and finance, a quadratic exponential smoothing algorithm is embedded to provide potential risks associated with increased resource usage over the next 3 to 7 days.

[0077] Using quadratic exponential smoothing instead of the mainstream linear fitting algorithm solves the problem of algorithm inaccuracy after a sudden change from one steady state to another. However, it does not completely avoid interference from similar problems, so further optimization of the algorithm is needed: in this embodiment, the data before the most recent steady state is removed and no longer used as the basis for predicting the trend of subsequent data, which better avoids the unreasonable data that may be caused by the sudden change of recent data.

[0078] Furthermore, the criteria for identifying abrupt change points are that the change relative to the previous time point does not exceed 1%, the criteria for steady-state points are that there is no change compared to the previous time point, and the elimination logic is a reverse traversal, where data before the steady-state point is eliminated and no longer included in the valid dataset. Generally, cloud products using trend prediction algorithms employ linear fitting algorithms, which cannot solve the problem of inaccurate budgets caused by abrupt changes in steady-state conditions. Therefore, quadratic exponential smoothing is introduced. However, directly introducing this method also has many drawbacks. The optimization points of this solution are:

[0079] 1. Criteria for identifying a sudden change point: The change in the current time point relative to the previous time point exceeds 1%;

[0080] 2. Criteria for determining a steady state: The current time point has not changed abruptly relative to the previous time point;

[0081] 3. Sudden change elimination logic: Perform reverse traversal, eliminate data before the most recent steady-state time point to obtain the effective dataset, and then use the quadratic exponential smoothing algorithm for prediction;

[0082] 4. Determination of weighting parameters in the quadratic exponential smoothing algorithm: For the effective dataset, let the variance of the complete dataset be a and the variance of the tail-removed dataset be b. When a÷b>0.05 (indicating that the last point of the effective dataset is relatively active, in order to better connect the starting point of the predicted data with the current true value, the weighting parameter is 0.8, otherwise the weighting parameter is 0.05).

[0083] Furthermore, the status of key system physical resources such as CPU, memory, disk, and network can be accurately acquired, monitored, and predicted, providing more detailed input data for the next step of dynamic adjustment of virtual machine resources.

[0084] Intelligent adjustment of QoS limits for various resource items of the business virtual machine:

[0085] In the editing of business virtual machines, the mainstream design also provides a corresponding QoS setting edit box in the corresponding resource tab, allowing users to statically input specific QoS values. This design addresses the drawback of static input leading to insufficient resource utilization.

[0086] The business virtual machine operates on a maximum capacity limited by the host's remaining processing power. It installs tools internally and updates relevant resource utilization and other monitoring information promptly through proactive push notifications and system data collection. The internal tools installation enables proactive push notifications.

[0087] If a single resource of a business virtual machine is nearing its limit or becoming a bottleneck, aside from the resource QoS limits automatically applied by the monitoring system, which will not affect the overall system stability, other resource items can be dynamically adjusted to ensure the virtual machine's business operations as much as possible. For example, when CPU resource utilization is high, memory usage is generally also high. Memory reservation can be dynamically triggered to pre-allocate memory or memory bubbles can be triggered to release more available memory, thus dynamically adjusting memory-related QoS to ensure the normal operation of the business virtual machine. (The QoS limits applied by the monitoring system refer to the various resource thresholds set in our system, which can be directly entered and configured.)

[0088] The dynamic adjustment method is the core of our invention. It makes timely adjustments based on accurate predictions, rather than waiting for resource utilization to reach a set threshold (at which point the system load will be relatively high and the overall performance will be affected). Simply put, if the system predicts that resource utilization will increase in a short period of time and that the overall product performance will decline significantly, it will make corresponding adjustments.

[0089] The QoS restrictions that the monitoring system is in effect refer to the resource thresholds set by various settings in our system, which can be directly entered into the settings;

[0090] The dynamic adjustment method makes timely adjustments based on accurate predictions, rather than waiting for resource utilization to reach a set threshold (at which point the system load will be relatively high and overall performance will be affected). Simply put, if the system predicts that resource utilization will increase in a short period of time and that overall product performance will decline significantly, it will make corresponding adjustments.

[0091] The present invention has the following beneficial effects: by providing a trend prediction algorithm to predict when resources will be exhausted, and making performance-related adjustments based on the trend prediction, it can effectively avoid the problem that system-level resource items can only be triggered when they are reached, and cannot be predicted in advance. This can effectively ensure the stability and performance of the system software, i.e., the user's virtual machine.

[0092] like Figure 2 As shown, another aspect of the present invention proposes an optimization system for dynamically adjusting QoS, comprising:

[0093] Monitoring module 1 is configured to acquire various performance parameters in real time or at regular intervals based on a predetermined method;

[0094] Prediction module 2, configured to determine basic performance parameters and provide trend prediction values ​​based on the basic performance parameters according to a predetermined algorithm;

[0095] Adjustment module 3 is configured to issue control instructions or early warning information based on the trend prediction value according to a predetermined strategy.

[0096] In some embodiments of the present invention, the monitoring module 1 is further configured to:

[0097] Install a testing tool in the virtual machine to monitor various performance parameters of the virtual machine system in real time;

[0098] The management node receives various performance parameters sent by the detection tool or actively polls each virtual machine to obtain the corresponding performance parameters of the virtual machine.

[0099] In some embodiments of the present invention, the prediction module 2 is further configured to:

[0100] The basic performance parameters are determined according to a predetermined method, and a dynamic early warning threshold is given based on the basic performance parameters using a quadratic exponential smoothing algorithm.

[0101] In some embodiments of the present invention, the dynamic early warning threshold is given based on basic performance parameters using a quadratic exponential smoothing algorithm, including:

[0102] The effective dataset is obtained by traversing back in reverse order to remove all performance parameters before the most recent steady-state point, and then the effective dataset is used to predict the dynamic early warning threshold based on the quadratic exponential smoothing algorithm.

[0103] In some embodiments of the present invention, it further includes:

[0104] Determine whether the performance parameters at the current time point have changed abruptly compared to the performance parameters at the previous time point. If no abrupt change has occurred, the current time point is taken as the steady state point.

[0105] In some embodiments of the present invention, the prediction module 2 is further configured to:

[0106] Determine whether the changes in the performance parameters at the current time point compared to the previous time point exceed the predetermined values. If the changes exceed the predetermined values, then the current time point is designated as the point of sudden change.

[0107] In some embodiments of the present invention, the prediction module 2 is further configured to:

[0108] The weighting parameter values ​​for the quadratic exponential smoothing algorithm are determined by comparing the variance of the complete dataset with the variance of the tailed data.

[0109] like Figure 3 As shown, another aspect of the present invention also provides a computer device comprising:

[0110] At least one processor 21; and

[0111] Memory 22 stores computer instructions 23 executable on processor 21. When executed by processor 21, the instructions 23 implement an optimization method for dynamically adjusting QoS, including:

[0112] Various performance parameters are acquired in real time or at scheduled intervals based on a predetermined method;

[0113] Determine the basic performance parameters and, based on these parameters, provide trend prediction values ​​according to a predetermined algorithm;

[0114] Based on the predicted trend values, control instructions or early warning information are issued according to a predetermined strategy.

[0115] In some embodiments of the present invention, acquiring various performance parameters in real time or at regular intervals based on a predetermined method includes:

[0116] Install a testing tool in the virtual machine to monitor various performance parameters of the virtual machine system in real time;

[0117] The management node receives various performance parameters sent by the detection tool or actively polls each virtual machine to obtain the corresponding performance parameters of the virtual machine.

[0118] In some embodiments of the present invention, determining basic performance parameters and providing trend prediction values ​​based on the basic performance parameters according to a predetermined algorithm includes:

[0119] The basic performance parameters are determined according to a predetermined method, and a dynamic early warning threshold is given based on the basic performance parameters using a quadratic exponential smoothing algorithm.

[0120] In some embodiments of the present invention, the dynamic early warning threshold is given based on basic performance parameters using a quadratic exponential smoothing algorithm, including:

[0121] The effective dataset is obtained by traversing back in reverse order to remove all performance parameters before the most recent steady-state point, and then the effective dataset is used to predict the dynamic early warning threshold based on the quadratic exponential smoothing algorithm.

[0122] In some embodiments of the present invention, the method further includes:

[0123] Determine whether the performance parameters at the current time point have changed abruptly compared to the performance parameters at the previous time point. If no abrupt change has occurred, the current time point is taken as the steady state point.

[0124] In some embodiments of the present invention, determining whether the performance parameters at the current time point have changed abruptly compared to the performance parameters at the previous time point includes:

[0125] Determine whether the changes in the performance parameters at the current time point compared to the previous time point exceed the predetermined values. If the changes exceed the predetermined values, then the current time point is designated as the point of sudden change.

[0126] In some embodiments of the present invention, the method of providing a dynamic early warning threshold based on basic performance parameters using a quadratic exponential smoothing algorithm further includes:

[0127] The weighting parameter values ​​for the quadratic exponential smoothing algorithm are determined by comparing the variance of the complete dataset with the variance of the tailed data.

[0128] like Figure 4 As shown, another aspect of the present invention provides a computer-readable storage medium 401, which stores a computer program 402. When executed by a processor, the computer program 402 implements an optimization method for dynamically adjusting QoS, comprising:

[0129] Various performance parameters are acquired in real time or at scheduled intervals based on a predetermined method;

[0130] Determine the basic performance parameters and, based on these parameters, provide trend prediction values ​​according to a predetermined algorithm;

[0131] Based on the predicted trend values, control instructions or early warning information are issued according to a predetermined strategy.

[0132] In some embodiments of the present invention, acquiring various performance parameters in real time or at regular intervals based on a predetermined method includes:

[0133] Install a testing tool in the virtual machine to monitor various performance parameters of the virtual machine system in real time;

[0134] The management node receives various performance parameters sent by the detection tool or actively polls each virtual machine to obtain the corresponding performance parameters of the virtual machine.

[0135] In some embodiments of the present invention, determining basic performance parameters and providing trend prediction values ​​based on the basic performance parameters according to a predetermined algorithm includes:

[0136] The basic performance parameters are determined according to a predetermined method, and a dynamic early warning threshold is given based on the basic performance parameters using a quadratic exponential smoothing algorithm.

[0137] In some embodiments of the present invention, the dynamic early warning threshold is given based on basic performance parameters using a quadratic exponential smoothing algorithm, including:

[0138] The effective dataset is obtained by traversing back in reverse order to remove all performance parameters before the most recent steady-state point, and then the effective dataset is used to predict the dynamic early warning threshold based on the quadratic exponential smoothing algorithm.

[0139] In some embodiments of the present invention, the method further includes:

[0140] Determine whether the performance parameters at the current time point have changed abruptly compared to the performance parameters at the previous time point. If no abrupt change has occurred, the current time point is taken as the steady state point.

[0141] In some embodiments of the present invention, determining whether the performance parameters at the current time point have changed abruptly compared to the performance parameters at the previous time point includes:

[0142] Determine whether the changes in the performance parameters at the current time point compared to the previous time point exceed the predetermined values. If the changes exceed the predetermined values, then the current time point is designated as the point of sudden change.

[0143] In some embodiments of the present invention, the method of providing a dynamic early warning threshold based on basic performance parameters using a quadratic exponential smoothing algorithm further includes:

[0144] The weighting parameter values ​​for the quadratic exponential smoothing algorithm are determined by comparing the variance of the complete dataset with the variance of the tailed data.

[0145] Finally, it should be noted that those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc. The embodiments of the computer program can achieve the same or similar effects as any of the corresponding foregoing method embodiments.

[0146] Furthermore, typically, the devices and equipment disclosed in the embodiments of this invention can be various electronic terminal devices, such as mobile phones, personal digital assistants (PDAs), tablet computers (PADs), smart TVs, etc., or they can be large terminal devices, such as servers. Therefore, the scope of protection disclosed in the embodiments of this invention should not be limited to a specific type of device or equipment. The client disclosed in the embodiments of this invention can be applied to any of the above-mentioned electronic terminal devices in the form of electronic hardware, computer software, or a combination of both.

[0147] Furthermore, the method disclosed in the embodiments of the present invention can also be implemented as a computer program executed by a CPU, which may be stored in a computer-readable storage medium. When the computer program is executed by the CPU, it performs the functions defined in the method disclosed in the embodiments of the present invention.

[0148] Furthermore, the above-described method steps and system units can also be implemented using a controller and a computer-readable storage medium for storing a computer program that enables the controller to perform the functions of the above-described steps or units.

[0149] Furthermore, it should be understood that the computer-readable storage medium (e.g., memory) described herein can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. By way of example, and not limitation, non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM), which can act as external cache memory. By way of example, and not limitation, RAM may be available in various forms, such as synchronous RAM (DRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), and direct Rambus RAM (DRRAM). The storage devices disclosed herein are intended to include, but are not limited to, these and other suitable types of memory.

[0150] Those skilled in the art will also understand that the various exemplary logic blocks, modules, circuits, and algorithm steps described in conjunction with the disclosure herein can be implemented as electronic hardware, computer software, or a combination of both. To clearly illustrate this interchangeability between hardware and software, the functionality of the various illustrative components, blocks, modules, circuits, and steps has been generally described. Whether this functionality is implemented as software or as hardware depends on the specific application and the design constraints imposed on the overall system. Those skilled in the art can implement the described functionality in various ways for each specific application, but such implementation decisions should not be construed as departing from the scope of the embodiments disclosed herein.

[0151] The various exemplary logic blocks, modules, and circuits described herein can be implemented or performed using the following components designed to perform the functions described herein: general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic, discrete hardware components, or any combination of these components. A general-purpose processor may be a microprocessor, but alternatively, the processor may be any conventional processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors combined with a DSP, and / or any other such configuration.

[0152] The steps of the methods or algorithms described herein can be directly incorporated into hardware, into a software module executed by a processor, or a combination of both. The software module can reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor, enabling the processor to read information from or write information to the storage medium. In an alternative, the storage medium can be integrated with the processor. The processor and storage medium can reside in an ASIC. The ASIC can reside in a user terminal. In an alternative, the processor and storage medium can reside as discrete components in the user terminal.

[0153] In one or more exemplary designs, the functionality may be implemented in hardware, software, firmware, or any combination thereof. If implemented in software, the functionality may be stored as one or more instructions or code on or transmitted via a computer-readable medium. A computer-readable medium includes computer storage media and communication media, including any medium that facilitates the transfer of a computer program from one location to another. A storage medium may be any available medium accessible to a general-purpose or special-purpose computer. By way of example, and not limitation, the computer-readable medium may include RAM, ROM, EEPROM, CD-ROM or other optical disc storage devices, disk storage devices or other magnetic storage devices, or any other medium that may be used to carry or store the required program code in the form of instructions or data structures and is accessible to a general-purpose or special-purpose computer or a general-purpose or special-purpose processor. Furthermore, any connection may be appropriately referred to as a computer-readable medium. For example, if software is transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, then the aforementioned coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are all included in the definition of a medium. As used herein, disks and optical discs include compact discs (CDs), laser discs, optical discs, digital multifunction discs (DVDs), floppy disks, and Blu-ray discs, where disks typically reproduce data magnetically, while optical discs reproduce data optically using lasers. Combinations of the above should also be included within the scope of computer-readable media.

Claims

1. A method for dynamically adjusting QoS, characterized in that, include: Various performance parameters are acquired in real time or at scheduled intervals based on a predetermined method; Determine basic performance parameters and give trend prediction values ​​based on basic performance parameters according to a predetermined algorithm, including: determining basic performance parameters according to a predetermined method, and giving dynamic early warning thresholds based on basic performance parameters through a quadratic exponential smoothing algorithm. Specifically, the effective dataset is obtained by removing all performance parameters before the most recent steady state point through a reverse traversal method, and the dynamic early warning threshold is predicted based on the quadratic exponential smoothing algorithm using the effective dataset. Determine whether the performance parameters at the current time point have changed abruptly compared to the performance parameters at the previous time point, including: determining whether the change in the performance parameters at the current time point exceeds a predetermined value; if it exceeds the predetermined value, then the current time point is taken as the point of abrupt change. In response to the absence of sudden changes, the current time points are taken as steady state points; Based on the predicted trend values, control instructions or early warning information are issued according to a predetermined strategy.

2. The method according to claim 1, characterized in that, The acquisition of various performance parameters in real time or at regular intervals based on a predetermined method includes: Install a testing tool in the virtual machine to monitor various performance parameters of the virtual machine system in real time; The management node receives various performance parameters sent by the detection tool or actively polls each virtual machine to obtain the corresponding performance parameters of the virtual machine.

3. The method according to claim 1, characterized in that, The method of providing a dynamic early warning threshold based on basic performance parameters using a quadratic exponential smoothing algorithm also includes: The weighting parameter values ​​for the quadratic exponential smoothing algorithm are determined by comparing the variance of the complete dataset with the variance of the tailed data.

4. An optimization system for dynamically adjusting QoS, characterized in that, include: The monitoring module is configured to acquire various performance parameters in real time or at regular intervals based on a predetermined method. The prediction module is configured to determine basic performance parameters and provide trend prediction values ​​based on the basic performance parameters according to a predetermined algorithm. An adjustment module is configured to issue control commands or early warning information based on the trend prediction value according to a predetermined strategy. The prediction module is further used to: determine basic performance parameters according to a predetermined method, and give a dynamic early warning threshold based on the basic performance parameters through a quadratic exponential smoothing algorithm. Specifically, the effective dataset is obtained by removing all performance parameters before the most recent steady state point through a reverse traversal method, and the dynamic early warning threshold is predicted based on the quadratic exponential smoothing algorithm using the effective dataset. And a module for the following functions: determining whether the performance parameters at the current time point have changed drastically compared to the performance parameters at the previous time point, including: determining whether the change in the performance parameters at the current time point compared to the performance parameters at the previous time point exceeds a predetermined value, and in response to exceeding the predetermined value, taking the current time point as the point of sudden change; In response to the absence of sudden changes, the current time points are taken as steady state points.

5. A computer device, characterized in that, include: At least one processor; as well as A memory storing computer instructions executable on the processor, which, when executed by the processor, implement the steps of the method according to any one of claims 1-3.

6. A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method according to any one of claims 1-3.

Images