CN102929693B - Performance evaluation method and device for servers of whole equipment cabinet - Google Patents

Abstract

The invention relates to the communication field and discloses a performance evaluation method and a device for servers of the whole equipment cabinet. The evaluation method comprises the steps that sever call commands are transmitted to virtual servers in any one load unit of the evaluated servers of the whole equipment cabinet, and the server call command comprises information of tasks to be respectively executed by the virtual servers, and the load units are in one-to-one correspondence with users; and the actual execution result information of the tasks executed by the virtual servers to the tasks is obtained, and the performance of the load units and the virtual servers are determined according to the actual execution result information of the tasks. the performance parameters of the servers of the whole equipment cabinet, obtained by the technical scheme, is more close to user application scenarios, and the obtained evaluation results are more accurate and further accord with actual application.

Description

The performance evaluation methodology of whole machine cabinet server and device

Technical field

The present invention relates to Server performance field, particularly a kind of performance evaluation methodology of whole machine cabinet server and device.

Background technology

Along with the development of computer hardware, powerful server becomes main flow.And with the development that computer technology solution and cloud computing technology are applied, just expedite the emergence of functional powerful whole machine cabinet server.

Particularly, whole machine cabinet server is, the underlying hardware frameworks such as conformity calculation node (also known as S-Blade), disk storage cabinet, network structure and a kind of integrating server formed.In this whole machine cabinet server, pass through Intel Virtualization Technology, multiple independently virtual server can be created in a physical server, although the underlying physical resource of each virtual server (such as disk storage cabinet etc.) be shared, but each virtual server is operationally completely independent, each virtual server can perform independently operating system and independently application software respectively, each independently virtual server provides different services to user independently of one another, user side seem just as multiple " hardware " server jointly for this user provide service.

The application of whole machine cabinet server improves the utilization rate of traditional physical server greatly, and is conducive to the abundant application of Internet resources and integrated.

Carrying out in research process of the present invention, the present inventor finds that prior art at least exists following defect:

Although whole machine cabinet server has obtained the accreditation in market, support that the performance evaluating technology of the performance for assessment of this whole machine cabinet server of whole machine cabinet Server development and marketing is but current difficult point.

At present traditional performance evaluating instrument is still generally adopted to the performance evaluating of whole machine cabinet server, such as: VMmark (Virtual Machine mark), SPECvirt (Standard PerformanceEvaluation Corporation, standard performance appraisal agency) etc. evaluate and test, and traditional performance evaluating is the evaluating tool traditional physical server being carried out to performance evaluating, and each virtual server in whole machine cabinet server may for belonging on a physical hardware independently to the virtualized server that user provides real network to serve, the performance evaluating demand that cannot meet current whole machine cabinet server to Server performance instrument far away traditional at present, this complete machine equipment cabinet server cannot be obtained when practical application to the service performance that user provides.

Summary of the invention

The embodiment of the present invention first object is the performance evaluation methodology providing a kind of whole machine cabinet server, applies the whole machine cabinet server performance parameter that this technical scheme obtains, application scenarios of being more close to the users, and the evaluation result obtained is more accurate, more realistic application.

The embodiment of the present invention second object is the performance evaluating apparatus providing the first whole machine cabinet server, apply the whole machine cabinet server performance parameter that this technical scheme obtains, more be close to the users application scenarios, the evaluation result obtained is more accurate, more realistic application.

The embodiment of the present invention the 3rd object is the performance evaluating apparatus providing the second whole machine cabinet server, apply the whole machine cabinet server performance parameter that this technical scheme obtains, more be close to the users application scenarios, the evaluation result obtained is more accurate, more realistic application.

The embodiment of the present invention the 4th object is the performance evaluating apparatus providing the third whole machine cabinet server, apply the whole machine cabinet server performance parameter that this technical scheme obtains, more be close to the users application scenarios, the evaluation result obtained is more accurate, more realistic application.

The embodiment of the present invention the 5th object is the performance evaluating apparatus providing the 4th kind of whole machine cabinet server, apply the whole machine cabinet server performance parameter that this technical scheme obtains, more be close to the users application scenarios, the evaluation result obtained is more accurate, more realistic application.

The performance evaluation methodology of a kind of whole machine cabinet server that the embodiment of the present invention provides, comprising:

Each virtual server in tested whole machine cabinet server in arbitrary load unit sends server calls order,

Described server calls order comprises: the information requiring the task that each described virtual server performs respectively,

Wherein, each described load unit respectively with each user's one_to_one corresponding;

Obtain the actual execution result information of each described virtual server to described task,

According to the actual execution result information to each described task, determine the performance of described load unit and each described virtual server.

Alternatively, the actual execution result information of described virtual server to described task comprises: the actual execution number of times and the execution time that perform described task.

Alternatively, step: each virtual server in tested whole machine cabinet server in arbitrary load unit sends server calls order, specifically:

According to the IP address of each described virtual server or the domain name of each described virtual server, described server calls order is sent to respectively each described virtual server in described load unit.

Alternatively, described load unit comprises: at least one server cluster, each described server cluster respectively by: at least two described virtual servers form;

When described virtual server belongs to arbitrary described server cluster, step: each virtual server in tested whole machine cabinet server in arbitrary load unit sends server calls order, specifically:

According to the floating IP address of each described server cluster or the domain name of each described server cluster, respectively to each described server cluster, send described server calls order,

For described server cluster according to predetermined cluster balance policy, each described virtual server is specified to perform described task;

In step: obtain each described virtual server to after the actual execution result information of described task, also comprise:

According to the actual execution result information to described task of virtual server described in each server cluster, determine the cluster performance of each described server cluster respectively.

Alternatively, described server cluster is: by least two mail server clusters for providing the described virtual server of mail service to form to described user;

Or described server cluster is: by least two web server clusters for providing the described virtual server of Web service to form to described user.

Alternatively, in step: each virtual server in tested whole machine cabinet server in arbitrary load unit also comprises after sending server calls order:

The utilization factor of the virtual cpu of each described virtual server in load unit described in periodic monitor,

There is overloaded cpu virtual server if current, then further:

According to the utilization factor of each described virtual cpu, by least one virtual cpu unappropriated in the resource pool of described whole machine cabinet server, distribute to described overloaded cpu virtual server,

Wherein, described overloaded cpu virtual server is: the utilization factor of described virtual cpu is equal to or greater than the virtual server of the predetermined cpu busy percentage upper limit corresponding to described virtual server;

And/or, in step: each virtual server in tested whole machine cabinet server in arbitrary load unit also comprises after sending server calls order:

The utilization factor of the virtual memory of each described virtual server in load unit described in periodic monitor,

There is internal memory overload virtual server if current, then further:

According to the utilization factor of each described virtual memory, by the part of virtual memory unappropriated in described resource pool, distribute to described internal memory overload virtual server,

Wherein, described internal memory overload virtual server is: the utilization factor of described virtual memory is equal to or greater than the virtual server of the predetermined virtual memory utilization factor upper limit corresponding to described virtual server.

Alternatively, in step: according to the utilization factor of each described virtual cpu, by least one virtual cpu unappropriated in the resource pool of described whole machine cabinet server, after distributing to described overloaded cpu virtual server,

And/or, in step: according to the utilization factor of each described virtual memory, by the part of virtual memory unappropriated in described resource pool, after distributing to described internal memory overload virtual server, also comprise:

According to the historical record of the batch operation of described virtual cpu, and/or, according to the historical record of the batch operation of described virtual memory, determine the resource allocation ability of described whole machine cabinet server, described resource comprises: the arbitrary or combination of described virtual cpu, virtual memory.

Alternatively, the utilization factor of the virtual cpu of each described virtual server is respectively:

Quantity is taken, with the ratio value taking quantity of the current described virtual cpu of described virtual server what start that initialization completes the described virtual cpu of virtual server described in the moment;

The utilization factor of the virtual memory of each described virtual server is respectively:

The occupancy of the described virtual memory of virtual server described in the moment is completed, with the ratio value of the occupancy of the current virtual memory of described virtual server in startup initialization.

Alternatively, in step: obtain each described virtual server to after the actual execution result information of described task, also comprise:

Store the performance of described load unit and each described virtual server, for the performance according to all described load units and all described virtual servers, determine the combination property of described whole machine cabinet server.

Alternatively, also comprise:

Framework call instruction is sent to framework manager, comprise in described framework call instruction: require described framework manager to the clone of described virtual server specified and the mission bit stream of deployment and require described framework manager to the mission bit stream of the migration of the described virtual server of specifying and require that described framework manager is to the mission bit stream of the migration of the data on storer described in specify at least two, wherein each described storer is connected with each described virtual server respectively

Obtain the practical task execution result information of described framework manager to the clone of described virtual server and the task of deployment,

Obtain the practical task execution result information of described framework manager to the mission bit stream of the migration of described virtual server,

Obtain the practical task execution result information of described framework manager to the mission bit stream of the migration of the data on described storer,

According to each described practical task execution result information, determine the framework managerial ability of described framework manager;

Wherein said framework manager is connected respectively with all described load units.

The performance evaluating apparatus of a kind of whole machine cabinet server that the embodiment of the present invention provides, comprising:

Call instruction transmitting element, for sending server calls order to each virtual server in arbitrary load unit in tested whole machine cabinet server,

Described server calls order comprises: the information requiring the task that each described virtual server performs respectively,

Wherein, each described load unit respectively with each user's one_to_one corresponding;

Execution result determining unit, for obtaining the actual execution result information of each described virtual server to described task;

Performance evaluating unit, is connected with described execution result determining unit, for according to the actual execution result information to each described task, determines the performance of described load unit and each described virtual server.

Alternatively, described call instruction transmitting element, specifically for according to the IP address of each described virtual server or the domain name of each described virtual server, is sent to each described virtual server respectively by described server calls order.

Alternatively, described call instruction transmitting element, concrete also for: when described virtual server belongs to arbitrary described server cluster, according to the floating IP address of each described server cluster or the domain name of each described server cluster, respectively to each described server cluster, send described server calls order

For described server cluster according to predetermined cluster balance policy, each described virtual server is specified to perform described task;

Described performance evaluating unit, also for the actual execution result information to described task according to virtual server described in each server cluster, determines the cluster performance of each described server cluster respectively.

Alternatively, also comprise:

The utilization factor determining unit of virtual cpu, for the utilization factor of the virtual cpu of described virtual server each in load unit described in periodic monitor;

Virtual cpu deployment unit, be connected with the utilization factor determining unit of described virtual cpu, for the utilization factor according to each described virtual cpu, by least one virtual cpu unappropriated in the resource pool of described whole machine cabinet server, distribute to described overloaded cpu virtual server

Wherein, described overloaded cpu virtual server is: the utilization factor of described virtual cpu is equal to or greater than the virtual server of the predetermined cpu busy percentage upper limit corresponding to described virtual server.

Alternatively, also comprise:

The utilization factor determining unit of virtual memory, for the utilization factor of the virtual memory of described virtual server each in load unit described in periodic monitor;

Virtual memory deployment unit, connects with the utilization factor determining unit of described virtual memory, for the utilization factor according to each described virtual memory, by the part of virtual memory unappropriated in described resource pool, distributes to described internal memory overload virtual server,

Wherein, described internal memory overload virtual server is: the utilization factor of described virtual memory is equal to or greater than the virtual server of the predetermined virtual memory utilization factor upper limit corresponding to described virtual server.

Alternatively, described performance evaluating unit, also be connected respectively with described virtual cpu deployment unit, virtual memory deployment unit, also for the historical record of the batch operation according to described virtual cpu, and/or, according to the historical record of the batch operation of described virtual memory, determine the resource allocation ability of described whole machine cabinet server, described resource comprises: the arbitrary or combination of described virtual cpu, virtual memory.

Alternatively, described call instruction unit, also send framework call instruction for the framework manager to described whole machine cabinet server, comprise in described framework call instruction: require that described framework manager is to the clone of described virtual server specified and the mission bit stream of deployment, and require that described framework manager is to the mission bit stream of the migration of the described virtual server of specifying, and require that described framework manager is to the mission bit stream of the migration of the data on storer described in specify at least two, wherein each described storer is connected with each described virtual server respectively, wherein said framework manager is connected respectively with all described load units,

Framework manager information determining unit, for obtaining the following information of described framework manager: to the practical task execution result information of the practical task execution result information of the clone of described virtual server and the task of deployment, the practical task execution result information to the mission bit stream of the migration of described virtual server, the mission bit stream to the migration of the data on described storer;

Described performance evaluating unit is also connected with described framework manager information determining unit, for according to each described practical task execution result information, determines the framework managerial ability of described framework manager.

The performance evaluating apparatus of a kind of whole machine cabinet server that the embodiment of the present invention provides, comprising:

Main test cell and at least one the auxiliary test cell be connected with described main test cell;

Described main test cell comprises: the first configuration information obtains subelement, the first call instruction sends subelement, execution result determining unit, the first storing sub-units and performance parameter determination subelement,

Described first configuration information obtains subelement, for obtaining the configuration information of each virtual server in the described load unit that is connected with described main test cell in tested whole machine cabinet server, described configuration information is stored to described first storing sub-units, and the configuration information of described virtual server comprises: the routing information of each described virtual server;

Described first call instruction sends subelement, obtain subelement with described first configuration information respectively, described load unit is connected, for sending server calls order to each described virtual server in described load unit, described server calls order comprises: the information requiring the task that each described virtual server performs respectively, wherein, each described load unit respectively with each user's one_to_one corresponding, and

For sending subtest call instruction to each described auxiliary test cell, each described load unit connected to described auxiliary test cell according to described call instruction for described auxiliary test cell sends call instruction;

First execution result determining unit, for obtaining each described virtual server to the actual execution result information of described task and the execution result information to described subtest call instruction obtaining described auxiliary test cell;

Described first storing sub-units, for storing information, described information comprises: the configuration information of each described virtual server and each described virtual server are to the execution result information to described subtest call instruction of the actual execution result information of described task and described auxiliary test cell;

Described performance parameter determination subelement, respectively with described first storing sub-units, described second storing sub-units in described auxiliary test machine is connected, for according to the data in described first storing sub-units and each described second storing sub-units, according to the actual execution result information of each described virtual server in each described load unit to each described task, determine the performance of each described load unit and each described virtual server, according to the execution result information to described subtest call instruction of each described auxiliary test cell, determine the performance of described auxiliary test cell,

Described auxiliary test cell is the unit of exerting pressure of simulation client terminal access behavior,

Described auxiliary test cell comprises: the second configuration information obtains subelement, the second call instruction sends subelement, the second execution result determining unit and the second storing sub-units;

Described second configuration information obtains subelement, for the control of the call instruction according to described master controller, the configuration information of each described virtual server in the described load unit that acquisition is connected with described auxiliary test cell, the configuration information of described each described virtual server is stored to described second storing sub-units, and described configuration information comprises: the routing information of each described virtual server;

Described second call instruction sends subelement, obtain subelement with described second configuration information respectively, described load unit is connected, for the control of the call instruction according to described master controller, server calls order is sent to each described virtual server in described load unit, described server calls order comprises: the information requiring the task that each described virtual server performs respectively

Second execution result determining unit, for obtaining the actual execution result information of each described virtual server to described task;

Described second storing sub-units, for storing information, described information comprises: the configuration information of each described virtual server and each described virtual server are to the actual execution result information of described task.

The performance evaluating apparatus of a kind of whole machine cabinet server that the embodiment of the present invention provides, comprising:

Call instruction transmitter, for sending server calls order to each virtual server in arbitrary load unit in tested whole machine cabinet server,

Described server calls order comprises: the information requiring the task that each described virtual server performs respectively,

Wherein, each described load device respectively with each user's one_to_one corresponding;

Execution result determiner, for obtaining the actual execution result information of each described virtual server to described task;

Performance evaluating device, is connected with described execution result determiner, for according to the actual execution result information to each described task, determines the performance of described load device and each described virtual server.

Alternatively, described call instruction transmitter, concrete also for: when described virtual server belongs to arbitrary described server cluster, according to the floating IP address of each described server cluster or the domain name of each described server cluster, respectively to each described server cluster, send described server calls order

For described server cluster according to predetermined cluster balance policy, each described virtual server is specified to perform described task;

Described performance evaluating device, also for the actual execution result information to described task according to virtual server described in each server cluster, determines the cluster performance of each described server cluster respectively.

Alternatively, also comprise:

The utilization factor determiner of virtual cpu, for the utilization factor of the virtual cpu of described virtual server each in load device described in periodic monitor;

Virtual cpu tuner, is connected with the utilization factor determiner of described virtual cpu, for the utilization factor according to each described virtual cpu, by least one virtual cpu unappropriated in the resource pool of described whole machine cabinet server, distributes to described overloaded cpu virtual server,

Wherein, described overloaded cpu virtual server is: the utilization factor of described virtual cpu is equal to or greater than the virtual server of the predetermined cpu busy percentage upper limit corresponding to described virtual server.

Alternatively, also comprise:

The utilization factor determiner of virtual memory, for the utilization factor of the virtual memory of described virtual server each in load device described in periodic monitor;

Virtual memory tuner, connects with the utilization factor determiner of described virtual memory, for the utilization factor according to each described virtual memory, by the part of virtual memory unappropriated in described resource pool, distributes to described internal memory overload virtual server,

Wherein, described internal memory overload virtual server is: the utilization factor of described virtual memory is equal to or greater than the virtual server of the predetermined virtual memory utilization factor upper limit corresponding to described virtual server.

Alternatively, described performance evaluating device, also be connected respectively with described virtual cpu tuner, virtual memory tuner, also for the historical record of the batch operation according to described virtual cpu, and/or, according to the historical record of the batch operation of described virtual memory, determine the resource allocation ability of described whole machine cabinet server, described resource comprises: the arbitrary or combination of described virtual cpu, virtual memory.

Alternatively, described call instruction device, also send framework call instruction for the framework manager to described whole machine cabinet server, comprise in described framework call instruction: require that described framework manager is to the clone of described virtual server specified and the mission bit stream of deployment, and require that described framework manager is to the mission bit stream of the migration of the described virtual server of specifying, and require that described framework manager is to the mission bit stream of the migration of the data on storer described in specify at least two, wherein each described storer is connected with each described virtual server respectively, wherein said framework manager is connected respectively with all described load devices,

Framework manager information determiner, for obtaining the following information of described framework manager: to the practical task execution result information of the practical task execution result information of the clone of described virtual server and the task of deployment, the practical task execution result information to the mission bit stream of the migration of described virtual server, the mission bit stream to the migration of the data on described storer;

Described performance evaluating device is also connected with described framework manager information determiner, for according to each described practical task execution result information, determines the framework managerial ability of described framework manager.

The performance evaluating apparatus of a kind of whole machine cabinet server that the embodiment of the present invention provides, comprising:

Main tester and at least one the auxiliary tester be connected with described main tester;

Described main tester comprises: the first configuration information getter, the first call instruction transmitter, execution result determiner, first memory and performance parameter determiner,

Described first configuration information getter, for obtaining the configuration information of each described virtual server in the described load device that is connected with described main tester in tested whole machine cabinet server, described configuration information is stored to described first memory, and the configuration information of described virtual server comprises: the routing information of each described virtual server;

Described first call instruction transmitter, be connected with described first configuration information getter, described load device respectively, for sending server calls order to each described virtual server in described load device, described server calls order comprises: the information requiring the task that each described virtual server performs respectively, wherein, each described load device respectively with each user's one_to_one corresponding, and

For sending subtest call instruction to each described auxiliary tester, each described load device connected to described auxiliary tester according to described call instruction for described auxiliary tester sends call instruction;

First execution result determiner, for obtaining each described virtual server to the actual execution result information of described task and the execution result information to described subtest call instruction obtaining described auxiliary tester;

Described first memory, for storing information, described information comprises: the configuration information of each described virtual server and each described virtual server are to the execution result information to described subtest call instruction of the actual execution result information of described task and described auxiliary tester;

Described performance parameter determiner, be connected with the described second memory in described first memory, described auxiliary test machine respectively, for according to the data in described first memory and each described second memory, according to each described virtual server in each described load device to the actual execution result information of each described task, determine the performance of each described load device and each described virtual server, according to the execution result information to described subtest call instruction of each described auxiliary tester, determine the performance of described auxiliary tester;

Described auxiliary tester is the pressure applicator of simulation client terminal access behavior,

Described auxiliary tester comprises: the second configuration information getter, the second call instruction transmitter, the second execution result determiner and second memory;

Described second configuration information getter, for the control of the call instruction according to described master controller, the configuration information of each described virtual server in the described load device that acquisition is connected with described auxiliary tester, the configuration information of described each described virtual server is stored to described second memory, and described configuration information comprises: the routing information of each described virtual server;

Described second call instruction transmitter, be connected with described second configuration information getter, described load device respectively, for the control of the call instruction according to described master controller, server calls order is sent to each described virtual server in described load device, described server calls order comprises: the information requiring the task that each described virtual server performs respectively

Second execution result determiner, for obtaining the actual execution result information of each described virtual server to described task;

Described second memory, for storing information, described information comprises: the configuration information of each described virtual server and each described virtual server are to the actual execution result information of described task.

Therefore, technical scheme in application the present invention, the embodiment of the present invention is when carrying out the performance evaluating of whole machine cabinet server, server calls order is sent respectively in this whole machine cabinet server, with each virtual server in user one to one load unit, after transmission call instruction, according to the actual execution result information of each virtual server to the task in call instruction, obtain the performance of each virtual server in the performance of each load unit and each load unit.Due to the load unit that object tested in the present embodiment technical scheme is corresponding with user, and when performance evaluating, the mode issuing call instruction to each virtual server in load unit by test machine is simulated when user uses, the tasks carrying actual result of each virtual server in the load unit that this user is corresponding, the performance parameter obtained according to the tasks carrying actual result of each virtual server in the load unit that user is corresponding more can reflect the reality of user's application scenarios, therefore the evaluation result obtained is more accurate, more realistic application.

Accompanying drawing explanation

In order to be illustrated more clearly in the embodiment of the present invention or technical scheme of the prior art, be briefly described to the accompanying drawing used required in embodiment or description of the prior art below, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.

The performance evaluation methodology schematic flow sheet of a kind of whole machine cabinet server that Fig. 1 provides for the embodiment of the present invention 1;

A kind of method flow schematic diagram obtaining virtual server CPU and virtual memory resource utilization that Fig. 2 provides for the embodiment of the present invention 2;

A kind of method flow schematic diagram obtaining virtual server CPU and virtual memory resource utilization that Fig. 3 provides for the embodiment of the present invention 3;

A kind of method flow schematic diagram obtaining virtual server CPU and virtual memory resource utilization that Fig. 4 provides for the embodiment of the present invention 4;

A kind of method flow schematic diagram obtaining virtual server CPU and virtual memory resource utilization that Fig. 5 provides for the embodiment of the present invention 5;

The performance evaluating apparatus structural representation of a kind of whole machine cabinet server that Fig. 6 provides for the embodiment of the present invention 6;

The performance evaluating apparatus structural representation of a kind of whole machine cabinet server that Fig. 7 provides for the embodiment of the present invention 7;

The performance evaluating apparatus structural representation of a kind of whole machine cabinet server that Fig. 8 provides for the embodiment of the present invention 8;

The performance evaluating apparatus structural representation of a kind of whole machine cabinet server that Fig. 9 provides for the embodiment of the present invention 9;

The performance evaluating apparatus structural representation of a kind of whole machine cabinet server that Figure 10 provides for the embodiment of the present invention 10;

The performance evaluating apparatus structural representation of a kind of whole machine cabinet server that Figure 11 provides for the embodiment of the present invention 11;

The performance evaluating apparatus structural representation of a kind of whole machine cabinet server that Figure 12 provides for the embodiment of the present invention 12;

The performance evaluating apparatus structural representation of a kind of whole machine cabinet server that Figure 13 provides for the embodiment of the present invention 13;

The performance evaluating apparatus structural representation of a kind of whole machine cabinet server that Figure 14 provides for the embodiment of the present invention 14;

The performance evaluating apparatus structural representation of a kind of whole machine cabinet server that Figure 15 provides for the embodiment of the present invention 15.

Embodiment

Below in conjunction with the accompanying drawing in the embodiment of the present invention, be clearly and completely described the technical scheme in the embodiment of the present invention, obviously, described embodiment is only the present invention's part embodiment, instead of whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art, not making the every other embodiment obtained under creative work prerequisite, belong to the scope of protection of the invention.

Embodiment 1:

Shown in Figure 1, present embodiments provide a kind of performance evaluation methodology of whole machine cabinet server.This whole machine cabinet server provides virtualizing performance service, comprises at least two load units, and each load unit comprises at least two virtual servers respectively,

In the present embodiment, the performance evaluating of whole machine cabinet is mainly comprised the following steps:

Step 101: each virtual server in whole machine cabinet server in arbitrary load unit sends server calls order.

In the present embodiment in tested whole machine cabinet server respectively with user one to one each load unit be connected, wherein this load unit can be, but not limited to self-defined division, each load unit is corresponding with a user, it is equivalent to the server load of this user, and each load unit is made up of one or more virtual server.

In the present embodiment, can according to the configuration information of each virtual server in the tested load unit obtained in advance, this configuration information comprises: the routing information of this virtual server in this network architecture in this whole machine cabinet server, specifically according to these configuration informations, send to each virtual server in tested load unit: server calls order.

This server calls order is including but not limited to comprising following information: the information requiring the task that each virtual server performs respectively.This task is specially the service that when analog subscriber is applied, this virtual server provides to user, such as but be not limited to basis: require that virtual server one performs N mail transmission/reception, or require that virtual server two performs M web site chaining service etc., or other.

Wherein each virtual server can be mail server, web server or electronic transaction server etc., specifically can preset according to various application scenario.

In the present embodiment, can be, but not limited to, according to the IP address in the network architecture of each virtual server in tested load unit in this whole machine cabinet server or domain name, send above-mentioned call instruction to each virtual server.

Step 102: to obtain in tested load unit each virtual server to the actual execution result information of corresponding task.

After send the call instruction being used for analog subscriber application to tested load unit, according to virtual server each in load unit feedack in the process of executing the task or after executing the task, according to feedack, determine that in tested load unit, each virtual server is to the actual execution result information of corresponding task, each virtual server to the actual execution result information of corresponding task can be for example: the actual execution number of times and the execution time that perform corresponding task, or also comprises time delay and error rate etc.

In addition, in tested load unit, each virtual server can by the virtual server active feedback in each tested load unit to the actual execution result information of task, also can by this test subject to each virtual server initiatively by inquire about or mode of asking etc. obtains.

Step 103: according to the actual execution result information of each virtual server to corresponding task, determine the performance of each virtual server in tested load unit and this tested load unit.

In the present embodiment, according to each virtual server in tested load unit (set of all virtual servers namely corresponding to the user that simulates) according to call instruction to the execution of task (namely in analog subscriber application process, virtual server provides the simulation of service for user) structure: actual execution result information, can determine that this tested load unit is when practical application scene, this complete machine equipment cabinet server that user side obtains provides the service performance of service to user, the performance of each virtual server namely in the performance of tested load unit and this load unit.

The performance parameter of each virtual server in the present embodiment in the performance of tested load unit and this load unit, can be, but not limited to be determined by the comparison of actual execution result information and predetermined gross data, such as but be not limited to:

Virtual server can be adopted in the schedule time the actual execution number of times of preplanned mission, as the actual execution result information to corresponding task of virtual server, if performing number of times actual is A1, it is A0 that this virtual server is performed number of times to the theory of corresponding task within this schedule time, in the present embodiment can using the performance parameter of following ratio value: A1/A0 as this virtual server.

It is that A0 can be made a reservation for by user that this wherein above-mentioned virtual server performs number of times to the theory of corresponding task within this schedule time, also can for the numerical value recorded under laboratory fc-specific test FC environment.

Therefore, technical scheme in application the present invention, the embodiment of the present invention is when carrying out the performance evaluating of whole machine cabinet server, server calls order is sent respectively in this whole machine cabinet server, with each virtual server in user one to one load unit, after transmission call instruction, according to the actual execution result information of each virtual server to the task in call instruction, obtain the performance of each virtual server in the performance of each load unit and each load unit.Due to the load unit that object tested in the present embodiment technical scheme is corresponding with user, and when performance evaluating, the mode issuing call instruction to each virtual server in load unit by test machine is simulated when user uses, the tasks carrying actual result of each virtual server in the load unit that this user is corresponding, the performance parameter obtained according to the tasks carrying actual result of each virtual server in the load unit that user is corresponding more can reflect the reality of user's application scenarios, therefore the evaluation result obtained is more accurate, more realistic application.

When applying the present embodiment technical scheme, the present embodiment technical scheme can be adopted respectively at arbitrary user side, the performance of each virtual server in the load unit corresponding to arbitrary user and this load unit is tested, obtain the performance parameter of this whole machine cabinet server that each user side obtains respectively, and the performance parameter of this whole machine cabinet server that can obtain according to all user sides, the comprehensive performance parameter of this whole machine cabinet server is obtained according to predetermined performance synthesis estimating algorithm.

Embodiment 2:

The present embodiment and embodiment 1 different dwelling mainly comprise: load unit virtual memory tested in the whole machine cabinet server of this enforcement is at one or more server cluster, namely the multiple virtual servers in this load unit form a server cluster, when there being user task, can by cluster according to predetermined cluster protocol, by the principle of the arbitrary or multiple virtual server in cluster according to load balancing, for user provides service.This server cluster is conducive to the service provision capacity and the load balance ability that play complete machine equipment cabinet server further.

Shown in Figure 2, correspondingly, the whole machine cabinet Server performance flow process in this enforcement is as follows:

Step 201: each virtual server in whole machine cabinet server in arbitrary load unit sends server calls order.

If when some virtual server in tested load unit does not belong to arbitrary server cluster, concrete the same as described in example 1 above, according to the IP address of this virtual server in the network architecture of this whole machine cabinet server or domain name, directly send this call instruction to this virtual server.

If when some virtual server in this tested load unit belongs to arbitrary server cluster, specifically can be, but not limited to the domain name (corresponding to cluster) of floating IP address (corresponding to cluster) according to this server cluster or this server cluster, respectively to this server cluster, send call instruction.After this server cluster receives this call instruction, in units of server cluster, for the purpose of the task of having specified in complete call order, according to predetermined cluster balance policy, the virtual server determined is specified to perform corresponding task.

Wherein this predetermined cluster balance policy can be made a reservation for by User Defined, also can by network management user when cluster builds, agreement fixing setting in advance in cluster.

Each server cluster in the present embodiment can be the cluster of various virtual server, such as can for by least two mail server clusters for providing the virtual server of mail service to form to user; Also can for by least two web server clusters for providing the virtual server of Web service to form to user.Do not repeat at this and limit.

Step 202: to obtain in tested load unit each virtual server to the actual execution result information of corresponding task.

Step 203: according to the actual execution result information of each virtual server to corresponding task, determine the performance of each virtual server in tested load unit and this tested load unit.

Above step 202-203 respectively with the step 102 in embodiment 1,103 in like manner.

Step 204: according to the actual execution result information to corresponding task of virtual server in each server cluster, determine the cluster performance of each server cluster respectively.

The present embodiment is for the whole machine cabinet server of presence server cluster, after the performance obtaining each virtual server in each load unit, also further according to the actual execution result information to corresponding task of virtual server each in each server cluster, determine the cluster performance of each server cluster in whole machine cabinet server further, i.e. cluster equalization performance.

Therefore, the present embodiment technical scheme is except can obtaining the beneficial effect described in embodiment 1, the server cluster ability in whole machine cabinet server can also be obtained further, obtain the whole machine cabinet performance of more various dimensions, make evaluation result more accurate, more react real network framework situation.

Embodiment 3:

Shown in Figure 3, the present embodiment and embodiment 2 differences are mainly: after step 201, step 202,203, before arbitrary step of 204 or afterwards, also comprise further

Step 301: periodically or aperiodically, the utilization factor of the virtual cpu of each virtual server in the tested load unit of periodic monitor.

Wherein, the virtual cpu of virtual server utilization factor can but do not limit and adopt following sign: take quantity, with the ratio value taking quantity of the virtual cpu of the current time in virtual server what start that initialization completes the virtual cpu of moment virtual server.

Step 302: according to monitoring result: the utilization factor of the virtual cpu of each virtual server in tested load unit, determines whether there is overloaded cpu virtual server.

In the present embodiment, this overloaded cpu Virtual Service implement body is defined as: the utilization factor of virtual cpu is equal to or greater than the virtual server of the predetermined cpu busy percentage upper limit;

CPU non-overload virtual server is: the utilization factor of virtual cpu is less than the virtual server of the predetermined cpu busy percentage upper limit.

It is identical that the wherein above-mentioned cpu busy percentage upper limit can be set to all virtual servers, also can preset and be decided to be difference, and the virtual server being namely respectively different sets the different cpu busy percentage upper limits.

Wherein, the utilization factor of the virtual cpu of each virtual server is respectively: complete the moment in startup initialization, the virtual cpu of this virtual server take quantity, with the ratio value taking quantity of virtual cpu shown in current in this virtual server.

Therefore, in embodiments of the present invention, owing to monitoring and dynamic adaptation the virtual cpu of the virtual server in each load unit and virtual memory resource, virtual resource can be made to be fully utilized, the use scenes of real simulation user, makes the result of evaluation and test more accurate.

Step 303: if there is overloaded cpu virtual server, then perform step 304; Otherwise return, wait for the arrival in next monitoring moment, continue monitoring, do not perform virtual cpu allotment.

Step 304: the utilization factor obtaining each virtual cpu of each virtual server according to monitoring, carries out virtual cpu allotment.

Concrete virtual cpu allotment is, according to the utilization factor of each virtual cpu, by at least one virtual cpu unappropriated in the resource pool of whole machine cabinet server, distribute to current C PU overload virtual server, to alleviate the overload condition of this overloaded cpu virtual server, improve the task treatment effeciency of this virtual server.

After execution this step 304, return step 301, wait for the arrival in next monitoring moment, continue monitoring, until monitor complete, when occurring that above-mentioned appearance is above-mentioned when each monitoring at every turn and there is the situation of overloaded cpu virtual server, all perform this step and carry out virtual cpu adjustment.

The determination of the utilization factor of the virtual cpu of concrete each virtual server specifically can according to following scheme:

If tested load unit virtual memory is at following virtual server: the first mail server, the second mail server, the first web server, the second web server.

Start successfully at each virtual server, complete the moment (being now idle task load) in startup initialization, read the initial occupation value of each virtual server virtual cpu in tested load unit, if the virtual cpu number that a certain virtual server is assigned with is 10, the number completing moment in running order virtual cpu (i.e. occupied virtual cpu) in dynamic initialization is 1, then the initial occupation value of this virtual server virtual cpu is: 1/10, and the virtual cpu being in idle condition is 9.

After transmission call instruction, (now each virtual server is respectively according to call instruction, perform the corresponding task in call instruction), the virtual cpu utilization factor of each virtual server of periodic monitor, being located at the current number monitoring the in running order virtual cpu (i.e. occupied virtual cpu) of the first virtual server is 10, then the initial occupation value of this first virtual server virtual cpu is: 10/10, and the upper limit of the cpu busy percentage of the first virtual server is 9/10, then now the first virtual server is: overloaded cpu virtual server;

The number of the current current in running order virtual cpu of the second virtual server (i.e. occupied virtual cpu) monitored is 3, then the initial occupation value of this second virtual server virtual cpu is: 3/10, and the upper limit of the cpu busy percentage of the second virtual server is 9/10, then now the second virtual server is: the non-overload virtual server of CPU.

Then now can perform virtual cpu deployment operations, according to the allotment flow process preset, by a current unappropriated virtual cpu in the resource pool of whole machine cabinet server or multiple virtual cpu (also claiming residue virtual cpu), distribute to this overloaded cpu virtual server.For example in resource pool, current 2 virtual cpus being not yet assigned to arbitrary virtual server are allocated to the first virtual server use, to share the work on the first virtual server, improve treatment effeciency and speed, now the first virtual server has 12 virtual cpus.Visible, adopt this technical scheme can integrate the resource of whole machine cabinet server further, improve service processing efficiency and the speed of whole machine cabinet server substantially, be conducive to the stability improving service.

In order to improve the various dimensions of evaluation and test performance further, after each periodic monitor performs step 302, step 3011 can also performed further: store current monitoring result.

Further, after step 3011, step 305 is performed: store current allotment result.

Further, when the result of determination when step 304 is no, and after step 305, also step 3051 is performed: the utilization factor of virtual cpu has been monitored, if do not completed, then returns step 301; Otherwise perform step 306.

Step 306: according to virtual cpu deployment operations historical record, determines the resource allocation ability of whole machine cabinet server.

Therefore, adopt technique scheme can determine tested whole machine cabinet server resource allotment ability further, obtain the performance information of more various dimensions.

It should be noted that, Fig. 3 is only the signal of the present embodiment flow process, wherein, step 301-3051 can before the arbitrary step after step 201 or after.

Embodiment 4:

Shown in Figure 4, after step 201, step 202,203, before arbitrary step of 204 or afterwards, also comprise further

The whole machine cabinet server evaluating method of the present embodiment mainly comprises following flow process:

Step 401: periodically or aperiodically, the utilization factor of virtual cpu of each virtual server and the utilization factor of virtual memory in the tested load unit of periodic monitor.

The utilization factor of the virtual memory of virtual server can be, but not limited to adopt following sign: the occupancy completing the virtual memory of moment virtual server in startup initialization, with the ratio value of the occupancy of the current virtual memory in virtual server.

Step 402: store current monitoring result.

Step 403: according to monitoring result: the utilization factor of the virtual cpu of each virtual server in tested load unit, determines whether there is internal memory overload virtual server, determines whether overloaded cpu virtual server.

In the present embodiment, this internal memory overload virtual server is: the utilization factor of virtual memory is equal to or greater than the virtual server of the predetermined virtual memory utilization factor upper limit corresponding to virtual server;

Internal memory non-overload virtual server is: the utilization factor of virtual memory is less than the virtual server of the predetermined virtual memory utilization factor upper limit corresponding to virtual server.

It is identical that the utilization factor upper limit of wherein above-mentioned virtual memory can be set to all virtual servers, and also can preset and be decided to be difference, the virtual server being namely respectively different sets the utilization factor upper limit of different virtual memorys.

Step 404: if there is overloaded cpu virtual server, then perform step 405; Otherwise perform step 407 and do not perform the virtual memory allotment between virtual server.

Step 405: the utilization factor obtaining each virtual cpu of each virtual server according to monitoring, carries out virtual cpu allotment.

Specifically can be, but not limited to the record see step 304 in embodiment 2.

Step 306: store current allotment result.

Step 407: if there is internal memory overload virtual server, then perform step 408; Otherwise perform step 410 and do not perform the overloaded cpu virtual server allotment between virtual server.

Step 408: the utilization factor obtaining the virtual memory of each virtual server according to monitoring, carries out virtual memory allotment.

The determination of the utilization factor of the virtual memory of concrete each virtual server, specifically can determine according to the scheme of the utilization factor with reference to virtual cpu:

If there is following virtual server in tested load unit: the first mail server, the second mail server, the first web server, the second web server.

Start successfully at each virtual server, complete the moment (being now idle task load) in startup initialization, read the initial occupation value of each virtual server virtual memory in tested load unit, if the virtual memory size that a certain virtual server is assigned with is 1G, complete the initial occupation value 5% of moment in running order virtual memory in dynamic initialization.

After transmission call instruction, (now each virtual server is respectively according to call instruction, perform the corresponding task in call instruction), the utilization factor of the virtual memory of each virtual server of periodic monitor, if the current occupation value monitoring the virtual memory of the first virtual server current is: 98%, the upper limit of the utilization factor of the virtual memory of the first virtual server is 80%, then now the first virtual server is: internal memory overload virtual server.

If the total virtual memory of the second virtual server monitored current is 1G, the occupation value of current virtual internal memory is: 10%, the upper limit of the utilization factor of the virtual virtual memory of the second virtual server is 80%, then now the second virtual server is: the non-overload virtual server of internal memory.

Now, following virtual memory deployment operations can be performed: according to the allotment flow process preset, by the part being not yet assigned to the virtual memory (also known as residue virtual memory) of arbitrary virtual server current in resource pool, distribute to current internal memory overload virtual server.For example the 200M of unappropriated virtual memory current in resource pool can be distributed to above-mentioned internal memory overload virtual server: the first virtual server, to share the virtual memory load of the first virtual server, improve treatment effeciency and speed, now the virtual memory of the first virtual server is 1.2G.

Therefore, the present embodiment is except can obtaining the beneficial effect in embodiment 3, also due in the monitoring to virtual cpu and carry out according to current monitoring outside dynamic virtual cpu allotment, dynamic monitoring and dynamic adaptation are carried out to the virtual memory of each virtual server in load unit, therefore can virtual resource be made further to be utilized more fully, the more use scenes of real simulation user, makes the result of evaluation and test more accurate.

Step 409: store current allotment result.

Step 410: whether the utilization factor monitoring of virtual cpu and virtual memory completes, if do not completed, then returns step 401; Otherwise perform step 410.

Step 411: according to virtual cpu deployment operations historical record and virtual memory deployment operations historical record, determine the resource allocation ability of whole machine cabinet server.

When occurring that above-mentioned appearance is above-mentioned when each monitoring at every turn and there is overloaded cpu virtual server situation, all perform virtual cpu adjustment, each occur above-mentioned there is internal memory overload virtual server situation time, all perform virtual memory adjustment,

To sum up, adopt this technical scheme can integrate the resource of whole machine cabinet server further, dynamically improve service processing efficiency and the speed of whole machine cabinet server substantially, be conducive to the stability improving service.

The embodiment of the resource allocation ability of this whole machine cabinet server can for being deployed into power or allotment quality reaction to virtual cpu and virtual memory.

Therefore, adopt technique scheme can determine tested whole machine cabinet server resource allotment ability further, obtain the performance information of more various dimensions.

Wherein, the resource pool in various embodiments of the present invention is the set of various resource in whole machine cabinet server, and this resource includes but not limited to virtual cpu and virtual memory.

When carrying out the deployment of virtual server, each virtual cpu in resource pool and each virtual memory can be distributed to respectively each virtual server and using.

The resource pool of whole machine cabinet server can realize on a physical hardware, also can realize on multiple physical hardware.

Embodiment 5:

Shown in Figure 5, the difference of the present embodiment and embodiment 4 is mainly, go back in the present embodiment virtual server clone with deployment, virtual server dynamic migration, store the task operating of dynamic migration, specifically after step 204 in embodiment 1, also comprise further:

Step 501: send framework call instruction to framework manager.

Wherein framework manager is connected respectively with each load unit in each load unit in tested whole machine cabinet server, for carrying out the operation of framework class in whole machine cabinet server: such as the clone of virtual server and deployment (comprising information and network information setting), virtual server migration (virtualizing performance of the first virtual server is realized elsewhere), storage Data Migration between the storer that in whole machine cabinet server, each virtual server is shared etc.

Comprise in framework call instruction: require framework manager to the clone of the virtual server of specifying and the mission bit stream of deployment and require framework manager to the mission bit stream of the migration of the virtual server of specifying and require that framework manager is to the mission bit stream of the migration of the data at least two storeies of specifying.

For example: clone's virtual server is specially, create each virtual server in each load unit, as created each virtual server in load unit N, clone's mail server is the first mail server N, second mail server N, clone's standard server is standard server N, clone web data storehouse server is web data storehouse server N, clone's Web server is the first Web server N, second Web server N, clone's electronic trade database server is electronic trade database server N, clone's electronic trade Web server is the first electronic trade Web server N, second electronic trade Web server N, 3rd electronic trade Web server N, its N represents 1, 2, 3, 4......, information setting is configured after clone, namely dispose.

Step 502: obtain framework manager to the framework result of virtual server.

The present embodiment specifically, the practical task execution result information to the clone of virtual server and the task of deployment cloned by acquisition framework manager, obtain framework manager to the practical task execution result information of the mission bit stream of the migration of virtual server, obtain framework manager to the practical task execution result information of the mission bit stream of the migration of the data on storer.

The practical task execution result of framework manager to each framework task can be determined in the present embodiment by the number of times and time performing the operation of above-mentioned framework.

Step 503: according to framework manager to above-mentioned each practical task execution result information, determine the framework managerial ability of framework manager.

It should be noted that, be described to perform step 501-503 after step 204 in the present embodiment, but be not limited to this, in the actual flow process that can perform step 501-503 before or after step 201 or arbitrary step.

Therefore, in embodiments of the present invention, due to virtual server clone with deployments, virtual server dynamic migration, store the simulated operation of dynamic migration, can the use scenes of real simulation user, make the result of evaluation and test more accurate.

In sum, adopt the evaluating method of above-described embodiment can obtain the multidimensional performance of this whole machine cabinet server, concrete sees table.

Table one: the performance parameter table of virtual server

The wherein performance of above-mentioned each virtual server, i.e. shown under the computing method of so-called relative score:

1: in load unit 1, total relative score of each virtual server is $\mathrm{\Σw}=\sqrt[7]{(c1*c2*c3*c5*c6*c7)}*y1;$

2: the total relative score of framework is $\mathrm{\Σi}=\sqrt[3]{(c8*c9*c10)}*y2;$

3: monitoring and the total relative score of dynamic resource allocation are ∑ v=c11*y3;

4: the total relative score of cluster is $\mathrm{\Σc}=\sqrt[2]{(c11*c12)}*y4;$

Y1+y2+y3+y4=100% wherein in above-mentioned 1,2,3,4;

5: the overall scores of load unit 1 is

$\mathrm{\Σ}1=\sqrt[7]{(c1*c2*c3*c5*c6*c7*)}*y1+\sqrt[3]{(c8*c9*c10)}*y2+c11*y3+\sqrt[2]{(c11*c12)}*y4;$

The overall scores of 6:N load unit is

$\mathrm{\Σ}1=\sqrt[7]{(c1*c2*c3*c5*c6*c7)}*y1+K\sqrt[7]{(\mathrm{cN}*\mathrm{cN}*\mathrm{cN}*\mathrm{cN}*\mathrm{cN}*\mathrm{cN})}*y1+\sqrt[3]{(c8*c9*c10)}*y2+$

$c11*y3+\sqrt[2]{(c11*c12)}*y4$

In Table 1, " reference value " specifically, by the theoretical value calculated this loaded server according to each payload traffic model, this theoretical reference value is the value recorded under laboratory fc-specific test FC environment;

" actual score " in above-mentioned table one specifically, characterizes with the actual number of times of executing the task of each virtual server.

The first mail server in upper table one, the second mail server, the first web server, the second web server, the first electronic trade web server, the second electronic trade web server, the 3rd electronic trade web server, be actual virtual server;

Standard server in upper table one is: the load virtual server of standard, and namely without the virtual server of any business load, object is the true use scenes of analog subscriber;

Framework operation in upper table one, comprises virtual server clone and deployment, virtual server dynamic migration, stores dynamic migration;

In upper table one, server cluster comprises: mail server cluster and web server cluster;

Monitoring in upper table one comprises the monitoring to monitoring and resource allocation server;

In upper table two, y1, y2, y3, y4 are respectively percentages, and y1+y2+y3+y4=100%, and in account form, the total relative score of each virtual server in load unit, framework operates the total relative score of total relative score, monitoring and resource allocation, the total relative score of cluster gets percent value again, and object is reduced as far as possible error of calculation rate;

In upper table one, the overall scores of each load unit is " in load unit, the total relative score of each virtual server, framework always operate the total relative score of relative score, monitoring and resource allocation, the total relative score of cluster " sum;

The overall scores of N number of load unit is " in each load unit, total relative score of each virtual server, framework operate the total relative score of total relative score, monitoring and resource allocation, the total relative score of cluster " sum;

In upper table one, overall scores presentation format is: " X.X Y VMs (Partitions Z) ", wherein: X.X is a numerical value, represent the overall scores score value of N number of load unit, and numerical value X.X is accurate to radix point 1; Y represents in this load unit to be provided with several virtual server, for example load unit is interior to be provided with N number of virtual server, then Y=N*10+1, Vms are tag, and the Z in " Partitions Z) " represents that whole machine cabinet server has divided how many subregions and evaluated and tested.

Achievement as comprised the load unit of 12 virtual servers is: 50.9@121VMs (Partitions2), and wherein the whole machine cabinet at its place is that 2 subregions are evaluated and tested by Huawei.

When the physical cpu of whole machine cabinet and virtual memory resource are utilized completely, total performance of each load unit reaches maximal value.

Embodiment 6:

Shown in Figure 6, present embodiments provide a kind of performance evaluating apparatus 600 of whole machine cabinet server, it comprises: call instruction transmitting element 601, execution result determining unit 602, performance evaluating unit 603.Annexation and the principle of work of each several part are as follows:

Call instruction transmitting element 601, sends server calls order with each virtual server in the north side load unit in tested whole machine cabinet server respectively.Comprise in server calls order: the information requiring the task that each virtual server performs respectively.Wherein, each load unit respectively with each user's one_to_one corresponding.

In the present embodiment, as a kind of alternatives, this call instruction transmitting element 601, specifically for according to the IP address of each virtual server or the domain name of each virtual server, is sent to each virtual server respectively by above-mentioned server calls order.Further can see the record in embodiment 2.

In the present embodiment, as a kind of alternatives, when virtual server belongs to arbitrary server cluster, this call instruction transmitting element 601 specifically can according to the domain name of the floating IP address of each server cluster or each server cluster, respectively to each server cluster, send server calls order, for each server cluster according to predetermined cluster balance policy, specify each virtual server to execute the task.Further can see the record in embodiment 2.

Execution result determining unit 602, for after call instruction transmitting element 601 sends call instruction, determines the actual execution result information of each virtual server to corresponding task.

Performance evaluating unit 603, is connected with execution result determining unit 602, for according to the actual execution result information of virtual server to corresponding task, determines the performance of each virtual server in tested load unit and tested load unit.

Correspondingly, if include server cluster in this whole machine cabinet server, then above-mentioned performance evaluating unit 603, also for the actual execution result information to corresponding task according to virtual server in each server cluster, determines the cluster performance of each server cluster respectively.Further can see the record in embodiment 2.To the whole machine cabinet server of presence server cluster, after the performance obtaining each virtual server in each load unit, also further according to the actual execution result information to corresponding task of virtual server each in each server cluster, determine the cluster performance of each server cluster in whole machine cabinet server further, i.e. cluster equalization performance.

Further principle of work can be, but not limited to the record see embodiment 1,2.

Therefore, technical scheme in application the present invention, the embodiment of the present invention is when carrying out the performance evaluating of whole machine cabinet server, call instruction transmitting element 601 in the whole machine cabinet Server performance device of the present embodiment is in this whole machine cabinet server, server calls order is sent with each virtual server in the load unit that user is tested one to one, after transmission call instruction, execution result determining unit 602 determines the actual execution result information of each virtual server to the task in call instruction, performance evaluating unit 603 is according to the actual execution result information of each virtual server to the task in call instruction, obtain the performance of each virtual server in the performance of each load unit and each load unit.

Due to the load unit that object tested in the present embodiment technical scheme is corresponding with user, and when performance evaluating, the mode issuing call instruction to each virtual server in load unit by test machine is simulated when user uses, the tasks carrying actual result of each virtual server in the load unit that this user is corresponding, the performance parameter obtained according to the tasks carrying actual result of each virtual server in the load unit that user is corresponding more can reflect the reality of user's application scenarios, therefore the evaluation result obtained is more accurate, more realistic application.

Embodiment 7:

Shown in Figure 7, the performance evaluating apparatus 700 of the whole machine cabinet server that the present embodiment provides also comprises relative to embodiment 6 evaluating apparatus:

The utilization factor determining unit 701 of virtual cpu, for the utilization factor of the virtual cpu of each virtual server in the tested load unit of periodic monitor;

Virtual cpu deployment unit 702, is connected with the utilization factor determining unit 701 of virtual cpu, for the utilization factor according to each virtual cpu, by least one virtual cpu unappropriated in the resource pool of whole machine cabinet server, distributes to overloaded cpu virtual server,

Wherein, overloaded cpu virtual server is: the utilization factor of virtual cpu is equal to or greater than the virtual server of the predetermined cpu busy percentage upper limit corresponding to virtual server; CPU non-overload virtual server is: the utilization factor of virtual cpu is less than the virtual server of the predetermined cpu busy percentage upper limit corresponding to virtual server.

Further principle of work can be, but not limited to see the detailed description in embodiment 3,4.

Therefore, adopt technique scheme can determine tested whole machine cabinet server resource allotment ability further, obtain the performance information of more various dimensions.

Embodiment 8:

Shown in Figure 8, the performance evaluating apparatus 800 of the whole machine cabinet server that the present embodiment provides also comprises relative to embodiment 7 evaluating apparatus:

The utilization factor determining unit 801 of virtual memory, for the utilization factor of the virtual memory of each virtual server in the tested load unit of periodic monitor.

Virtual memory deployment unit 802, connect with the utilization factor determining unit 801 of virtual memory, for the utilization factor according to each virtual memory, by the part of unappropriated virtual memory current in the resource pool of whole machine cabinet server, distribute to current internal memory overload virtual server.

Wherein, internal memory overload virtual server is: the utilization factor of virtual memory is equal to or greater than the virtual server of the predetermined virtual memory utilization factor upper limit corresponding to virtual server; Internal memory non-overload virtual server is: the utilization factor of virtual memory is less than the virtual server of the predetermined virtual memory utilization factor upper limit corresponding to virtual server.

Performance evaluating unit 803, is also connected respectively with virtual cpu deployment unit, virtual memory deployment unit 802, also for according to virtual cpu deployment operations historical record and virtual memory deployment operations historical record, determines the resource allocation ability of whole machine cabinet server.

Further principle of work can be, but not limited to see the detailed description in embodiment 4.

Therefore, adopt technique scheme can determine tested whole machine cabinet server resource allotment ability further, obtain the performance information of more various dimensions.

Embodiment 9:

Shown in Figure 9, the performance evaluating apparatus 900 of the whole machine cabinet server that the present embodiment provides mainly is relative to embodiment 8 evaluating apparatus institute difference:

Call instruction transmitting element 901 in the performance evaluating apparatus of the whole machine cabinet server of the present embodiment, also sends framework call instruction for the framework manager to whole machine cabinet server.

Comprise in framework call instruction: require framework manager to the clone of the virtual server of specifying and the mission bit stream of deployment and require framework manager to the mission bit stream of the migration of the virtual server of specifying and require that framework manager is to the mission bit stream of the migration of the data at least two storeies of specifying, wherein each storer is connected with each virtual server respectively, and wherein framework manager is connected respectively with all load units;

The performance evaluating apparatus of the whole machine cabinet server of the present embodiment also comprises:

Framework manager information determining unit 902, for obtaining the following information of framework manager: to the practical task execution result information of the practical task execution result information of the clone of virtual server and the task of deployment, the practical task execution result information to the mission bit stream of the migration of virtual server, the mission bit stream to the migration of the data on storer;

The performance evaluating unit 603 of the performance evaluating apparatus of the whole machine cabinet server of the present embodiment is also connected with framework manager information determining unit 902, also for each practical task execution result information according to framework manager, determine the framework managerial ability of framework manager.

More detailed principle of work can be, but not limited to the record see embodiment 5.

Therefore, in embodiments of the present invention, due to virtual server clone with deployments, virtual server dynamic migration, store the simulated operation of dynamic migration, can the use scenes of real simulation user, make the result of evaluation and test more accurate.

Embodiment 10:

Shown in Figure 10, the whole machine cabinet server that the present embodiment provides its mainly comprise: main test cell 101 and at least one the auxiliary test cell 102 be connected with main test cell 101.

Main test cell 101 is mainly pressure test and controls master unit, for configuration testing parameter, and control and management test process.Auxiliary test cell 102 is the unit of exerting pressure of simulation client terminal access behavior.

Main test cell 101 comprises: the first configuration information obtains subelement 1011, first call instruction and sends subelement 1012, first execution result determining unit 1013, first storing sub-units 1014 and performance parameter determining unit 1015.

Auxiliary test cell 102 comprises: the second configuration information obtains subelement 1021, second call instruction and sends subelement 1022, second execution result determining unit 1023 and the second storing sub-units 1024.

Each modular construction in auxiliary test cell 102 and principle of work as follows:

Second configuration information obtains subelement 1021, for the control of the call instruction according to master controller, the configuration information of each virtual server in the tested load unit that acquisition is connected with auxiliary test cell 102, the configuration information of each virtual server is stored to the second storing sub-units 1024, configuration information comprises: the routing information of each virtual server;

Second call instruction sends subelement 1022, obtain subelement 1021 with the second configuration information respectively, tested load unit is connected, for the control of the call instruction according to master controller, server calls order is sent to each virtual server in load unit, comprise in server calls order: the information requiring the task that each virtual server performs respectively

Second execution result determining unit 1023, for obtaining the actual execution result information of each virtual server to corresponding task.

Second storing sub-units 1024, for storing information, information includes but not limited to: the configuration information of each virtual server and each virtual server are to the actual execution result information of corresponding task.

Structure in main test cell 101 is as follows:

First configuration information obtains subelement 1011, for obtaining the configuration information of each virtual server in the tested load unit that is connected with described main tester in whole machine cabinet server, configuration information is stored to the first storing sub-units 1014 in main test cell 101, the configuration information of virtual server is including but not limited to the routing information of each virtual server.

First call instruction sends subelement 1012, subelement 1011 is obtained respectively with the first configuration information, and tested load unit is connected, for sending server calls order to each virtual server in tested load unit, comprise in server calls order: the information requiring the task that each virtual server performs respectively, wherein, each load unit respectively with each user's one_to_one corresponding, and first call instruction send subelement 1012 also for sending subtest call instruction to each auxiliary test cell 102, server calls order is sent for each virtual server in the tested load unit that auxiliary test cell 102 connects to auxiliary test cell 102 according to subtest call instruction.

First execution result determining unit 1013, for obtaining each virtual server to the actual execution result information of corresponding task and the execution result information to subtest call instruction obtaining each auxiliary test cell 102.

First storing sub-units 1014, for storing information, information includes but not limited to: the configuration information of each virtual server and each tested virtual server are to the execution result information to subtest call instruction of the actual execution result information of corresponding task and each auxiliary test cell 102.

Performance parameter determining unit 1015, be connected with the second storing sub-units 1024 in the first storing sub-units 1014, auxiliary test machine respectively, for according to the data in the first storing sub-units 1014 and each second storing sub-units 1024, according to the actual execution result information of each virtual server in each tested load unit to corresponding task, determine the performance of each load unit and each virtual server, and, according to the execution result information to subtest call instruction of each auxiliary test cell 102, determine the performance of auxiliary test cell 102.

Therefore, the present embodiment is except having embodiment 1, outside beneficial effect in 6, because the performance evaluating apparatus of the whole machine cabinet server in originally implementing is divided into a main test cell 101 and multiple attached test cell, can be controlled by coordinated management when applying and improve testing efficiency, and can when arbitrary test cell breaks down, substituted by other test cells (main test cell 101 or attached test cell) and served as function, perform the test to tested load unit, make the robustness of whole machine cabinet Server performance device more.

Embodiment 11:

Embodiment 11:

Shown in Figure 11, present embodiments provide a kind of performance evaluating apparatus 110 of whole machine cabinet server, it comprises: call instruction transmitter 1101, execution result determiner 1102, performance evaluating device 1103.Annexation and the principle of work of each several part are as follows:

Call instruction transmitter 1101, sends server calls order with each virtual server in the north side load device in tested whole machine cabinet server respectively.Comprise in server calls order: the information requiring the task that each virtual server performs respectively.Wherein, each load device respectively with each user's one_to_one corresponding.

In the present embodiment, as a kind of alternatives, this call instruction transmitter 1101, specifically for according to the IP address of each virtual server or the domain name of each virtual server, is sent to each virtual server respectively by above-mentioned server calls order.Further can see the record in embodiment 2.

In the present embodiment, as a kind of alternatives, when virtual server belongs to arbitrary server cluster, this call instruction transmitter 1101 specifically can according to the domain name of the floating IP address of each server cluster or each server cluster, respectively to each server cluster, send server calls order, for each server cluster according to predetermined cluster balance policy, specify each virtual server to execute the task.Further can see the record in embodiment 2.

Execution result determiner 1102, for after call instruction transmitter 1101 sends call instruction, determines the actual execution result information of each virtual server to corresponding task.

Performance evaluating device 1103, is connected with execution result determiner 1102, for according to the actual execution result information of virtual server to corresponding task, determines the performance of each virtual server in tested load device and tested load device.

Correspondingly, if include server cluster in this whole machine cabinet server, then above-mentioned performance evaluating device 1103, also for the actual execution result information to corresponding task according to virtual server in each server cluster, determines the cluster performance of each server cluster respectively.Further can see the record in embodiment 2.To the whole machine cabinet server of presence server cluster, after the performance obtaining each virtual server in each load device, also further according to the actual execution result information to corresponding task of virtual server each in each server cluster, determine the cluster performance of each server cluster in whole machine cabinet server further, i.e. cluster equalization performance.

Further principle of work can be, but not limited to the record see embodiment 1,2.

Therefore, technical scheme in application the present invention, the embodiment of the present invention is when carrying out the performance evaluating of whole machine cabinet server, call instruction transmitter 1101 in the whole machine cabinet Server performance device of the present embodiment is in this whole machine cabinet server, server calls order is sent with each virtual server in the load device that user is tested one to one, after transmission call instruction, execution result determiner 1102 determines the actual execution result information of each virtual server to the task in call instruction, performance evaluating device 1103 is according to the actual execution result information of each virtual server to the task in call instruction, obtain the performance of each virtual server in the performance of each load device and each load device.

Due to the load device that object tested in the present embodiment technical scheme is corresponding with user, and when performance evaluating, the mode issuing call instruction to each virtual server in load device by test machine is simulated when user uses, the tasks carrying actual result of each virtual server in the load device that this user is corresponding, the performance parameter obtained according to the tasks carrying actual result of each virtual server in the load device that user is corresponding more can reflect the reality of user's application scenarios, therefore the evaluation result obtained is more accurate, more realistic application.

Embodiment 12:

Shown in Figure 12, the performance evaluating apparatus 120 of the whole machine cabinet server that the present embodiment provides also comprises relative to embodiment 11 evaluating apparatus:

The utilization factor determiner 1201 of virtual cpu, for the utilization factor of the virtual cpu of each virtual server in periodic monitor tested load device;

Virtual cpu tuner 1202, is connected with the utilization factor determiner 1201 of virtual cpu, for the utilization factor according to each virtual cpu, by least one virtual cpu unappropriated in the resource pool of whole machine cabinet server, distributes to overloaded cpu virtual server,

Wherein, overloaded cpu virtual server is: the utilization factor of virtual cpu is equal to or greater than the virtual server of the predetermined cpu busy percentage upper limit corresponding to virtual server; CPU non-overload virtual server is: the utilization factor of virtual cpu is less than the virtual server of the predetermined cpu busy percentage upper limit corresponding to virtual server.

Further principle of work can be, but not limited to see the detailed description in embodiment 3,4.

Therefore, adopt technique scheme can determine tested whole machine cabinet server resource allotment ability further, obtain the performance information of more various dimensions.

Embodiment 13:

Shown in Figure 13, the performance evaluating apparatus 130 of the whole machine cabinet server that the present embodiment provides also comprises relative to embodiment 12 evaluating apparatus:

The utilization factor determiner 1301 of virtual memory, for the utilization factor of the virtual memory of each virtual server in periodic monitor tested load device.

Virtual memory tuner 1302, connect with the utilization factor determiner 1301 of virtual memory, for the utilization factor according to each virtual memory, by the part of unappropriated virtual memory current in the resource pool of whole machine cabinet server, distribute to current internal memory overload virtual server.

Wherein, internal memory overload virtual server is: the utilization factor of virtual memory is equal to or greater than the virtual server of the predetermined virtual memory utilization factor upper limit corresponding to virtual server; Internal memory non-overload virtual server is: the utilization factor of virtual memory is less than the virtual server of the predetermined virtual memory utilization factor upper limit corresponding to virtual server.

Performance evaluating device 1303, is also connected respectively with virtual cpu tuner, virtual memory tuner 1302, also for according to virtual cpu deployment operations historical record and virtual memory deployment operations historical record, determines the resource allocation ability of whole machine cabinet server.

Further principle of work can be, but not limited to see the detailed description in embodiment 4.

Therefore, adopt technique scheme can determine tested whole machine cabinet server resource allotment ability further, obtain the performance information of more various dimensions.

Embodiment 14:

Shown in Figure 14, the performance evaluating apparatus 140 of the whole machine cabinet server that the present embodiment provides mainly is relative to embodiment 13 evaluating apparatus institute difference:

Call instruction transmitter 1401 in the performance evaluating apparatus of the whole machine cabinet server of the present embodiment, also sends framework call instruction for the framework manager to whole machine cabinet server.

Comprise in framework call instruction: require framework manager to the clone of the virtual server of specifying and the mission bit stream of deployment and require framework manager to the mission bit stream of the migration of the virtual server of specifying and require that framework manager is to the mission bit stream of the migration of the data at least two storeies of specifying, wherein each storer is connected with each virtual server respectively, and wherein framework manager is connected respectively with all load devices;

The performance evaluating apparatus of the whole machine cabinet server of the present embodiment also comprises:

Framework manager information determiner 1402, for obtaining the following information of framework manager: to the practical task execution result information of the practical task execution result information of the clone of virtual server and the task of deployment, the practical task execution result information to the mission bit stream of the migration of virtual server, the mission bit stream to the migration of the data on storer;

The performance evaluating device 1103 of the performance evaluating apparatus of the whole machine cabinet server of the present embodiment is also connected with framework manager information determiner 1402, also for each practical task execution result information according to framework manager, determine the framework managerial ability of framework manager.

More detailed principle of work can be, but not limited to the record see embodiment 5.

Therefore, in embodiments of the present invention, due to virtual server clone with deployments, virtual server dynamic migration, store the simulated operation of dynamic migration, can the use scenes of real simulation user, make the result of evaluation and test more accurate.

Embodiment 15:

Shown in Figure 15, the whole machine cabinet server that the present embodiment provides its mainly comprise: main tester 151 and at least one the auxiliary tester 152 be connected with main tester 151.

Main tester 151 is mainly pressure test and controls main device, for configuration testing parameter, and control and management test process.Auxiliary tester 152 is the pressure applicator of simulation client terminal access behavior.

Main tester 151 comprises: the first configuration information getter 15011, first call instruction transmitter 15012, first execution result determiner 15013, first memory 15014 and performance parameter determiner 15015.

Auxiliary tester 152 comprises: the second configuration information getter 15021, second call instruction transmitter 15022, second execution result determiner 15023 and second memory 15024.

Each modular construction in auxiliary tester 152 and principle of work as follows:

Second configuration information getter 15021, for the control of the call instruction according to master controller, the configuration information of each virtual server in the tested load device that acquisition is connected with auxiliary tester 152, the configuration information of each virtual server is stored to second memory 15024, and configuration information comprises: the routing information of each virtual server;

Second call instruction transmitter 15022, be connected with the second configuration information getter 15021, tested load device respectively, for the control of the call instruction according to master controller, server calls order is sent to each virtual server in load device, comprise in server calls order: the information requiring the task that each virtual server performs respectively

Second execution result determiner 15023, for obtaining the actual execution result information of each virtual server to corresponding task.

Second memory 15024, for storing information, information includes but not limited to: the configuration information of each virtual server and each virtual server are to the actual execution result information of corresponding task.

Structure in main tester 151 is as follows:

First configuration information getter 15011, for obtaining the configuration information of each virtual server in the tested load device that is connected with described main tester in whole machine cabinet server, configuration information is stored to the first memory 15014 in main tester 151, the configuration information of virtual server is including but not limited to the routing information of each virtual server.

First call instruction transmitter 15012, respectively with the first configuration information getter 15011, and tested load device is connected, for sending server calls order to each virtual server in tested load device, comprise in server calls order: the information requiring the task that each virtual server performs respectively, wherein, each load device respectively with each user's one_to_one corresponding, and first call instruction transmitter 15012 also for sending subtest call instruction to each auxiliary tester 152, server calls order is sent for each virtual server in the tested load device that auxiliary tester 152 connects to auxiliary tester 152 according to subtest call instruction.

First execution result determiner 15013, for obtaining each virtual server to the actual execution result information of corresponding task and the execution result information to subtest call instruction obtaining each auxiliary tester 152.

First memory 15014, for storing information, information includes but not limited to: the configuration information of each virtual server and each tested virtual server are to the execution result information to subtest call instruction of the actual execution result information of corresponding task and each auxiliary tester 152.

Performance parameter determiner 15015, be connected with the second memory 15024 in first memory 15014, auxiliary test machine respectively, for according to the data in first memory 15014 and each second memory 15024, according to each virtual server in each tested load device to the actual execution result information of corresponding task, determine the performance of each load device and each virtual server, and, according to the execution result information to subtest call instruction of each auxiliary tester 152, determine the performance of auxiliary tester 152.

Therefore, the present embodiment is except having the beneficial effect in embodiment 1,6, because the performance evaluating apparatus of the whole machine cabinet server in originally implementing is divided into a main tester 151 and multiple attached tester, can be controlled by coordinated management when applying and improve testing efficiency, and can when arbitrary tester breaks down, substituted by other testers (main tester 151 or attached tester) and served as function, perform the test to tested load device, make the robustness of whole machine cabinet Server performance device more.

It should be noted that, in various embodiments of the present invention, tested whole machine cabinet server can be various whole machine cabinet server, such as can for supporting that the whole machine cabinet server of subregion can for not supporting the whole machine cabinet server of subregion yet.

Such as, when tested whole machine cabinet server supports subregion, usually, server management tool can be passed through when practical application, according to user need whole machine cabinet server logic is divided into N number of subregion, each subregion arranges virtual server, and each virtual server is separately provided services on the Internet to user, one of them virtual server can be provided services on the Internet to a user, also can provide services on the Internet respectively to multiple user.In this whole machine cabinet server, the storage resources between each virtual server is shared mutually;

More specifically, if during this whole machine cabinet server support subregion, such as subregion number N=1, can all virtual servers be divided in this subregion jointly in the present embodiment, specifically the virtual memory of virtual cpu corresponding for each virtual server and its correspondence be all divided in this subregion.Further, storer can also be divided into n logical storage (n>1), and arrange n logical storage be virtualized all virtual servers a subregion share;

When subregion number N ≠ 1, in the present embodiment can by the N number of subregion of average division of all virtual servers, specifically that the virtual memory of virtual cpu corresponding for each virtual server and its correspondence is on average divided in N number of subregion, and, storage is divided into n logical storage (n=N), and n logical storage is set is shared by N number of subregion.

Such as, when tested object whole machine cabinet server support not subregion time, can whole machine cabinet server be regarded as traditional server to evaluate and test, storage is divided into n logical storage (n >=3), wherein 1 logical storage is used for virtual platform, and (n-1) individual logical storage configuration is shared.

Device embodiment described above is only schematic, the wherein said unit illustrated as separating component or can may not be and physically separates, parts as unit display can be or may not be physical location, namely can be positioned at a place, or also can be distributed in multiple network element.Some or all of module wherein can be selected according to the actual needs to realize the object of the present embodiment scheme.Those of ordinary skill in the art, when not paying performing creative labour, are namely appreciated that and implement.

Through the above description of the embodiments, those skilled in the art can be well understood to the mode that each embodiment can add required general hardware platform by software and realize, and can certainly pass through hardware.Based on such understanding, technique scheme can embody with the form of software product the part that prior art contributes in essence in other words, this computer software product can store in a computer-readable storage medium, as ROM/RAM, magnetic disc, CD etc., comprising some instructions in order to make a computer installation (can be personal computer, server, or network equipment etc.) perform some Part Methods of each embodiment or embodiment.

Above embodiment, does not form the restriction to this technical scheme protection domain.The amendment done within any spirit at above-mentioned embodiment and principle, equivalently to replace and improvement etc., within the protection domain that all should be included in this technical scheme.

Claims (18)

1. a performance evaluation methodology for whole machine cabinet server, is characterized in that, comprising:

Each virtual server in tested whole machine cabinet server in arbitrary load unit sends server calls order,

Described server calls order comprises: the information requiring the task that each described virtual server performs respectively, and this task is specially the service that when analog subscriber is applied, this virtual server provides to user,

Wherein, each described load unit respectively with each user's one_to_one corresponding;

Obtain the actual execution result information of each described virtual server to described task,

According to the actual execution result information to each described task, determine the performance of described load unit and each described virtual server.

2. the performance evaluation methodology of a kind of whole machine cabinet server according to claim 1, is characterized in that,

The actual execution result information of described virtual server to described task comprises: the actual execution number of times and the execution time that perform described task.

3. a kind of performance evaluation methodology of whole machine cabinet server according to claim 1 or 2, is characterized in that,

Step: each virtual server in tested whole machine cabinet server in arbitrary load unit sends server calls order, specifically:

According to the IP address of each described virtual server or the domain name of each described virtual server, described server calls order is sent to respectively each described virtual server in described load unit.

4. a kind of performance evaluation methodology of whole machine cabinet server according to claim 1 or 2, is characterized in that,

Described load unit comprises: at least one server cluster, each described server cluster respectively by: at least two described virtual servers form;

When described virtual server belongs to arbitrary described server cluster, step: each virtual server in tested whole machine cabinet server in arbitrary load unit sends server calls order, specifically:

According to the floating IP address of each described server cluster or the domain name of each described server cluster, respectively to each described server cluster, send described server calls order,

For described server cluster according to predetermined cluster balance policy, each described virtual server is specified to perform described task;

In step: obtain each described virtual server to after the actual execution result information of described task, also comprise:

According to the actual execution result information to described task of virtual server described in each server cluster, determine the cluster performance of each described server cluster respectively.

5. the performance evaluation methodology of a kind of whole machine cabinet server according to claim 4, is characterized in that,

Described server cluster is: by least two mail server clusters for providing the described virtual server of mail service to form to described user;

Or described server cluster is: by least two web server clusters for providing the described virtual server of Web service to form to described user.

6. the performance evaluation methodology of a kind of whole machine cabinet server according to claim 1, is characterized in that,

In step: each virtual server in tested whole machine cabinet server in arbitrary load unit also comprises after sending server calls order:

The utilization factor of the virtual cpu of each described virtual server in load unit described in periodic monitor,

There is overloaded cpu virtual server if current, then further:

According to the utilization factor of each described virtual cpu, by least one virtual cpu unappropriated in the resource pool of described whole machine cabinet server, distribute to described overloaded cpu virtual server,

Wherein, described overloaded cpu virtual server is: the utilization factor of described virtual cpu is equal to or greater than the virtual server of the predetermined cpu busy percentage upper limit corresponding to described virtual server;

And/or, in step: each virtual server in tested whole machine cabinet server in arbitrary load unit also comprises after sending server calls order:

The utilization factor of the virtual memory of each described virtual server in load unit described in periodic monitor,

There is internal memory overload virtual server if current, then further:

According to the utilization factor of each described virtual memory, by the part of virtual memory unappropriated in described resource pool, distribute to described internal memory overload virtual server,

Wherein, described internal memory overload virtual server is: the utilization factor of described virtual memory is equal to or greater than the virtual server of the predetermined virtual memory utilization factor upper limit corresponding to described virtual server.

7. the performance evaluation methodology of a kind of whole machine cabinet server according to claim 6, is characterized in that,

In step: according to the utilization factor of each described virtual cpu, by least one virtual cpu unappropriated in the resource pool of described whole machine cabinet server, after distributing to described overloaded cpu virtual server,

And/or, in step: according to the utilization factor of each described virtual memory, by the part of virtual memory unappropriated in described resource pool, after distributing to described internal memory overload virtual server, also comprise:

According to the historical record of the batch operation of described virtual cpu, and/or, according to the historical record of the batch operation of described virtual memory, determine the resource allocation ability of described whole machine cabinet server, described resource comprises: the arbitrary or combination of described virtual cpu, virtual memory.

8. the performance evaluation methodology of a kind of whole machine cabinet server according to claim 6, is characterized in that,

The utilization factor of the virtual cpu of each described virtual server is respectively:

Quantity is taken, with the ratio value taking quantity of the current described virtual cpu of described virtual server what start that initialization completes the described virtual cpu of virtual server described in the moment;

The utilization factor of the virtual memory of each described virtual server is respectively:

The occupancy of the described virtual memory of virtual server described in the moment is completed, with the ratio value of the occupancy of the current virtual memory of described virtual server in startup initialization.

9. the performance evaluation methodology of a kind of whole machine cabinet server according to claim 1 and 2, is characterized in that,

In step: obtain each described virtual server to after the actual execution result information of described task, also comprise:

Store the performance of described load unit and each described virtual server, for the performance according to all described load units and all described virtual servers, determine the combination property of described whole machine cabinet server.

10. the performance evaluation methodology of a kind of whole machine cabinet server according to claim 1 and 2, is characterized in that, also comprise:

Framework call instruction is sent to framework manager, comprise in described framework call instruction: require described framework manager to the clone of described virtual server specified and the mission bit stream of deployment and require described framework manager to the mission bit stream of the migration of the described virtual server of specifying and require that described framework manager is to the mission bit stream of the migration of the data at least two storeies of specifying, wherein each described storer is connected with each described virtual server respectively

Obtain the practical task execution result information of described framework manager to the clone of described virtual server and the task of deployment,

Obtain the practical task execution result information of described framework manager to the mission bit stream of the migration of described virtual server,

Obtain the practical task execution result information of described framework manager to the mission bit stream of the migration of the data on described storer,

According to each described practical task execution result information, determine the framework managerial ability of described framework manager;

Wherein said framework manager is connected respectively with all described load units.

The performance evaluating apparatus of 11. 1 kinds of whole machine cabinet servers, is characterized in that, comprising:

Call instruction transmitting element, for sending server calls order to each virtual server in arbitrary load unit in tested whole machine cabinet server,

Described server calls order comprises: the information requiring the task that each described virtual server performs respectively, and this task is specially the service that when analog subscriber is applied, this virtual server provides to user,

Wherein, each described load unit respectively with each user's one_to_one corresponding;

Execution result determining unit, for obtaining the actual execution result information of each described virtual server to described task;

Performance evaluating unit, is connected with described execution result determining unit, for according to the actual execution result information to each described task, determines the performance of described load unit and each described virtual server.

The performance evaluating apparatus of 12. whole machine cabinet servers according to claim 11, is characterized in that,

Described call instruction transmitting element, specifically for according to the IP address of each described virtual server or the domain name of each described virtual server, is sent to each described virtual server respectively by described server calls order.

The performance evaluating apparatus of 13. whole machine cabinet servers according to claim 11, is characterized in that,

Described call instruction transmitting element, concrete also for: when described virtual server belongs to arbitrary described server cluster, according to the floating IP address of each described server cluster or the domain name of each described server cluster, respectively to each described server cluster, send described server calls order

For described server cluster according to predetermined cluster balance policy, each described virtual server is specified to perform described task;

Described performance evaluating unit, also for the actual execution result information to described task according to virtual server described in each server cluster, determines the cluster performance of each described server cluster respectively.

The performance evaluating apparatus of 14. whole machine cabinet servers according to claim 11, is characterized in that, also comprise:

The utilization factor determining unit of virtual cpu, for the utilization factor of the virtual cpu of described virtual server each in load unit described in periodic monitor;

Virtual cpu deployment unit, be connected with the utilization factor determining unit of described virtual cpu, for the utilization factor according to each described virtual cpu, by least one virtual cpu unappropriated in the resource pool of described whole machine cabinet server, distribute to described overloaded cpu virtual server

Wherein, described overloaded cpu virtual server is: the utilization factor of described virtual cpu is equal to or greater than the virtual server of the predetermined cpu busy percentage upper limit corresponding to described virtual server.

The performance evaluating apparatus of 15. whole machine cabinet servers according to claim 14, is characterized in that, also comprise:

The utilization factor determining unit of virtual memory, for the utilization factor of the virtual memory of described virtual server each in load unit described in periodic monitor;

Virtual memory deployment unit, connects with the utilization factor determining unit of described virtual memory, for the utilization factor according to each described virtual memory, by the part of virtual memory unappropriated in described resource pool, distributes to described internal memory overload virtual server,

Wherein, described internal memory overload virtual server is: the utilization factor of described virtual memory is equal to or greater than the virtual server of the predetermined virtual memory utilization factor upper limit corresponding to described virtual server.

The performance evaluating apparatus of 16. whole machine cabinet servers according to claim 15, is characterized in that,

Described performance evaluating unit, also be connected respectively with described virtual cpu deployment unit, virtual memory deployment unit, also for the historical record of the batch operation according to described virtual cpu, and/or, according to the historical record of the batch operation of described virtual memory, determine the resource allocation ability of described whole machine cabinet server, described resource comprises: the arbitrary or combination of described virtual cpu, virtual memory.

17. according to claim 11 to 16 the performance evaluating apparatus of arbitrary described whole machine cabinet server, it is characterized in that,

Described call instruction unit, also send framework call instruction for the framework manager to described whole machine cabinet server, comprise in described framework call instruction: require that described framework manager is to the clone of described virtual server specified and the mission bit stream of deployment, and require that described framework manager is to the mission bit stream of the migration of the described virtual server of specifying, and require that described framework manager is to the mission bit stream of the migration of the data at least two storeies of specifying, wherein each described storer is connected with each described virtual server respectively, wherein said framework manager is connected respectively with all described load units,

Framework manager information determining unit, for obtaining the following information of described framework manager: to the practical task execution result information of the practical task execution result information of the clone of described virtual server and the task of deployment, the practical task execution result information to the mission bit stream of the migration of described virtual server, the mission bit stream to the migration of the data on described storer;

Described performance evaluating unit is also connected with described framework manager information determining unit, for according to each described practical task execution result information, determines the framework managerial ability of described framework manager.

The performance evaluating apparatus of 18. 1 kinds of whole machine cabinet servers, is characterized in that, comprising:

Main test cell and at least one the auxiliary test cell be connected with described main test cell;

Described main test cell comprises: the first configuration information obtains subelement, the first call instruction sends subelement, execution result determining unit, the first storing sub-units and performance parameter determination subelement,

Described first configuration information obtains subelement, for obtaining the configuration information of each virtual server in the load unit that is connected with described main test cell in tested whole machine cabinet server, described configuration information is stored to described first storing sub-units, and the configuration information of described virtual server comprises: the routing information of each described virtual server;

Described first call instruction sends subelement, obtain subelement with described first configuration information respectively, described load unit is connected, for sending server calls order to each described virtual server in described load unit, described server calls order comprises: the information requiring the task that each described virtual server performs respectively, this task is specially the service that when analog subscriber is applied, this virtual server provides to user, wherein, each described load unit respectively with each user's one_to_one corresponding, and

For sending subtest call instruction to each described auxiliary test cell, each described load unit connected to described auxiliary test cell according to described call instruction for described auxiliary test cell sends call instruction;

First execution result determining unit, for obtaining each described virtual server to the actual execution result information of described task and the execution result information to described subtest call instruction obtaining described auxiliary test cell;

Described first storing sub-units, for storing information, described information comprises: the configuration information of each described virtual server and each described virtual server are to the execution result information to described subtest call instruction of the actual execution result information of described task and described auxiliary test cell;

Described performance parameter determination subelement, respectively with described first storing sub-units, the second storing sub-units in described auxiliary test machine is connected, for according to the data in described first storing sub-units and each described second storing sub-units, according to the actual execution result information of each described virtual server in each described load unit to each described task, determine the performance of each described load unit and each described virtual server, according to the execution result information to described subtest call instruction of each described auxiliary test cell, determine the performance of described auxiliary test cell,

Described auxiliary test cell is the unit of exerting pressure of simulation client terminal access behavior,

Described auxiliary test cell comprises: the second configuration information obtains subelement, the second call instruction sends subelement, the second execution result determining unit and the second storing sub-units;

Described second configuration information obtains subelement, for the control of the call instruction according to described main test cell, the configuration information of each described virtual server in the described load unit that acquisition is connected with described auxiliary test cell, the configuration information of described each described virtual server is stored to described second storing sub-units, and described configuration information comprises: the routing information of each described virtual server;

Described second call instruction sends subelement, obtain subelement with described second configuration information respectively, described load unit is connected, for the control of the call instruction according to described main test cell, server calls order is sent to each described virtual server in described load unit, described server calls order comprises: the information requiring the task that each described virtual server performs respectively

Second execution result determining unit, for obtaining the actual execution result information of each described virtual server to described task;

Described second storing sub-units, for storing information, described information comprises: the configuration information of each described virtual server and each described virtual server are to the actual execution result information of described task.