KR100447041B1 - Method and system for making performance test scenarios and for testing, analyzing and motoring the performance of servers in multi-tier distributed computing environment. - Google Patents

Abstract

The present invention is a system for collecting, analyzing, and monitoring the performance of the server and the performance of the request resources of the server for each tier for the performance measurement of the server per tier in a multi-tier distributed environment. And a method, more specifically, a collection of request resources for testing the performance of a server using an agent server that is in communication with a server for each server of each tier, and the method and By using the collected request resources in the usual way, the performance of the servers in each tier is measured and analyzed, and the performance of the servers in the other tiers participating in the processing of the requested resources, A system and method for measuring, analyzing, and monitoring the performance of processed request resources. To this end, the present invention includes a controller server, an agent server, a repository server, a performance test monitor, and a user interface. In detail, the controller server includes a job scheduler that manages performance tests using collected request resources, and a workbench unit that collects request resources and performs performance tests on specific request resources. An analyzer unit that searches and analyzes performance information from the repository server, and generates a random number of virtual users to perform performance tests to deliver the performance information to the controller server, and to store it in the repository server. It includes a user tester and a performance test monitor that monitors the performance test status in real time, and the agent server can run on a system that is not the same as the server in charge, and also bridge the communication with the server in charge. Collecting performance information of the unit and the processed request resource and passing it to the repository It includes a management unit that, repository server is characterized in comprising a database portion for storing, managing the resource request and the job (Job) information and performance information.

Description

System and method for collecting client's request resource for performance measurement of server per tier in multi-tier distributed environment, and measuring, analyzing and monitoring performance information of server's performance and request resources for each tier {Method and system for making performance test scenarios and for testing, analyzing and motoring the performance of servers in multi-tier distributed computing environment.}

The present invention is a system for collecting, analyzing, and monitoring the performance of the server and the performance of the request resources of the server for each tier for the performance measurement of the server per tier in a multi-tier distributed environment. And a method, more specifically, a collection of request resources for testing the performance of a corresponding server by using an agent server that is in communication with a server for each server of each tier, and the above method. Using the collected request resources in a general way, the performance of the servers in each tier is measured and analyzed, and also the performance of the servers in other tiers participating in the processing of the requested request resources and the servers The present invention relates to a system and a method for measuring, analyzing, and monitoring the performance of request resources processed in.

Today's computing environment has evolved into a multi-tier, distributed environment of servers that serve clients, applications, databases, middleware, enterprise resource plans (ERPs), and customer relationship management (CRM) solutions. In the previous client-server two-stage computing environment, the performance management of the system was relatively easy, but the load of hundreds or thousands of concurrent users using the web can be irregularly generated. As the process takes place in a complex environment that handles the client's request and finally delivers the results to the client, system performance management becomes more difficult. In such a complex multi-tier distributed environment, there is a need for systems and methods that provide accurate testing, measurement and analysis of the performance of servers at each layer.

In this multi-tier distributed environment, performance management of each server mainly uses a method of measuring and monitoring performance information by analyzing the log of the server by running an agent in each server of each tier. However, this method requires that both the agent know the logging process and format of the target server, and satisfy the condition that the agent must be able to run on the same system as the target server. Recently, the server side measures the performance of distributed servers for processing a request resource of a client by using a log function of a Java application server using a Java virtual machine to servers communicating with the Java application server. This is done by measuring the performance of the server, and the client side is based on the web browser, but it is difficult to apply when the application on the server is a system implemented with CGI, PHP, Perl or database or other kind of server solution or middleware. On the other hand, when the client side is an application other than the web browser, it is difficult to collect, request, and measure and analyze the performance of the server for each of the requested resources. In addition, a method of measuring and analyzing the performance of the server by inserting a promised API (Application Programming Interface) into the source of the application program of the server and analyzing log information generated during execution by the inserted API is used. There is no standard way to do this, and there is a need to modify and recompile the source of the program.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and an object of the present invention is to use agent servers connected to each tier server and communication protocol in a multi-tier distributed environment in which multiple servers participate. By providing a system and method for the collection of request resources (resource) for the performance test of each tier server. Here, the execution system of each agent server does not have to be the same as the corresponding server.

In addition, another object of the present invention is to store the request resources of interest as a set in the repository server when collecting the request resources in the above method, to perform a performance test of the request resources of each set and a combination of these sets The present invention provides a system and method for assisting the creation of a new set of requested resources.

In addition, another object of the present invention is to perform a performance measurement by any number of virtual users using the request resources collected by any method, the server of the other tier participating in the performance of the request resources and processing of the request resources. It is to provide a system and method for measuring and analyzing the performance of these resources and the performance of the request resources processed in servers of different tiers.

In addition, another object of the present invention is to store a variety of performance information tested and measured by the above method in the repository server to query a variety of information (per day, by test conditions) information on the performance of the server for each tier of the server for each tier It is to provide a system and method to help improve performance.

In addition, another object of the present invention is to retrieve the request resources that serve as a reference to the performance of the server for each tier among the request resources collected by any method from the repository server to always perform performance measurements on the operating system performance of the operating system To provide a system and method for real-time monitoring in a sampling manner by the selected request resources.

In order to achieve the above objects, the present invention includes a controller server, an agent server, a repository server, a performance test monitor, and a user interface. In detail, the controller server includes a job scheduler that manages performance tests using collected request resources, and a workbench unit that collects request resources and performs performance tests on specific request resources. An analyzer unit that searches and analyzes performance information from the repository server, and generates a random number of virtual users to perform performance tests to deliver the performance information to the controller server, and to store it in the repository server. It includes a user tester and a performance test monitor for monitoring the performance test status in real time, and the agent server can be run on a system that is not the same as the server in charge, and also a bridge for communicating with the server in charge. And log management to collect performance information of the processed request resources and deliver them to the repository To include, and repository server is characterized in comprising a database portion for storing, managing the resource request and the job (Job) information and performance information.

1 is a schematic diagram of a system for collecting, managing, and measuring requested resources using a controller server, an agent server, a repository server, a performance test monitor, and a user interface, and managing and analyzing measured performance information according to the present invention. It is a block diagram which shows a structure.

Figure 2 is a block diagram illustrating an execution environment for a database server layer executed in one embodiment of the present invention.

3 is a block diagram illustrating a monitor execution environment for a three-tier server system executed in one embodiment of the present invention.

<Description of Symbols for Main Parts of Drawings>

10. Controller server

11. Virtual user 12. Workbench 13. Job scheduler 14. The analyzer

20. Repository

30. Performance test monitor

40. Agent

41. Web Agents 42. Middleware Agents 43. Database Agents

50. User Interface

60. Server

61. Web application server 62. Middleware server 63. Database server

70. Client Application

71. Web client 72. Middleware client 73. Database client

Preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. 1 is a diagram illustrating performance using a controller server 10, a repository server 20, a performance test monitor 30, an agent server 41, 42, 43, and a user interface 50 according to the present invention. This is a block diagram showing the general configuration of a system that collects and manages the client's request resources for testing and manages, analyzes, and monitors performance measurements and measured performance information using the collected request resources. As shown in Figure 1, the overall configuration of the server performance measurement, analysis, and monitoring system of the multi-tier distributed environment according to the present invention is the controller server 10, the repository server 20 and the performance test monitor 30 And an agent server 40 and a user interface 50, wherein the controller server 10 includes a virtual user 11 unit, a workbench 12 unit, a job scheduler 13 unit, and an analyzer 14 unit. The virtual user 11 unit includes a communication means as a client for each tier server, and the repository server 20 is composed of any relational database supporting standard SQL, and the performance test monitor 30 is composed of a window application, the user interface 50 is composed of a web browser. In addition, the agent servers 41, 42, and 43 for each tier are connected to the servers 61, 62, and 63 using the supporting protocols, and do not need to be executed on the same system as the connection server. Collecting performance information on the requested request resource, characterized in that for transmitting to the repository server (20). Using the system of the present invention having such a configuration, a user can measure and analyze server performance in a multi-tier distributed environment using only a web browser, which is a user interface 50, and a corresponding client application 70 of servers 60 for each tier. Can be monitored.

Collect and manage request resources for performance measurement

Performance measurement and analysis of server systems in a multi-tier distributed environment must first collect the request resources for performance measurement, in which case the measured system is accessible from the web client and not from the web client. Explain separately. First, the target multi-tier distributed server system can be accessed from a web client as shown in FIG. 1, and in this case, the user may use a web browser 50 and a workbench 12 as user interfaces. When the target web application server 61 is connected to the target web application server 61 through the web agent server 41 to perform a measurement target task, the web agent server 41 performs performance information of the user's request resources, that is, request URL, and each URL. Various information such as a stay time, a response time, a response code, a response size, an HTTP request header, an HTTP response header, and a cookie are stored in the repository server 20. The user may read the URLs stored in the repository server 20 using the user interface 50 and the workbench 12 to generate a set of request resources for various performance tests. Also, the web application server 61 connected to the web agent server 41 connects to the middleware server 62 connected to the middleware agent server 42 via the middleware agent server 42 to process the requested URLs. Procedure Call) method is requested. At this time, the performance information of each requested RPC method is stored in the repository server 20 by the middleware agent server 42. The user may again read and use a request resource (RPC method) to be used for performance measurement on the middleware server using the user interface 50 and the workbench 12. In addition, the middleware server 62 connected to the middleware agent server 42 requests a SQL from the database server 63 connected to the database server through the database agent server 43 to process the requested RPC methods. At this time, the performance information of each of the requested SQL is stored in the repository server 20 by the database agent server 43. The user may read and use the request resource SQL to be used for performance measurement of the database server 63 using the user interface 50 and the workbench 12 in various ways.

On the other hand, when the web client cannot access the multi-tier distributed system for performance measurement, the user may access and test the server through the agent server of the target server using the client application of the server of the tier to be accessed. Perform the function. For example, when only the database tier exists as shown in FIG. 2, in this case, the database agent server 43 may determine a stay time between SQLs, which are request resources of the database client 73, and each SQL request. Various performance information such as response time, response code, and response size are stored in the repository server 20. The user may read the SQLs stored in the repository server 20 using the user interface 50 and the workbench 12 to generate a set of request resources for performance testing.

The set of various request resources for each tier server collected by the above method is stored in the repository server 20 in a unit called task, and each request resource is the IP address of the request server and the number of executions. It also stores measurement conditions such as time intervals between request resources. For example, the HTTP request resource is described in detail, and the HTTP request header value such as URL, Query String, Cookie, Referer, User-Agent, Authorization, Content-Type, etc., Iteration, and Stay Time (between requests) Performance can be measured with various HTTP request header values and performance conditions by applying functions of sequence (from, to, increment), file ('data.txt'), random (from, to) to performance measurement conditions such as time). Let's do it. The sequence (from, to, increment) function is applied to the part of the string where this function is located while iterating, incrementing and decrementing from from to to value incrementally. The file ('data.txt') Function means to apply data by reading line by line while repeating data from tab-delimited data file (data.txt) passed as argument to the part of the string where this function is located. The random (from, to) function means that the function changes as it is applied to a random value between the from value and the to value while iterating over the portion of the string where the function is located. The job scheduler 13 combines the tasks to generate a job to perform a performance test. When combining tasks, set the number of executions for each task to enable various performance tests. In addition, tasks and jobs made by the above method are all stored in the repository server 20 for reuse, co-use, and management.

<Measurement and analysis of the processing performance of each request resource processed by other servers for processing the request resource for one request resource for a specific server using a user interface>

By using the request resources of the tier-specific server collected by any method, the user measures the performance of a specific request resource in a single user environment and the order and performance of the request resources processed by servers of different tiers to process the specific request resource. can do. In the case of FIG. 1, assuming that the web tier is the first tier accessed from the client, the user selects a request resource (URL) of interest from the collected request resources (URL) and selects the user interface 50 and the workbench ( 12) the performance measurement using the web agent server 41 transfers the performance information of the request resource (URL) processed by the web application server 61 to the repository server 20, the middleware agent server ( 42 transmits performance information of resources (RPC methods) requested to the middleware server 62 to the repository server 20 to process the resources (URL) requested by the web application server 61, and also the database The agent server 43 stores the performance information of the resources (SQL) requested by the database server 63 to process the resources (RPC method) requested by the middleware server 62. To pass on. Finally, the workbench 12 uses the analyzer 14 to perform performance information (contents) of request resources (SQL, PRC methods) executed in the server for each tier in order to process the resource (URL) initially requested by the user. , Number of times, processing time, order, etc.) are read and transmitted to the user interface 50. At this time, the workbench 12 and each of the agent servers 41, 42, and 43 set exclusive exclusive rights to ensure that only one performance test is executed, so that the first requested resource (URL) is a request resource of another tier. When deriving a request (RPC method, SQL), it prevents processing of resource requests even if they come from other users. On the other hand, even if the tier first accessed by the user is not a web tier, performance information for each tier can be measured and analyzed through a similar process as described above.

Measure and analyze the performance of servers in each tier, using performance tests for specific servers

Using the user interface and the requested resources for a particular server collected in any way, the user can measure the performance of that particular server using a means by which any number of virtual users can communicate as a client to that particular server. The performance of the specific server and the performance of the requested resources can be measured, analyzed and monitored. In this case, when resources are requested to servers of different tiers in order to process the requested resources from the specific server, agent servers for each tier collect performance information of resources requested from the server in charge of the repository. Will be delivered to the server. In this way, the performance of the server for each tier and the performance information of the requested resources can be read using the user interface. In this way, the user can perform a performance test on one particular server to measure, analyze, and monitor the performance of servers in different tiers and the performance of request resources processed in servers in each tier.

Preferred embodiments in this case will be described in detail with reference to the accompanying drawings. In the case of Fig. 1, it is assumed that the specific server is a web application server 61, and assuming that a set of request resources (URLs) for the web application server 61 are collected in the repository server 20. The user sets the virtual user 11 to perform a performance test on the web application server 61 by using the user interface 50 and the job scheduler 13, so that any number of virtual users can execute the web application server ( 61) to request resources (URLs). At this time, the web application server 61 requests the resources (RPC methods) to the middleware agent server 42 to process the requested resources (URLs), the middleware agent server 42 is requested resources The RPC methods are delivered to the responsible middleware server 62 for processing, and after processing, the performance information of each request resource (RPC method) is transmitted to the repository server 20, and also the middleware server. 62 requests resources (SQL) from the database agent server 43 to process the requested resources (RPC methods), and the database agent server 43 handles the requested resources (SQL) in charge of the database server 63. In this case, when the processing is completed by the database server 63, the performance information of the request resources SQL is transmitted to the repository server 20. Meanwhile, the virtual user 11 transfers the performance information of the requested resources (URLs) to the repository server 20 after the processing of each request resources (URLs). In this manner, since the repository server 20 transmits and stores performance information of request resources (URL, RPC method, and SQL) processed by the server for each tier, the user may use the user interface 50 and the analyzer 14. ) To query and analyze the performance information of the request resources (URL, RPC method, SQL) from the repository server 20. In addition, the performance test monitor 30 communicates with the controller 10, the performance information of each system delivered from the system of each agent server 40 and the test server 60 and the request transmitted from the virtual user 11 Receives the performance information of the resource and outputs it to the user interface 50 screen for monitoring.

Measure and analyze the performance of servers in each tier and the associated request resources processed by servers in each tier, using performance tests for specific servers.

As described above, the performance of a specific request resource in a single user environment and the order and performance of request resources processed by servers in different tiers to process the specific request resource may be measured. A user can perform a performance test on one particular server to measure, analyze, and monitor the performance of the servers in the other tiers involved and the performance of the request resources processed in the servers in each tier. The combined measurement of these two methods, i.e., using request resources collected by any method, allows the performance of a particular server's request resources by any virtual user and the use of other tiers for processing the request resources. It can measure, analyze, and monitor performance information such as response order, response code, response size and response time of request resources processed by server.

Preferred embodiments in this case will be described in detail with reference to the accompanying drawings. 1, the web application server 61, the middleware server 62, and the database server 63 exist, and the performance test is performed on the web application server 61 using any number of virtual users. In this case, first, the virtual user 11 performs a performance measurement of each request resource (URL) for the web application server 61 by using the workbench 12 in a single user environment. In order to process a specific request resource (URL), the repository server displays the order and contents of the request resources (RPC method) processed by the middleware server 62 and the request resources (SQL) processed by the database server 63. Save to 20. In this way, the virtual user 11 measures the performance of any number of user environments after measuring the single user environment. At this time, the agent server 40 of each tier transfers the performance information on the requested resources (URL, RPC method, SQL) to the repository server 20. After the measurement by any number of user environments is finished, the user can use the user interface 50 and the analyzer 14 to perform information on performance tests conducted in a single user environment for the first time and performance tests conducted in any number of user environments. The combined order of information, the processed order, response code, response size, and response of the request resources (RPC method, SQL) processed by the server for each tier for each request resource (URL) processed by the web application server 61. Read out performance information such as time.

<Management, reading and analysis of measured performance information>

All the performance information measured by the above method is stored in the repository server 20, so that the user can inquire various performance information using the user interface 50 and the analyzer 14. In detail, the performance information of a specific request resource measured using the workbench 12 in a single user environment can query performance information for each time zone such as time, date, month, and year, and all measured requests. The top-k ranked query can be searched for resources in order of performance information such as response time, response size, and bytes per response time, and performance is measured using the virtual user 11 in a multi-user environment. The information is also stored in the repository server 20, so that the performance information of the request resources of all the tier-specific servers are measured in order of performance information such as response time, response size, and bytes per response time. Changes in the performance information and the number of virtual users by time of measurement by using Top-k Ranked Query and merging performance information by performance test condition and merging of performance information Performance information, performance gap classified information, the continuity between the requested resource (Persistence) can be found for specific performance information. In addition, the performance of the server can be compared and analyzed using a comparison function of the performance information merged by the above method.

<Monitoring and analyzing performance by tier of operating system>

By executing the agent server 40 in the server 60 for each tier of the operating system, the performance information on the request resource delivered by each agent server 41, 42, 43 stored in the repository server 20 is inquired. Monitor and analyze the performance of servers 61, 62, and 63. If the agent server 40 for the operating system brings a degradation in the performance of the server 60 system, the agent server 40 is not used. Referring to the case of FIG. 3 as an example, among the request resources collected by any method, the request resources that have a significant effect on the performance of each server 60 are selectively read from the repository server 20, The scheduler 13 and the virtual user 11 perform constant performance measurements. In this case, the performance test monitor 30 may be monitored in a sampling method for predicting the performance of the target server 60 in real time only with respect to the performance of the selected request resource, and the repository server 20. Performance information on the requested resource stored in real time can be inquired using the analyzer 14 and the user interface 50 to monitor the performance of a more detailed operating system.

The present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the scope of the technical idea of the present invention. For example, the communication server of the agent server is described as an XML database server supporting HTTP (HTTPS), a web server, an ERP solution, a customer relationship management (CRM) solution, or a CORBA supporting Internet Inter-Object Protocol (IIOP). Common Object Request Broker Architecture (EJP) middleware or Enterprize Java Beans (EJB) server that supports Remote Method Invocation (RMI) or File Transfer Protocol (FTP), TELNET, Simple Mail Transfer Protocol (SMTP), Post Office Protocol 3 (POP3), Multiple servers exist that support various protocols such as Internet Message Access Protocol (IMAP), Network News Transfer Protocol (NNTP), Microsoft Media Service (MMS), Real Time Streaming Protocol (RTSP), and Firewall Protocol (SOCKS). Applicable for tier distributed systems

As described above, according to the present invention, the performance of the servers of each tier and the performance of the processed request resources are tested, measured, analyzed, and monitored in a complex multi-tier distributed environment of a plurality of servers. It can help improve the performance of tiers of servers. In addition, the agent server that collects the performance information of the servers of each tier and delivers it to the repository server and the servers of each tier are connected as long as the supporting protocol is satisfied. By measuring, analyzing, and monitoring the performance and the performance of the requested resources, it is possible to help improve the performance of the server for each tier. In addition, all request resources and performance measurement conditions and measured performance information about the server are all stored and managed in the repository server, which can be helpful for testing, measuring, and analyzing server performance. In addition, performance tests are always performed on the operating system of the server by selectively using request resources that have a significant influence on the performance of each server, so that performance monitoring and analysis of the operating system can be performed without putting a large load on the operating system. Can be.

Claims (5)

In a distributed multi-tier computing environment, a method of collecting clients and processing request resources (URL, SQL, RPC methods, etc.) generated by clients and servers used for measuring performance of servers or reproducing communication statuses, Executing each agent responsible for communicating with each server; The application running on each server is configured to use the IP address and port number of the agent in place of the communication IP address and port number of the other server, in order to communicate with other servers, Accessing a specific agent using a user interface or a client application, requesting a requested resource from a specific server in charge of the accessed specific agent; Transmitting, by the specific agent, the requested request resource to the server, and storing information about the requested resource processed by the repository server after receiving the response; Requesting another request resource from a second server in charge of a second agent by generating another request to another second agent to perform a request resource delivered from the specific agent; , Transmitting, by the second agent, information about the requested resource mediated to the second server to the repository server; In this way, using the agent for each server by using the agent to transfer information about the requested resources requested to the respective servers from the client or other servers to the repository server, and In this way, in a distributed multi-tier computing environment, when a specific client requests a request resource from a specific server, not only the request resource requested by the specific server, but also the specific server to another server to process the request resource, And collecting the other requested resources between the servers using each agent and storing them in the repository server. A method of collecting requested and processed request resources to servers in a distributed multi-tier computing environment, comprising: In a distributed multi-tier computing environment, when a specific client requests a request resource (such as a URL) from a specific server and receives a response, the specific server sends a request to another second server to process the requested request resource (such as a URL). Requesting another requesting resource, and the second server first requesting the specific server from the specific server, such as requesting another requesting resource from a third server to process the requested requesting resources. In order to receive the response of the request resource (URL, etc.), the processing order, response time, response size, etc. of request resources (URL, SQL, RPC method, etc.) serially requested and processed to each server and between the servers In how to find out the performance information, Executing each agent responsible for communicating with each server; The application running on each server is configured to use the IP address and port number of the agent in place of the communication IP address and port number of the other server, in order to communicate with other servers, Accessing the workbench using a user interface, bringing the workbench into a single user state, and requesting the workbench to perform performance tests on specific request resources (URLs, etc.); The workbench using a client means having a function of requesting the requesting resource (URL, etc.) to the corresponding server, accessing a specific agent in charge of communication with the corresponding server; The workbench transmits a request to the specific server by using the connected specific agent; Receiving, by the specific agent, a response to the requested request resource (URL, etc.), delivering performance information to a repository server, and storing the performance information; Generating, by the specific server, another request to the second agent to perform the request resource (URL, etc.) delivered from the specific agent, and communicating with a second server in charge of the second agent; Receiving, by the second agent, a response to the requested resource mediated to the second server, delivering the performance information to the repository server, and storing the performance information; In this way, a repository agent is used for each server to store performance information on request resources (URL, SQL, RPC method, etc.) requested and processed from the corresponding client or other servers to the respective servers. Forwarding it to the server and storing it; In this way, the performance information of the request resources processed in each tier server stored in the repository server is read, and the performance information such as the processing order, response time, response size, etc. of the request resources processed in each tier server is analyzed. Steps to Analyzing performance information such as processing order, response time, response size, etc. of request resources processed by each tier server to make a response to the first client's request in a distributed multi-tier computing environment Way. In a distributed multi-tier computing environment, the average response time, transactions per second (TPS), and bytes per second (BPS) are achieved using a set of request resources (URL, SQL, RPC methods, etc.), virtual users, and user interfaces, commonly referred to as load test scenarios. In the method of finding performance information such as server performance, response time, response size, etc. for each of the request resources processed by each server, Executing each agent responsible for communicating with each server; The application running on each server is configured to use the IP address and port number of the agent in place of the communication IP address and port number of the other server, in order to communicate with other servers, Using the user interface and the job scheduler, preparing and scheduling a test scenario Job to prepare for a load test, The virtual user delivers and loads a request to a specific server by using a client means having a function of requesting each requesting resource (URL, SQL, RPC method, etc.) to a corresponding server according to a scenario set in a job. Generating a load test and Generating a further request to another agent to communicate with a second server that the agent is responsible for, in order for the particular server to perform the requested resources requested from the virtual user, Collecting, by the agent, performance information of request resources made to be processed by the second server and forwarding the stored information to a repository server; In this way, each server, except for the server that received the first request from the virtual user, transfers the performance information of the requested resources requested and processed by the client or another server to the repository server by using a responsible agent. To do that, The virtual user receives a response to the request resources from the specific server and delivers the performance information to the repository server for storage; In this way, by reading the performance information of the request resources processed in each tier server stored in the repository server, the performance of each server, such as the average response time, TPS, BPS of each tier server, and each tier server Analyzing performance information such as processing order, response time, response size, etc. Method for measuring and analyzing the performance of the server of each tier in the distributed multi-tier computing environment, and the individual performance of the request resources (URL, SQL, RPC methods, etc.) processed by each tier server. delete A controller server 10, an agent server 40, a repository server 20, a performance test monitor 30, and a user interface 50, The controller server 10, Means (11) for generating any number of virtual users and communicating with servers for each tier; Means (12) for communicating exclusively with the agent server (40) of the server for each tier to enable performance measurement of request resources for each tier; Means (13) for reserving and performing a performance measurement scenario JOB of various conditions; Means for communicating with the performance test monitor 30 to transmit performance information of each server; Means for reading information of request resources stored in the repository server (20); Means (13) for generating a new set of requested resources by combining the set of requested resources collected for each tier-specific server; And means (14) for reading performance information in association with the repository server (20), The agent server 40, Means for connecting with a linked server using a supported protocol; Means for collecting the performance information of the requested resource to the connected server and transmitting it to the repository server (20); Means for making an exclusive exclusive setting to make a single user environment; The repository server 20, Means for communicating with the controller server (10) to store request resources and performance information; Means for communicating with the agent server 40 to store the requested resource and performance information; The performance test monitor 30, Means for communicating with the controller server 10 to receive performance information of each server for each tier; And means for displaying the performance information of the server received by communicating with the controller server 10 on the screen. The user interface 50, Information transfer means between the controller server 10, the agent server 40, the performance test monitor 30, and the user. Server performance measurement, analysis, and monitoring system for distributed multi-tier computing environment, including