Virtual service node performance testing method, device and computer-readable storage medium

Abstract

An embodiment of the present invention provides a virtual service node performance testing method, device and computer-readable storage medium. The method is applied to a single virtual service node, including: generating a jammer test script based on preset interference parameters to simulate the interference conditions when the virtual service node is running; under different interference conditions, determining the performance value of the virtual service node based on the statistical value of business traffic within a preset time.

Description

Technical Field

[0001] The present invention relates to the field of mobile communication technology, and in particular to a virtual service node performance testing method, device and computer-readable storage medium. Background Art

[0002] For service function nodes (SF) in traditional data centers, such as load balancers, firewalls, network address translation, etc., because they use proprietary hardware devices, when testing their performance, they are generally tested by simulating actual usage scenarios. This method is also used to test software service nodes deployed on virtual machines.

[0003] At present, there is little external interference in the environment where applications are deployed, and the deviation between the test value and the actual usage value is small. However, in a virtualized environment, business service nodes are deployed on virtual machines in the form of software. Although virtual machines are strictly isolated from each other and run together in an unknowing manner, and certain resource guarantees can be set, the performance of virtual machines is not completely isolated. All virtual machines on the same physical server actually share the underlying physical resources. When other virtual machines are not running, the ideal performance of the service node can be tested; but when other virtual machines are running, there will be serious competition for hardware resources, resulting in a decrease in the performance of the business service node, and its capacity is far from the ideal test value. As can be seen from the above, it is difficult for the relevant scheme to truly reflect the performance of the business service node under interference. Summary of the invention

[0004] In view of this, embodiments of the present invention are intended to provide a method, an apparatus, and a computer-readable storage medium for testing performance of a virtual service node.

[0005] To achieve the above object, the technical solution of the embodiment of the present invention is implemented as follows:

[0006] An embodiment of the present invention provides a virtual service node performance testing method, which is applied to a single virtual service node and includes:

[0007] Generate a jammer test script based on preset jamming parameters to simulate the jamming conditions when the virtual service node is running;

[0008] Under different interference conditions, the performance value of the virtual service node is determined based on the statistical value of the service traffic within a preset time.

[0009] Wherein, determining the performance value of the virtual service node based on the statistical value of the service traffic within a preset time under different interference conditions includes:

[0010] During the test, different interference conditions are obtained by modifying the values ​​of the preset interference parameters;

[0011] Under the corresponding interference condition, the performance value of the virtual service node under the interference condition is determined according to the statistical value of the service flow within the preset time.

[0012] The presets include but are not limited to one or more of the following:

[0013] The CPU usage of the physical server when the virtual service node is not started x1;

[0014] Storage usage on the physical server when the virtual service node is not started x2;

[0015] Input / output I / O usage on the physical server when the virtual service node is not started x3.

[0016] Optionally, the method further includes:

[0017] Based on the preset interference parameter and the performance value of the virtual service node, a performance model of the virtual service node under interference conditions is established.

[0018] Optionally, when establishing the performance model of the virtual service node under interference conditions, the method includes:

[0019] The segmented multiple linear regression model was used for fitting, and then the least square method was used to estimate the parameters in the model in each segmented interval;

[0020] Each segment interval is set according to a preset value of an interference parameter.

[0021] Wherein, when the preset interference parameters include: CPU occupancy x1, storage usage x2 and I / O usage x3 when the virtual service node on the physical server is not started, the equation corresponding to the multivariate linear regression model in each segment interval is:

[0022] Y=b0+b1x1+b2x2+b3x3+e;

[0023] Among them, Y is the performance of the virtual service node instance under interference conditions; b0 is the maximum performance value of the virtual service node instance under non-interference conditions; b1, b2, b3 are the parameters of the multivariate linear model; e is the prediction residual of the model, obtained by the least squares method.

[0024] Among them, based on the performance value of the virtual service node in the segmented interval and the equation corresponding to the multivariate linear regression model, the following matrix is ​​obtained:

[0025] Y = BX + E;

[0026] Among them, Y is an n×1 matrix, which is composed of various performance values ​​obtained by testing in the segmented interval; X is an n×3 matrix, which is composed of various interference parameters; B is the model parameter vector; E is an n×1 matrix, which is the residual term and obeys the normal distribution, and n is the number of performance values ​​in the segmented interval;

[0027] The value of B is obtained based on the above matrix, and the performance model of the virtual service node within the segment interval is obtained.

[0028] The embodiment of the present invention further provides a virtual service node performance testing device, which is applied to a single virtual service node and includes:

[0029] An interference module, used to generate a jammer test script based on preset interference parameters, used to simulate interference conditions when the virtual service node is running;

[0030] The determination module is used to determine the performance value of the virtual service node based on the statistical value of the business traffic within a preset time under different interference conditions.

[0031] The embodiment of the present invention further provides a virtual service node performance testing device, the device comprising: a processor and a memory for storing a computer program that can be run on the processor,

[0032] Wherein, the processor is used to execute the steps of the above method when running the computer program.

[0033] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the above method are implemented.

[0034] The virtual service node performance test method, device and computer-readable storage medium provided by the embodiments of the present invention generate a jammer test script based on preset interference parameters to simulate the interference conditions when the virtual service node is running; under different interference conditions, the performance value of the virtual service node is determined based on the statistical value of the business traffic within a preset time. The embodiments of the present invention are applied to a single virtual service node, and environmental interference is considered when testing the performance of the virtual service node, thereby improving the effectiveness of the business service node performance test under interference conditions, and the test value more truly reflects the performance of the business service node.

[0035] In addition, by establishing a virtual service node performance model to predict the performance of a virtual service node under certain environmental interference conditions, there is no need to traverse all interference conditions, and the availability is high. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 The virtual service node performance testing method according to the embodiment of the present invention is shown in FIG. Figure 1 ;

[0037] Figure 2 The structure of the virtual service node performance device according to the embodiment of the present invention is shown as follows Figure 1 ;

[0038] Figure 3 The structure of the virtual service node performance device according to the embodiment of the present invention is shown as follows Figure 2 ;

[0039] Figure 4 A schematic diagram of a performance test system for a virtual service node according to an embodiment of the present invention;

[0040] Figure 5 The virtual service node performance testing method process described in the embodiment of the present invention is as follows: Figure 2 ;

[0041] Figure 6 This is a schematic diagram of the virtual service node deployment structure described in an embodiment of the present invention. DETAILED DESCRIPTION

[0042] The present invention is described below in conjunction with the accompanying drawings and embodiments.

[0043] The embodiment of the present invention provides a virtual service node performance testing method, such as Figure 1 As shown, the method is applied to a single virtual service node, including:

[0044] Step 101: Generate a jammer test script based on preset jamming parameters to simulate the jamming conditions when the virtual service node is running;

[0045] Step 102: Under different interference conditions, determine the performance value of the virtual service node based on the statistical value of the service traffic within a preset time.

[0046] The embodiment of the present invention takes environmental interference into consideration when testing the performance of a virtual service node, thereby improving the effectiveness of the performance test of the business service node under interference conditions, and the test value more truly reflects the performance of the business service node.

[0047] In the embodiment of the present invention, determining the performance value of the virtual service node based on the statistical value of the service traffic within a preset time under different interference conditions includes:

[0048] During the test, different interference conditions are obtained by modifying the values ​​of the preset interference parameters;

[0049] Under corresponding interference conditions, the performance value of the virtual service node under the interference conditions is determined according to the statistical value of the service flow within a preset time.

[0050] In the embodiment of the present invention, the preset interference parameters include but are not limited to one or more of the following:

[0051] The CPU usage of the physical server when the virtual service node is not started x1;

[0052] Storage usage on the physical server when the virtual service node is not started x2;

[0053] Input / output I / O usage on the physical server when the virtual service node is not started x3.

[0054] In one embodiment, the method further comprises:

[0055] Based on the preset interference parameter and the performance value of the virtual service node, a performance model of the virtual service node under interference conditions is established.

[0056] In the embodiment of the present invention, when establishing the performance model of the virtual service node under interference conditions, the method includes:

[0057] The segmented multiple linear regression model was used for fitting, and then the least square method was used to estimate the parameters in the model in each segmented interval;

[0058] Each segment interval is set according to a preset value of an interference parameter.

[0059] In the embodiment of the present invention, when the preset interference parameters include: CPU occupancy x1, storage usage x2 and I / O usage x3 when the virtual service node on the physical server is not started, the equation corresponding to the multivariate linear regression model in each segment interval is:

[0060] Y=b0+b1x1+b2x2+b3x3+e;

[0061] Among them, Y is the performance of the virtual service node instance under interference conditions; b0 is the maximum performance value of the virtual service node instance under non-interference conditions; b1, b2, b3 are the parameters of the multivariate linear model; e is the prediction residual of the model, obtained by the least squares method.

[0062] In the embodiment of the present invention, based on the performance value of the virtual service node in the segmented interval and the equation corresponding to the multivariate linear regression model, the following matrix is ​​obtained:

[0063] Y = BX + E;

[0064] Among them, Y is an n×1 matrix, which is composed of various performance values ​​obtained by testing in the segmented interval; X is an n×3 matrix, which is composed of various interference parameters; B is the model parameter vector; E is an n×1 matrix, which is the residual term and obeys the normal distribution, and n is the number of performance values ​​in the segmented interval;

[0065] The value of B is obtained based on the above matrix, and the performance model of the virtual service node within the segment interval is obtained.

[0066] In order to implement the above method embodiment, the embodiment of the present invention also provides a virtual service node performance testing device, such as Figure 2 As shown, the device is applied to a single virtual service node, including:

[0067] The interference module 201 is used to generate a jammer test script based on preset interference parameters, so as to simulate the interference conditions when the virtual service node is running;

[0068] The determination module 202 is used to determine the performance value of the virtual service node based on the statistical value of the service traffic within a preset time under different interference conditions.

[0069] In the embodiment of the present invention, the determining module 202 determines the performance value of the virtual service node based on the statistical value of the service traffic within a preset time under different interference conditions, including:

[0070] During the test, different interference conditions are obtained by modifying the values ​​of the preset interference parameters;

[0071] Under corresponding interference conditions, the performance value of the virtual service node under the interference conditions is determined according to the statistical value of the service flow within a preset time.

[0072] In the embodiment of the present invention, the preset interference parameters include but are not limited to one or more of the following:

[0073] The CPU usage of the physical server when the virtual service node is not started x1;

[0074] Storage usage on the physical server when the virtual service node is not started x2;

[0075] Input / output I / O usage on the physical server when the virtual service node is not started x3.

[0076] In one embodiment, if Figure 3 As shown, the device further includes: an establishing module 203;

[0077] The establishing module 203 is used to establish a performance model of the virtual service node under interference conditions based on the preset interference parameters and the performance value of the virtual service node.

[0078] In the embodiment of the present invention, when the establishing module 203 establishes the performance model of the virtual service node under interference conditions, it includes:

[0079] The segmented multiple linear regression model was used for fitting, and then the least square method was used to estimate the parameters in the model in each segmented interval;

[0080] Each segment interval is set according to a preset value of an interference parameter.

[0081] In the embodiment of the present invention, when the preset interference parameters include: CPU occupancy x1, storage usage x2 and I / O usage x3 when the virtual service node on the physical server is not started, the equation corresponding to the multivariate linear regression model in each segment interval is:

[0082] Y=b0+b1x1+b2x2+b3x3+e;

[0083] Among them, Y is the performance of the virtual service node instance under interference conditions; b0 is the maximum performance value of the virtual service node instance under non-interference conditions; b1, b2, b3 are the parameters of the multivariate linear model; e is the prediction residual of the model, obtained by the least squares method.

[0084] In the embodiment of the present invention, the establishment module 203 obtains the following matrix based on the performance value of the virtual service node in the segmented interval and the equation corresponding to the multivariate linear regression model:

[0085] Y = BX + E;

[0086] Among them, Y is an n×1 matrix, which is composed of various performance values ​​obtained by testing in the segmented interval; X is an n×3 matrix, which is composed of various interference parameters; B is the model parameter vector; E is an n×1 matrix, which is the residual term and obeys the normal distribution, and n is the number of performance values ​​in the segmented interval;

[0087] The value of B is obtained based on the above matrix, and the performance model of the virtual service node within the segment interval is obtained.

[0088] The embodiment of the present invention further provides a virtual service node performance testing device, the device comprising: a processor and a memory for storing a computer program that can be run on the processor,

[0089] Wherein, when the processor is used to run the computer program, it executes:

[0090] Generate a jammer test script based on preset jamming parameters to simulate the jamming conditions when the virtual service node is running;

[0091] Under different interference conditions, the performance value of the virtual service node is determined based on the statistical value of the service traffic within a preset time.

[0092] When the performance value of the virtual service node is determined based on the statistical value of the service flow within a preset time under different interference conditions, the processor is further configured to execute, when running the computer program:

[0093] During the test, different interference conditions are obtained by modifying the values ​​of the preset interference parameters;

[0094] Under corresponding interference conditions, the performance value of the virtual service node under the interference conditions is determined according to the statistical value of the service flow within a preset time.

[0095] The preset interference parameters include but are not limited to one or more of the following:

[0096] The CPU usage of the physical server when the virtual service node is not started x1;

[0097] Storage usage on the physical server when the virtual service node is not started x2;

[0098] Input / output I / O usage on the physical server when the virtual service node is not started x3.

[0099] The processor is further configured to execute, when running the computer program:

[0100] Based on the preset interference parameter and the performance value of the virtual service node, a performance model of the virtual service node under interference conditions is established.

[0101] When the performance model of the virtual service node under interference conditions is established, the processor is further configured to execute, when running the computer program:

[0102] The segmented multiple linear regression model was used for fitting, and then the least square method was used to estimate the parameters in the model in each segmented interval;

[0103] Each segment interval is set according to a preset value of an interference parameter.

[0104] Wherein, when the preset interference parameters include: CPU occupancy x1, storage usage x2 and I / O usage x3 when the virtual service node on the physical server is not started, the equation corresponding to the multivariate linear regression model in each segment interval is:

[0105] Y=b0+b1x1+b2x2+b3x3+e;

[0106] Among them, Y is the performance of the virtual service node instance under interference conditions; b0 is the maximum performance value of the virtual service node instance under non-interference conditions; b1, b2, b3 are the parameters of the multivariate linear model; e is the prediction residual of the model, obtained by the least squares method.

[0107] Among them, based on the performance value of the virtual service node in the segmented interval and the equation corresponding to the multivariate linear regression model, the following matrix is ​​obtained:

[0108] Y = BX + E;

[0109] Among them, Y is an n×1 matrix, which is composed of various performance values ​​obtained by testing in the segmented interval; X is an n×3 matrix, which is composed of various interference parameters; B is the model parameter vector; E is an n×1 matrix, which is the residual term and obeys the normal distribution, and n is the number of performance values ​​in the segmented interval;

[0110] The value of B is obtained based on the above matrix, and the performance model of the virtual service node within the segment interval is obtained.

[0111] It should be noted that: when the device provided in the above embodiment performs a virtual service node performance test, only the division of the above program modules is used as an example. In actual applications, the above processing can be assigned to different program modules as needed, that is, the internal structure of the device is divided into different program modules to complete all or part of the processing described above. In addition, the device provided in the above embodiment and the corresponding method embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0112] In an exemplary embodiment, an embodiment of the present invention further provides a computer-readable storage medium, which may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, Flash Memory, magnetic surface memory, optical disk, or CD-ROM; or may be various devices including one or any combination of the above memories, such as a mobile phone, a computer, a tablet device, a personal digital assistant, etc.

[0113] The embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the computer program performs:

[0114] Generate a jammer test script based on preset jamming parameters to simulate the jamming conditions when the virtual service node is running;

[0115] Under different interference conditions, the performance value of the virtual service node is determined based on the statistical value of the service traffic within a preset time.

[0116] When the performance value of the virtual service node is determined based on the statistical value of the service flow within a preset time under different interference conditions, when the computer program is executed by the processor, the following is further performed:

[0117] During the test, different interference conditions are obtained by modifying the values ​​of the preset interference parameters;

[0118] Under corresponding interference conditions, the performance value of the virtual service node under the interference conditions is determined according to the statistical value of the service flow within a preset time.

[0119] The preset interference parameters include but are not limited to one or more of the following:

[0120] The CPU usage of the physical server when the virtual service node is not started x1;

[0121] Storage usage on the physical server when the virtual service node is not started x2;

[0122] Input / output I / O usage on the physical server when the virtual service node is not started x3.

[0123] When the computer program is executed by a processor, it also performs:

[0124] Based on the preset interference parameter and the performance value of the virtual service node, a performance model of the virtual service node under interference conditions is established.

[0125] When the performance model of the virtual service node under interference conditions is established, when the computer program is executed by the processor, the following steps are further performed:

[0126] The segmented multiple linear regression model was used for fitting, and then the least square method was used to estimate the parameters in the model in each segmented interval;

[0127] Each segment interval is set according to a preset value of an interference parameter.

[0128] Wherein, when the preset interference parameters include: CPU occupancy x1, storage usage x2 and I / O usage x3 when the virtual service node on the physical server is not started, the equation corresponding to the multivariate linear regression model in each segment interval is:

[0129] Y=b0+b1x1+b2x2+b3x3+e;

[0130] Among them, Y is the performance of the virtual service node instance under interference conditions; b0 is the maximum performance value of the virtual service node instance under non-interference conditions; b1, b2, b3 are the parameters of the multivariate linear model; e is the prediction residual of the model, obtained by the least squares method.

[0131] Among them, based on the performance value of the virtual service node in the segmented interval and the equation corresponding to the multivariate linear regression model, the following matrix is ​​obtained:

[0132] Y = BX + E;

[0133] Among them, Y is an n×1 matrix, which is composed of various performance values ​​obtained by testing in the segmented interval; X is an n×3 matrix, which is composed of various interference parameters; B is the model parameter vector; E is an n×1 matrix, which is the residual term and obeys the normal distribution, and n is the number of performance values ​​in the segmented interval;

[0134] The value of B is obtained based on the above matrix, and the performance model of the virtual service node within the segment interval is obtained.

[0135] The present invention is described below in conjunction with scenario embodiments.

[0136] Service-chain technology enables data packets to selectively and sequentially pass through various service nodes in the network according to business needs, providing users with flexible and stable network services. The service nodes in traditional data centers are proprietary hardware devices, which are deployed to the network through physical connections to complete various business services. The software-defined network (SDN) and network function virtualization (NFV) technologies of current cloud data center networks have led to a new development of service-chain technology. Business service nodes (such as load balancers, firewalls, network address translation, etc.) are gradually transformed from traditional proprietary hardware to virtual business resource pools built by general physical servers. Proprietary software deployed on virtual machines on physical servers provides business services, which greatly improves the flexibility of business deployment, shortens business deployment time, and makes resource utilization more efficient.

[0137] When software business service nodes are deployed on the virtual machines of the server, there will be serious competition for shared computing resources and performance interference between virtual machines, which will cause a significant drop in load performance and fluctuations, greatly affecting the service level agreement (SLA) of user load applications. Real-time understanding of the performance of business service nodes when running under interference is the key to ensuring user service quality. By accurately estimating the performance of business service nodes, the resource utilization of physical servers can be improved, thereby reducing the energy consumption of data centers.

[0138] Therefore, the embodiment of the present invention provides a performance testing method and system for a virtual service node in a service chain, and its system architecture is as follows: Figure 4 As shown, it includes: a device under test (i.e., a virtual service node), an interference sending unit for sending a jammer test script, a service flow generating unit and a service monitoring unit on the client side, and a service flow receiving unit and a service monitoring unit on the server side.

[0139] Combine the following Figure 5 The performance test flow chart of the virtual service node shown in the figure describes the application of the above system. The specific steps are as follows:

[0140] Step 1: Build a performance test topology for virtual service nodes;

[0141] like Figure 6 Is a virtual service node (i.e. Figure 4 The factors that determine the performance of the service node are its software and hardware configuration. The hardware configuration includes: the physical server hardware model and the hardware resources allocated to the service node, such as CPU, memory, and network card. The software configuration is relatively complex, which can be mainly attributed to the Host OS software and version, Hypervisor software and version, network card delivery mode (SR-IOV or OVS), network card driver version, etc. Under interference-free conditions, the performance tested with fixed software and hardware configurations is unchanged. Therefore, when setting up the test environment, simulate the actual operation scenario and use the same physical server configuration and software configuration.

[0142] Step 2: Deploy virtual service nodes;

[0143] Here, you can pull up a test virtual machine on a physical server and deploy the software business service node on it. The hardware resources allocated to the virtual machine refer to the actual operation scenario.

[0144] Step 3: Send interference test script;

[0145] Here, a test jammer can be pulled up on a physical server and an interference generation tool can be deployed on it. The corresponding jammer test script is obtained according to the set interference parameters, and the script is sent to the interference generation tool on the jammer to simulate the interference when the virtual service node is running.

[0146] The performance of business service nodes is mainly composed of three indicators: the maximum number of new connections per second, the maximum number of concurrent users per second, and the maximum throughput per second. The maximum number of new connections per second mainly consumes CPU resources, the maximum number of concurrent users per second mainly consumes storage resources (including cache and memory), and the maximum throughput per second mainly consumes network card I / O resources.

[0147] The performance interference of other virtual machines on the same physical server can be divided into three types: CPU interference, storage interference, and network card I / O interference. The CPU interference parameter is defined as the CPU occupancy rate x1 when the virtual service node on the physical server is not started, the storage interference parameter is the storage usage rate x2 when the virtual service node on the physical server is not started, and the network card I / O interference parameter is the I / O usage rate x3 when the virtual service node on the physical server is not started. The actual interference is the combined value of x1, x2 and x3. Taking the CPU occupancy rate x1 as an example, the configuration parameters of the jammer test script can be obtained, and this parameter can be sent to the jammer so that when the CPU occupancy rate of the physical server reaches the set value, the actual deployment environment of the virtual service node is fully simulated.

[0148] Step 4: Sending business traffic;

[0149] Figure 4 The service traffic generation unit shown simulates user behavior and initiates service requests of a certain stress model to the device under test. According to the service performance indicators, it can be the maximum number of new connections per second, the maximum number of concurrent users per second, the maximum throughput per second, or the superposition of the three. The service monitoring unit is connected to the system on the client and server ends to count the number of new connections, concurrent users, throughput, and service success and failure of the service processing on the node in real time.

[0150] Step 5: Stop sending business traffic;

[0151] After a certain period of continuous transmission, the traffic is stopped. According to the statistical value of the service monitoring unit during the continuous period of traffic, the maximum performance value of the service node under the interference condition is calculated.

[0152] Step 6: Modify interference parameters;

[0153] The modification of the interference parameters is to modify the CPU occupancy x1, the storage usage x2 and the I / O usage x3, and repeat steps 3 to 5 until the performance value test under all interference conditions is completed.

[0154] Step 7: Based on the interference parameters and the performance values ​​obtained from the test, a performance model of the service node under interference conditions is established.

[0155] The performance of a business node instance under interference is a nonlinear relationship with the interference. Considering the cost and accuracy of establishing the model, a piecewise multivariate linear regression model is used for fitting, and then the least squares method is used to estimate the parameters in the model in each interval. The so-called segmentation in this patent is defined by CPU occupancy x1, storage usage x2, and I / O usage x3. For example, assuming that in the interval of 30%≤CPU occupancy x1≤40%, 40%≤storage usage x2≤50%, and 40%≤I / O usage x3≤50%, the performance of a business node instance under interference and the interference can be described by a multivariate linear relationship. The smaller the strength of the segmentation, the higher the accuracy of the model, and it can be selected according to actual needs.

[0156] In each segmented interval, the linear regression equation of the application performance under interference and the runtime system characteristic parameters is established through the regression model as follows:

[0157] Y=b0+b1x1+b2x2+b3x3+e;

[0158] Among them, Y is the performance of the service node instance under interference, b0 is the maximum performance value of the service node instance without interference, b1, b2, b3 are the parameters of the multivariate linear model, and e is the prediction residual of the model. The parameters of the multivariate linear model can be calculated using the least squares method. The basic idea is to make the residual sum of squares ∑e 2 Reach the minimum.

[0159] ∑e 2 =∑(Y-b0-b1x1-b2x2-b3x3) 2

[0160] By bringing the performance values ​​obtained in the segmented interval in step 6 into the model, we can get the following matrix:

[0161] Y = BX + E;

[0162] in

[0163]

[0164] Where Y is an n×1 matrix, which is composed of the various performance values ​​obtained by testing in the segmented interval; X is an n×3 matrix, which is composed of various interference parameters (x1, x2, x3 in each segmented interval); B is the model parameter vector; E is an n×1 matrix, which is the residual term and obeys the normal distribution, and n is the number of performance values ​​in the segmented interval.

[0165] The resulting parameter estimates are:

[0166] B=(X′X) -1 X′Y;

[0167] Where X′ is the transposed matrix of X; (X′X) -1 is the inverse matrix of the matrix (X′X).

[0168] After finding the value of matrix B, the performance prediction model in the segmented interval can be determined. When in use, it is only necessary to observe the interference value of the business node instance deployment environment, that is, CPU occupancy x1, storage usage x2, and I / O usage x3, to obtain the performance of the business node instance deployed under this interference.

[0169] The embodiment of the present invention takes environmental interference into consideration when testing the performance of a virtual service node, thereby improving the effectiveness of the performance test of the business service node under interference conditions, and the test value more truly reflects the performance of the business service node.

[0170] In addition, by establishing a virtual service node performance model to predict the performance of a virtual service node under certain environmental interference conditions, there is no need to traverse all interference conditions, and the availability is high.

[0171] The above description is only a preferred embodiment of the present invention and is not intended to limit the protection scope of the present invention.

Claims

1. A virtual service node performance testing method, characterized in that: The method is applied to a single virtual service node and includes: Generate a jammer test script based on preset jamming parameters to simulate the jamming conditions when the virtual service node is running; Under different interference conditions, determining the performance value of the virtual service node based on the statistical value of the service traffic within a preset time; The method further comprises: Based on the preset interference parameter and the performance value of the virtual service node, establishing a performance model of the virtual service node under interference conditions; The determining, under different interference conditions, the performance value of the virtual service node based on the statistical value of the service traffic within a preset time, includes: During the test, different interference conditions are obtained by modifying the values ​​of the preset interference parameters; Under the corresponding interference condition, determining the performance value of the virtual service node under the interference condition according to the statistical value of the service flow within a preset time; When establishing the performance model of the virtual service node under interference conditions, the method includes: The segmented multiple linear regression model was used for fitting, and then the least square method was used to estimate the parameters in the model in each segmented interval; Each segment interval is set according to a preset value of an interference parameter.

2. The method according to claim 1, characterized in that The preset interference parameters include but are not limited to one or more of the following: The CPU usage of the physical server when the virtual service node is not started x1; Storage usage on the physical server when the virtual service node is not started x2; Input / output I / O usage on the physical server when the virtual service node is not started x3.

3. The method according to claim 1, characterized in that When the preset interference parameters include: CPU occupancy x1, storage usage x2 and I / O usage x3 when the virtual service node on the physical server is not started, the equation corresponding to the multivariate linear regression model in each segment interval is: Y=b0+b1x1+b2x2+b3x3+e; Among them, Y is the performance of the virtual service node instance under interference conditions; b0 is the maximum performance value of the virtual service node instance under non-interference conditions; b1, b2, b3 are the parameters of the multivariate linear model; e is the prediction residual of the model, obtained by the least squares method.

4. The method according to claim 3, characterized in that Based on the performance values ​​of the virtual service nodes in the segmented intervals and the equation corresponding to the multivariate linear regression model, the following matrix is ​​obtained: Y = BX + E; Among them, Y is an n×1 matrix, which is composed of various performance values ​​obtained by testing in the segmented interval; X is an n×3 matrix, which is composed of various interference parameters; B is the model parameter vector; E is an n×1 matrix, which is the residual term and obeys the normal distribution, and n is the number of performance values ​​in the segmented interval; The value of B is obtained based on the above matrix, and the performance model of the virtual service node within the segment interval is obtained.

5. A virtual service node performance testing device, characterized in that: The device is applied to a single virtual service node and includes: An interference module, used to generate a jammer test script based on preset interference parameters, used to simulate interference conditions when the virtual service node is running; A determination module, configured to determine the performance value of the virtual service node based on the statistical value of the service flow within a preset time under different interference conditions; The device further includes an establishing module for establishing a performance model of the virtual service node under interference conditions based on the preset interference parameter and the performance value of the virtual service node; The determination module is further used to obtain different interference conditions by modifying the values ​​of preset interference parameters during the test process; under the corresponding interference conditions, determine the performance value of the virtual service node under the interference conditions according to the statistical value of the business traffic within a preset time; The establishment module is also used to adopt a segmented multivariate linear regression model for fitting, and then use the least squares method to estimate the parameters in the model in each segment interval; each segment interval is set according to the value of the preset interference parameter.

6. A virtual service node performance testing device, characterized in that: The apparatus comprises: a processor and a memory for storing a computer program capable of running on the processor, Wherein, when the processor is used to run the computer program, the steps of the method described in any one of claims 1-4 are executed.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images