WO2002082727A1 - Method for collecting a network performance information, computer readable medium storing the same, and an analysis system and method for network performance - Google Patents

Abstract

This invention relates to a method for collecting network performance information, a computer readable recording medium storing the same, and a system and method for analyzing network performance. A client system including a manager program and agent programs receives schedule information from a network performance analysis server system, starts driving using the information, performs network testing, accesses a network or a site to execute a service test operation, and provides testing information to the network performance analysis server system. An ICMP unit of the agent program installs a trace-router function to transmit an ICMP echo message to a router on the target URL's path, and checks a response to find the router's existence on the path and its delay time. A WinInet unit of the agent program measures performance of CPs or ISPs under general Internet user environments. An analyzer of the agent program outputs measurement results using data measured by a measurement unit.

Description

METHOD FOR COLLECTING A NETWORK PERFORMANCE

INFORMATION, COMPUTER READABLE MEDIUM STORING THE SAME,

AND AN ANALYSIS SYSTEM AND METHOD FOR NETWORK

PERFORMANCE

CROSS REFERENCE TO RELATED APPLICATION

This application is based on Korea Patent Application No. 2001 -18292

filed on April 6, 2001 in the Korean Intellectual Property Office, the content of

which is incorporated herein by reference.

BACKGROUND OF THE INVENTION

(a) Field of the Invention

The present invention relates to a system and method for analyzing

network performance. More specifically, the present invention relates to a

method for collecting network performance information, a computer readable

recording medium installing the same, and a system and method for analyzing

network performance using the medium.

(b) Description of the Related Art

In general, Internet-related business categories include web contents

providers such as network service providers, portal service providers, and

broadcasting service providers; sales service providers or brokers for electronic

commerce and shopping malls; and equipment and solution providers.

In addition to these categories, they further include data center hosting

(IDC) service providers, web contents delivery (CDN) service providers, and performance measurement portal (PMP) service providers for measuring

Internet performance and web server performance, which complement each

other.

Recently, various protocols and services have been developed under

the open TCP/IP network systems, and the Internet and web environments

have also been greatly developed. Accordingly, methods for measuring

network performance together with the development of various and new

services are now becoming an important issue to the service providers and

service users.

Conventional web service providers also perceive the need for

monitoring the access performance and managing the network from the clients'

viewpoint as well as for the conventional internal system management.

Further, Internet equipment providers (e.g., PCs, servers, load

balancers, and repeaters) are gradually becoming more aware of business

methods that require pre/post services for clients through an analysis of

accurate performance and status of client systems on the PMP basis so as to

strengthen their competitiveness.

However, since access performance according to the current client

management methods used by various network service providers is far

different from access performance measured while using the network service,

there is an urgent need to manage network access performance that the clients

actually experience.

As the number of network service users has greatly increased, heavy loads have occurred in network servers, and the users have sometimes been

unable to access a server because of the bottleneck phenomenon on the

network, and hence, users' complaints have grown. Accordingly, it is now

required to manage the service satisfaction of users in addition to simply

providing the network services.

However, the management of companies that provide services to

network users are mainly focused on the management of databases provided

to the users and the management of the networks.

Since performance evaluated by the management of the networks is

different from the network service performance actually experienced by the

users, problems exist in that the network access performance that the users

generally experience is less than that which is ascertained by such

management.

SUMMARY OF THE INVENTION

It is an object of the present invention to provide a computer readable

recording medium storing a network performance information collecting

program for more fluently collecting basic data for managing networks such as

the Internet and PC networks and basic data for managing contents services

for users to thus easily obtain network using states, and for managing access

performance to the Internet or the PC networks.

It is another object of the present invention to provide a method for

collecting network performance information using the network performance information collecting program.

It is still another object of the present invention to provide a network

performance analysis system using the network performance information

collecting method.

It is yet another object of the present invention to provide a network

performance analysis method using the network performance information

collecting method.

In one aspect of the present invention, a method for collecting network

performance information comprises: (a) accessing a network performance

analysis server according to program starting, and downloading schedule

information including a starting time, a measuring target medium, a URL

(uniform resource location), and a test location used for testing the network

performance; (b) analyzing service response times for respective network

intervals and disconnection intervals on the basis of the schedule information

to test network or Internet devices; (c) accessing a corresponding service to

test homepage and file downloading; and (d) generating a log file on the basis

of respective test results of (b) and (c), and transmitting the log file to the

network performance analysis server.

In another aspect of the present invention, a computer readable

recording medium with a network performance information collecting program

comprises: a schedule module, including at least one test from among a

predetermined time range and a number established by a system manager of a

measuring side, for managing schedule information on performance measurement; an ICMP (Internet control message protocol) module having a

trace-route function to transmit an ICMP echo message to at least one network

or Internet device provided on a path of a measuring target URL on the basis of

the schedule information, and checking its response to check an existence

state and its delay time of the network or Internet device on the path; a Winlnet

module for measuring performance of CPs (contents providers) or ISPs

(Internet service providers) under an environment identical with that of normal

Internet use; and an analysis module for calculating measuring results on the

basis of data measured by the ICMP module and the Winlnet module.

The recording medium further comprises a manager program including

a TCP client module for transmitting the analysis log information using a

predetermined protocol.

The manager program controls generation and removal of at least one

agent program including an ICMP module, a Winlnet module, an analysis

module, a TCP client module, and a schedule module.

In still another aspect of the present invention, a network performance

analysis system comprises: a plurality of client systems for using a client

program installed in a client PC (personal computer) or a server-level computer

of a plurality of Internet users, accessing a measured system through the

Internet or the network, receiving network performance test information on web

sites and network performance, and outputting the network performance test

information; and a network performance analysis server system, including a log

server and a database, for providing predetermined schedule instructions for a network performance analysis to clients that are locally scattered and installed

on the system, receiving data collected according to its response, and

providing various categories of statistical data on the basis of the data.

In still yet another aspect of the present invention, a network

performance analysis method using a plurality of client systems for collecting

network performance test information on the basis of predetermined schedule

information and outputting the information and a network performance analysis

server system for providing various types of statistical information on the basis

of the information comprises: (a) downloading the schedule information of

items for measuring the network performance to the agent program when a

task number and user information are input to the manager program of the

client system and at least one agent program is generated; (b) checking

whether a format of log information is a previously established data format

when the log information of test data collected and provided on the basis of the

downloaded schedule information is transmitted; (c) outputting an error

message and going to (b) when the data format is not previously established;

(d) storing the data format in a predetermined database when the data format

is previously established; and (e) transmitting an ACK message to the manager

program and going to (b) when the data format is stored in the database.

The (b) comprises: (1 ) attempting to access a corresponding medium

on the basis of the schedule information; (2) testing a network or Internet

device through a service response time for each network interval and an

analysis of a disconnection interval; (3) accessing a corresponding service and testing downloading of a homepage and a file; (4) generating a log file on the

basis of each test result; and (5) disconnecting the access to the corresponding

medium when the test of downloading the homepage and the file is finished,

and going to (1 ) when a unit test process is provided in the schedule

information.

According to the network performance information collecting method,

the computer readable recording medium storing the method, and the network

performance analysis system and method using the recording medium,

network service performance and contents service performance may be

checked on the basis of network performance information used for accessing a

predetermined web site and service performance information provided to the

corresponding web site through an agent program, each installed in a client

personal computer.

BRIEF DESCRIPTION OF THE DRAWINGS

The accompanying drawings, which are incorporated in and constitute

a part of the specification, illustrate an embodiment of the invention, and,

together with the description, serve to explain the principles of the invention:

FIG. 1 shows a block diagram for a network performance analysis

system according to a preferred embodiment of the present invention;

FIG. 2 shows a client program according to a preferred embodiment of

the present invention;

FIG. 3 shows a signal flowchart between a TCP client module and a log server for describing data communication protocol;

FIG. 4 shows a flowchart for an operation of a manager program

according to a preferred embodiment of the present invention;

FIG. 5(a) shows a flowchart for an operation of an agent program

according to a preferred embodiment of the present invention;

FIG. 5(b) shows a flowchart for an operation of an agent program

according to another preferred embodiment of the present invention;

FIG. 6 shows a flowchart for describing a network performance test by

the client program according to a preferred embodiment of the present

invention;

FIG. 7 shows a flowchart for an access to a server according to a

preferred embodiment of the present invention;

FIG. 8 shows a flowchart for describing packet flows of a trace-route

according to a preferred embodiment of the present invention;

FIG. 9 shows a diagram for measuring a receiving speed and a cut

ratio according to a preferred embodiment of the present invention;

FIG. 10 shows a flowchart for a network performance analysis method

by a log server according to a preferred embodiment of the present invention;

and

FIG. 1 1 shows a network performance analysis server system

according to a preferred embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

In the following detailed description, only the preferred embodiment of

the invention has been shown and described, simply by way of illustration of

the best mode contemplated by the inventor(s) of carrying out the invention. As

will be realized, the invention is capable of modification in various obvious

respects, all without departing from the invention. Accordingly, the drawings

and description are to be regarded as illustrative in nature, and not restrictive.

FIG. 1 shows a block diagram for a network performance analysis

system according to a preferred embodiment of the present invention.

Referring to FIG^". 1 , the network performance analysis system

comprises a network performance analysis server system (NPASS) 100; a

plurality of client systems 200; and a plurality of measured systems 300.

The NPASS 100 comprises a log server 1 10, a database server 120,

and a web server 130, and it provides schedule instructions for performance

analysis to the clients, each of which is locally scattered and installed on the

system, and it receives test information that is collected in response to the

schedule instructions from them.

Also, the NPASS 100 generates various types of statistical information

on the basis of the test information, and provides the statistical information to a

user who requests a network performance analysis, through the web server

130.

Namely, it is desirable to provide statistical information to a network

manager for providing various services, and it is also preferable for the network manager to be able to process the statistical information such that it can be

used for comparison within their own company or within the same industry.

The client system 200 that is accessible to the Internet or the networks

comprises a manager program 210 and an agent program 220, and it is a

computer system positioned at a home for the network performance analysis

test, or a computer system or a server system installed at a predetermined

place of business for the network performance analysis.

Regarding its operation, the client system 200 receives the schedule

information from the NPASS 100, starts its operation on the basis of the

schedule information, performs a network test for measuring how the user

experiences the network, accesses a specific service provider or a site to

perform a service test for measuring how the user experiences the contents

service, and provides corresponding test information to the NPASS 100.

During the network test, as an example, the client system 200 receives

test information at hops that are provided on a path from the client system to a

specific site, and tests network service states such as network efficiencies on

the basis of the test information. In this instance, the hops include network

devices such as a server, a RAS (remote access server), or a router.

Also, during the service test, the client system accesses a

predetermined site via a plurality of hops, and downloads the first web page,

that is, a homepage, to the client system to thereby test contents service states.

The measured system 300 may be a system for managing a

homepage and a plurality of web pages. The Internet is exemplified for the measured system 300, but the

measured system may further comprise wire and wireless communication

networks, cable networks, optical communication networks, business intranets,

public broadband networks, and other public networks.

A realization method and an operation principle of respective modules

of the client program according to the preferred embodiment of the present

invention will now be described.

FIG. 2 shows a client program according to a preferred embodiment of

the present invention.

Referring to FIGs. 1 and 2, the client program 200 comprises: a

manager program 210 including a TCP client module 212; and an agent

program 220 including a schedule module 221 , an ICMP (Internet control

message protocol) module 222, a Winlnet module 223, an analysis module 224,

and a media connection module 225. Functions of the respective modules will

be separately described for ease of explanation.

The client program 200 accesses a web site according to a schedule,

downloads the first web page to the client program 200 or downloads a

predetermined file to the client program 200 according to the schedule, to

automatically measure service performance of the corresponding site and the

performance of the network through which the client program 200 accesses the

corresponding web site, and it transmits measurement results to the log server

110 of the NPASS 100.

It is desirable to install the client program in a computer system provided to the home of a high use Internet user, or in a server level system, or

to install the client program in various local regions (e.g., an IDC (Internet data

center) under various environments so as to measure in many ways how the

clients experience the network performance. In this instance, a single agent is

provided, and a plurality of client servers may also be set within the

requirements of the client system.

The respective modules in the client program 200 will now be

described in detail.

The TCP client module 212 transmits various types of analysis

information generated by the client program 200 to the log server 110 that is an

assigned management server using a protocol such as a TCP socket

communication. The log server 110 provides various categories of statistical

information to a system that requests a network performance analysis, on the

basis of data collected by the clients which are locally scattered and installed

on the system.

The log data tested by the client, that is, a kind of a protocol, are

transmitted to the log server 1 10 of the NPASS 100 to be stored in a database.

The client program 200 generates log data when each test is

terminated. The log data are stored in a client PC's HDD or a floppy disk and

are used for processing abnormal program errors and system errors and for

safely transmitting test results to the log server 110 through the Internet.

In order to remove various errors that may occur during transmission of

the log information, the log server 1 10 checks whether the log data transmitted by the client program 200 are correct, and transmits checking results to the

corresponding client program 200. That is, when receiving log data from the

client program 200, the log server 110 checks the log data and notifies the

client program 200 of an ACK or NAK response.

In the above-described preferred embodiment of the present invention,

the TCP client module 212 is installed in the manager program to perform TCP

socket communication, and it is also obvious that a UDP client module is

installed in the manager program to perform UDP socket communication.

In addition, a manager program and an agent program are individually

installed in the client program in the embodiment of the present invention, and

further it is obvious that they may be concurrently installed in the client program.

FIG. 3 shows a signal flowchart between a TCP client module and a

log server for describing data communication protocol.

Referring to FIGs. 1 and 3, the client program 200 waits for an ACK

response on the transmitted log data from the log server 1 10. If receiving the

ACK response from the log server 1 10, the client program 200 determines that

information is safely transmitted.

The log server 110 performs several tests on the transmitted log

information to check whether the data are lost or damaged on the network or

the Internet, and transmits test results to the client through an ACK or NAK

response. When the client receives a NAK message from the log server 1 10,

the client assumes that a problem is generated while a corresponding log

frame is transmitted, and the client again transmits the log information to the log server 1 10. This process is repeated for each identical time frame until the

client receives an ACK response from the log server 1 10. During this process,

subsequent errors may be assumed.

First, when the client transmits log information to the log server but

fails to receive an ACK/NAK signal from the log server 1 10 because of

disconnection of data communication with the log server 1 10, the client

program stands by indefinitely while waiting to receive the ACK/NAK signal.

To solve the first problem, the client program drives a timer, which is

timing out software. That is, when the client program transmits log information,

resets the timer, and receives no ACK/NAK signal from the log server 110

within a predetermined time frame, the client program processes this as time¬

out, and determines that no log information is transmitted to the log server 110.

This time-out function creates benefits, but it may cause the following

problems.

It is assumed that the client transmits log information and drives the

timer and this log information is transmitted through a low-speed network.

The log information is transmitted to the log server 1 10 through the

low-speed network, and it is processed to thereby generate a corresponding

ACK/NAK frame. The ACK/NAK frame is then transmitted to the client through

the low-speed network. During this process, however, a time-out may occur in

the client program because of a lot of time spent in the process.

A second problem may exist in that the frame synchronized state

between the log information of the client program and the ACK/NAK frame transmitted by the log server 110 may not be guaranteed because of the

ACK/NAK frame transmitting after the time-out.

In order to solve the second problem, a piggyback method is used for

the ACK/NAK frame. That is, "log time" information in the log information

transmitted to the log server 110 is provided to the ACK/NAK frame so that the

log time information and the ACK/NAK may be transmitted together.

Accordingly, by using the information, the client program checks to what log the

ACK/NAK frame transmitted by the log server 110 is matched.

Referring to FIG. 2, in order for the client program 200 to measure the

performance of the Internet or the network in various and effective ways, the

schedule module 221 receives schedule information on a test procedure and a

corresponding method from the log server 110 when driving the agent program

220. Using the schedule information, the NPASS controls the client programs

each of which is locally scattered and installed on the system.

For example, respective schedule items that a manager may control

using the NPASS may be configured so as to 1 ) perform a test during

predetermined time periods, 2) add a location of a measuring target URL and

modify it, 3) set various categories of information used for tests such as test

areas and media types, and 4) provide a different schedule to each client.

In particular, so as to perform the functions of 3) and 4), it is desirable

to separate the client program from the NPASS. To perform the separation,

each client program is assigned a serial number, and the NPASS controls the

clients' measuring operation on the basis of the serial numbers. The ICMP module 222 has a trace-route function to transmit an ICMP

echo message to respective network or Internet devices such as routers that

are provided on the path of the measuring target URL, and check

corresponding responses, thereby testing the routers' existence states on the

paths as well as corresponding delay times. Namely, the ICMP module 222

uses the trace-route function to find information on addresses of the hops (e.g.,

routers) on the corresponding paths, time delay factors between the hops, and

packet loss rates.

The client program may define ICMP items including TTL (time to live),

lengths of ICMP packets, number of ICMP packets, and time-out, and manage

the ICMP items.

The packet loss rates generated from the respective routers may be

measured by adjusting the number of ICMP packets. That is, when the ICMP

packet is transmitted to each router provided on the path of a predetermined

URL, each hop is required to echo the ICMP packet in normal cases. The loss

rate may be measured by counting the number of ICMP packets safely

transmitted to the client program by the router after being transmitted to the

hops.

For example, it is assumed that the user sets the ICMP items such that

TTL = 30, the length of the ICMP packet = 64 Bytes, the number of ICMP

packets = 10, and Time-out = 3000ms.

Under this state, when a trace-route operation is performed on a

predetermined URL, the ICMP packets are transmitted to the hops each provided on the corresponding path by the predetermined number of ICMP

packets (the number is set to be ten in this example). If a predetermined hop

transmits nine echo responses to the client program, the corresponding packet

loss rate is calculated as follows.

100% - (9/10)*100 = 10%

The Winlnet module 223 measures the users' network performance on

the networks or the Internet, and uses Winlnet functions of the Internet

Explorer (IE) installed in the Windows O/S (operating system) as a default. By

performing performance measurement on the basis of the Winlnet functions,

subsequent merits may be obtained.

That is, in the case specifications of general-purpose application

protocols (e.g., HTTP or FTP) of the TCP/IP are changed, the performance

measurement may be continuously performed using the updated IE Winlnet

functions without additional modification of the client program.

Also, the performance measurement may be performed under the

identical IE environments used by general Internet users.

The client program 200 reads the schedule set by the NPASS 100

through the schedule module 221 , and performs a test operation according to

the schedule.

The analysis module 224 generates various measurement result data

on the basis of the data measured by the respective measurement modules of

the client program.

The media connection module 225 performs a media access operation to wire and wireless terminals or modems in each test stage. In this instance,

the media access operation includes a first network access method for

accessing the network such as the Internet through computer communication

(e.g., the 104XY-phone number method via a modem) or through a wireless

Internet that requires additional access attempts by a user, and a second

network access method through a high-speed communication network or an

exclusive line that needs no additional access attempts by the user.

Through the above-described media access operation, the actual

access rate of the user receiving desired services through the network or the

Internet may be calculated. Furthermore, the client program separately

transmits notification of access failures that may occur and access failure

reasons to the measurement server, thereby providing notification of problems

that exist between wire and wireless intervals of the networks or the Internet,

and suggesting corresponding improvement ideas.

To execute data communication through the wire and wireless

terminals or the wire and wireless modems, an additional process for making

the Windows O/S recognize the above-noted terminals and modems is needed.

Accordingly, after making the Windows O/S recognize driver programs

according to individual methods corresponding to the respective media, the

user may drive respective wire and wireless devices through the "phone

access network" Windows program with no concern regarding the

characteristics of the corresponding media, which have been widely applied to

the conventional modem usage and recent ADSL devices. The client program uses the access methods of the wire and wireless

terminals to use the Windows standard RASAPI (remote access service

application programming interface) for directly controlling the "phone access

network" Windows program, and it also uses the characteristics of the modem

provided to each device to find the modem's access speed or problems

caused by access failure.

Items to be measured and managed by the client program that

manages services of an Internet or network management system will now be

described in detail. In particular, the measurement and management items

according to the first network access method through computer communication

such as the 014XY phone number method or through the wireless Internet, and

those according to the second network access method through the high-speed

communication network or the exclusive line, may be separately described.

First, information to be collected and processed by the client program

through the first network access method includes (1) access information, (2)

data communication information, (3) ICMP trace-route information, and (4)

other categories of information according to contents and characteristics of the

corresponding items.

Corresponding contents and configuration of each information set will

now be described.

(1 ) Access information

Probable error factors and various measurement parameters that may

be generated during attempting data communication by using a wire and wireless terminal or a wire and wireless modem, are provided in Table 1

Table 1

In particular, messages transmitted to an upper program from the TAPI

when an accessing state is disconnected will be subsequently summarized in

the case of the access error code.

For reference, the client program processes cases of when the

subsequent messages are transmitted from the Windows O/S as errors, and

leaves a log message for each corresponding case.

The client program is programmed to check an RAS access when the

access is abnormally disconnected while the subsequent messages are not

transmitted from the TAPI.

Also, when determining that the RAS access has failed, the client program leaves an error message of "User authentication has failed," and in

another case, it leaves a log of "The line has been disconnected normally."

LINEDISCONNECTMODE_NORMAL // Process as "Disconnect

access from remote place"

LINEDISCONNECTMODEJJNKOWN // Process as "Disconnect

unknown call connection"

LINEDISCONNECTMODE_BUSY // Process as "Line is busy"

LINEDISCONNECTMODE_NOANSWER // Process as "No answer

from another user's repeater"

LINEDISCONNECTMODE_NODIALTONE // Process as "No dial tone

detected"

RASDISCONNECTMODE_AUTHENTICATEFAIL // Process as "User

authentication failed"

The definitions and descriptions of the above-noted errors are

provided as documents to the MSDN (Microsoft Developer Network)

periodically issued by Microsoft Corporation.

(2) Data communication information

After successful access, the client program performs corresponding

detailed tests according to the test procedure transmitted from the log server

1 10 of the NPASS 100. Table 2 describes respective items generated through

the test procedure. Table 2

The transmission speed will now be described in further detail.

The download count represents a number of files that form the first

page, and the client processes text-based HTML files or general-purpose

images files (e.g., those of GIF and JPG formats) as download target files.

The download size indicates a number of total bytes of the files that

configure the first page of the test target URL.

The download speed shows a mean down speed (in bytes/sec)

calculated after actually receiving the contents of all the files that configure the

first page of the test target URL.

The error code will now be described in detail.

As an example, problems may be generated to the actual data

communication because of error factors that may be generated from a

predetermined server provided in a corresponding medium after access

through the medium is successfully performed. These error factors need to be

processed differently from the error factors generated when accessing the

medium. Namely, when the client program uses a GET instruction of the HTTP

so as to receive a predetermined file from a measurement server and no such

file is found from the measurement server, the measurement server transmits

an error code (e.g., 404) to the client program instead of contents of the file.

The error code transmitted by the server becomes a measuring target

item. Generally, the error code is dynamically processed by the corresponding

server or within the range of a protocol to be thereby transmitted to the client

program. The DNS Lookup will now be described.

The DNS count represents a number of times the DNS used for

modifying a URL that directs the location of the files (e.g., HTML files or image

files) into an IP (Internet protocol) address is called.

The DNS time indicates a total time spent calling the DNS that is

required for downloading the first page.

The server connection will now be further described.

The server count represents a number of HTTP server connection

tasks required for downloading files that configure the first page of a test target

URL.

The server time shows a total time used for a HTTP server connection

task required for receiving the first page.

(3) ICMP trace-route information

In order to make the relations of the hops provided on the data

transmission paths numbers, the client program uses an ICMP protocol at each

performance of the test to perform the trace-route, thereby creating items of

Table 3. The measuring basis of the respective items will refer to the ICMP

module description.

Table 3

(4) Other categories of information

Table 4 shows items that are transmitted to the NPASS and used as

important index keys for processing various statistical analyses and databases,

though they are not related to the actual measurement. Table 4

In the above, various types of information that the client program collects and processes through the first network access method are described.

In the case of the second network access method, an additional access

operation by a user is not required, and accordingly, it may be desirable to omit

the description of the first item, that is, the (1) access information.

FIG. 4 shows a flowchart for an operation of the manager program

according to the preferred embodiment of the present invention.

Referring to FIG. 4, when the manager program is executed in step

S110, a number of tasks and other categories of information are input in step

S112. In this instance, the number of tasks represents the number of agent

programs managed by the manager program. The other categories of

information include personal information (e.g., a name, an address, and a

telephone number) of the agent program's manager; network environments

(e.g., an IP address, an access medium, and a network speed in the case of a

fixed IP) under which the agent program will be operated; and information (e.g.,

an IP address of a main log server, and a URL of a schedule file) for

communication with the log server.

It is checked whether the number of tasks is provided in step S114,

and when it is not found, the agent program is executed in step S1 16, and it

returns to step S1 14.

When the number of tasks is provided, the manager program 210

performs a multitasking operation with the log server 110 and the agent

program 220.

An operation between the manager program 210 and the log server 1 10 will now be described.

When the number of tasks is provided in step S1 14, it is checked

whether test data to be transmitted to the log server 1 10 are provided in step

S120, and when the test data are not found, a standby mode is executed, and

when they are found, it is checked whether a network connection to the log

server 1 10 (e.g., a socket connection) is performed in step S122.

When it is found that the socket is not connected, a standby mode is

performed until the socket is connected, and when the socket is found to be

connected, the corresponding test data are transmitted to the log server 1 10 in

step S124.

It is checked whether an ACK message is received from the log server

1 10 in step S126, and when the ACK message is not received, an operation for

going to the step S124 and transmitting the test data again is repeated, and

when the ACK message is received, an operation for going to the step S120

and transmitting other test data to the log server 1 10 is repeated. In this

instance, the number of the process for going to the step S124 may be

controlled to be a predetermined number by the manager.

An operation between the manager program 210 and the agent

program 220 will now be described.

When the number of tasks is provided in the previous step S1 14, a

predetermined message for checking a driving state is transmitted to the agent

program in step S130, and it is checked whether an ACK message

corresponding to a message transmission is received from the corresponding agent program in step S132.

When the ACK message is not received, it is checked that the agent

program is not executed, and accordingly, the agent program is executed in

step S134. When the ACK message is received, it is checked whether a

predetermined time is passed in step S136. In this instance, the predetermined

time includes a period from a start time to an end time of driving the agent

program.

When the time is not passed, it goes to the previous step S130, and

when the time is passed, an end message is transmitted to the agent program

in step S138, and it goes to the previous step S130.

Referring to FIG. 2, a plurality of agent programs 220 may be installed

in the client PC, and when a client program that includes a manager program

and an agent program is installed in a server-level computer of huge capacity

and high-speed, the manager program may generate at least two agent

programs.

In the case at least two agent programs are generated as described

above, an operation of the log server 1 10 and at least two agent programs is

possible as well as an operation of the log server 1 10 and one agent program.

Here, it is obvious that the steps S130 through S138 perform a multitask

operation with each generated agent program.

FIG. 5(a) shows a flowchart for an operation of the agent program

according to the preferred embodiment of the present invention, and in detail, it

is a flowchart for an operation of the agent program that uses the second network access method through a very high-speed communication network or

an exclusive line that does not require an additional access attempt operation

by the user.

Referring to FIG. 5(a), the agent program is executed in step S210,

and a ping message is transmitted to check an activation state of the network

in step S212. In this instance, the ping message is used to test transmission

and receiving states between systems using the ICMP.

When the network is not activated, a standby mode is performed until

the corresponding network is activated, and when it is activated, schedule

information is downloaded from the log server 1 10 in step S214, and it is

checked whether the agent program is required to be refreshed before the

items set by the schedule are tested in step S216. In this instance, the

refreshing represents a function for terminating the agent program for each

predetermined time interval (e.g., a one-hour time interval), and the terminated

agent program is re-executed by the manager program. The refreshing function

previously protects the agent program's abnormal execution that may be

generated because of various reasons, and minimizes occurrences of the

abnormal execution.

When the agent program requires refreshing, the execution of the

agent program is terminated and reset, and when it requires no refreshing, a

ping message is transmitted to check the network activation state in step S218.

When the network is found to be not activated, a standby mode is

performed until it is activated, and when the network is activated, a trace-route is performed according to schedule information in step S220, and a URL

according to the schedule information is downloaded to receive test information

in step S222. For example, when the schedule information includes an

instruction to access a predetermined site, network test information may be

received from the hops used for accessing the corresponding site, and test

information may be received through a downloading operation of images and

tests that configure a homepage of the accessed site.

It is checked whether a manager program is started in step S224, and

when it is not started, the corresponding manager program is started in step

S226, the test information generated by performing the step S220 and the test

information generated by executing the step S222 is transmitted to the started

manager program in step S228, and it feeds back to the previous step S214. In

this instance, the process of obtaining the test information through the

feedback may be determined by the schedule information, and when the

schedule information includes no further information on the measured targets

to be tested, the process is obviously terminated.

FIG. 5(b) shows a flowchart for an operation of the agent program

according to a second preferred embodiment of the present invention, and in

detail, it is a flowchart for an operation of the agent program that uses the first

network access method through the 014XY telephone method or wireless

media that require an additional access attempt operation by the user.

Referring to FIG. 5(b), an agent program is executed in step S250, and

a schedule and operating environments are input in step S252. An access to a medium is attempted to check whether the access to

the corresponding medium is successful in step S254, and when the access

fails, the access is attempted until the access is successful, and when the

access is successful, an access to the main log server is attempted to check

whether the access to the log server is successful in step S256.

When the log-in access in step S256 is successful, corresponding log

information is transmitted, and a trace-route test is performed in steps S258

and S260. When the log-in access is checked and found to have failed, the

trace-route test is performed in step S260.

A URL download test operation is performed in step S262, test log

information is calculated in step S264, the access to the medium is cancelled,

and it goes to the step S254 to attempt an access to a medium according to

schedule information in step S266.

FIG. 6 shows a packet flow of the trace-route according to the

preferred embodiment of the present invention, describing the step S220 of

FIG. 5 in detail.

Referring to FIG. 6, when the agent program 220 provides an ICMP

message to the hops on the Internet or the network according to a schedule so

as to access a predetermined Internet site, the corresponding hops determine

the paths of the packets provided to the hops and reduce items of the TTL that

relates to the packets' lifetime.

As a result, when the TTL value of the packet provided to the hop is

set to be zero, the corresponding packet is no longer transmitted to the destination, but a time-out packet, preferably a time exceeded error ICMP

packet is transmitted to the corresponding packet's transmitter as an echo. The

response by the time exceeded error ICMP packet is based on the RFC792

standard documents.

Accordingly, when the terminal that performs the trace-route sets the

TTL value of the first ICMP echo message to be '1 ' and transmits it, the

terminal may receive a time exceeded error ICMP packet from the first router

on the path.

When the terminal transmits an ICMP echo message having the TTL

value as '2' to the destination, the terminal may receive a time exceeded error

ICMP packet from the second hop such as a router.

When the above method is repeated, the trace-route operation may be

performed up to the target hop connected to a predetermined Internet site,

information and delay time of all the hops (e.g., routers) provided on the paths

to the destination may be obtained, and test information such as a packet loss

rate may also be found.

FIG. 7 shows a flowchart for a network performance test by a client

program according to the preferred embodiment of the present invention, in

detail, describing a network performance test according to a client program that

uses the second network access method through a very high-speed

communication network or an exclusive line that requires no additional access

attempt operation by a user.

Referring to FIG. 7, when the client program is started, the NPASS is accessed to download schedule information including a starting time, a

measuring target, a URL, and a test location required for the test in step S310.

An attempt to access a medium according to the corresponding

schedule through a wire/wireless modem, a terminal, or a LAN card installed in

the client system in which the client program is installed is executed in step

S320.

An ICMP test is performed to analyze network service contents

including a service response time and a disconnection interval for each

network interval (i.e., a hop), thereby receiving test results in step S330.

An access to a corresponding service (e.g., a predetermined Internet

site) is executed to test a homepage and a file downloading process and

receive test results in step S340.

Various test results received through the steps S330 and S340 are

used to generate log information, and the generated log information is

transmitted to the log server 1 10 to store it in a database in step S350.

When the test of the networks or the service performance for receiving

a predetermined service according to the schedule information is finished, the

access to the corresponding site is disconnected in step S360, and it goes to

the previous step S320 so as to perform a next test according to the schedule

information, and thereby repeat a unit test process under an established

schedule.

In the above, a network performance test through the client program

that uses the first network access method using a computer communication (e.g., 014XY) or a wireless Internet that requires an additional access operation

by a user is described, and the network performance test through the client

program that uses the second network access method through a very high¬

speed network or an exclusive line that requires no additional access attempt

by the user excludes the steps S320 and S360 from the whole steps of FIG. 7,

which will no longer be described.

FIG. 8 shows a server access diagram according to the preferred

embodiment of the present invention, describing the step S320 of FIG. 7 in

detail.

Referring to FIG. 8, an access to a corresponding measured server

(i.e., "request HTTP connect") is performed so as to download a file installed in

the measured server by a test program, and the measured server transmits an

access permission signal (i.e., "accept HTTP connect") to the test program that

is responsive to the above-noted request.

When receiving the access permission signal, the measuring program

requests a test file from the measured server (i.e., "get test file"), and the

measured server provides a corresponding file to the measuring program (i.e.,

"ACK").

The above-described process is referred to as a "server connection,"

and when the process is successfully executed, it is defined as "server

connection succeeded," and the corresponding time spent is measured as a

server access spending time.

When an access to a file of the measured server from the test program is attempted and no response is provided from the corresponding server, it is

processed as "server connection failed."

The above-noted access process to the measured server is applied to

the HTTP type and general TCP type communication environments. Namely,

the communication protocol may apply identical methods to the Telnet, the FTP,

and the SMTP, as well as the HTTP.

FIG. 9 shows a process for measuring a receiving speed and a

disconnection rate, describing the step S340 of FIG. 7 in detail.

Referring to FIG. 9, when an access to the measured server is

successfully performed, information on the file to be downloaded is requested

from the measured server (i.e., "get test file header"), and file size information

(i.e., "ACK & header data") is received from the measured server, that is, an

HTTP server according to a corresponding response. The test program finds

the actual size of the file provided to the server using the file size information.

In this instance, the information is used so as to measure receiving speeds and

disconnection rates.

When collecting the basic file information, an actual download of the

corresponding file is requested on the HTTP protocol basis, and the file is

transmitted according to its response (i.e., "get test file connect," and "ACK &

real data.")

Regarding the actual downloading of the file, the test program

downloads the file contents to the corresponding server until the transmitted file

size information is satisfied. When the connection state to the server is disconnected or the server

does not respond for a long time when the size of the currently received file

has not reached the total file size at the time of receiving the file data, it is

processed to be a file download error.

When the total file is successfully downloaded, a receiving speed (or a

transmission rate) is measured in reference to Equation 1.

Equation 1

. . . , , . . file size receiving speed (or transmission rate) = downloading time

FIG. 10 shows a flowchart for a network performance analysis method

by the log server according to the preferred embodiment of the present

invention.

Referring to FIG. 10, the log server 1 10 is operated in step S410 to

reset a socket connected to the client in step S412.

Log information according to test data provided from the manager

program of the client program is received in step S414.

It is checked in step S416 whether the log information is formatted

data, and when the log information is not formatted, an error message is output

to the manager program of the client program in step S418, and it goes to the

step S414. In this instance, the reason for receiving formatted data is for the

client program and the log server 110 to transmit and receive established

protocols through a predetermined setting.

When the log information is found to be formatted data in step S416, it is transmitted to a database to be stored therein in step S420, an ACK

message that reports a normal receipt state is transmitted to the corresponding

manager program in step S422, and it goes to the previous step S414 to

receive log information according to test information.

FIG. 1 1 shows an NPASS according to the preferred embodiment of

the present invention.

Referring to FIG. 1 1 , the NPASS comprises a log server 1 10, a

database server 120, and a web server 130. The database server 120

comprises a daily report generator 121 , an access time analyzer 122, a

download time analyzer 123, a download speed analyzer 124, an agent

information analyzer 125, a homepage analyzer 126, a hop and error

information analyzer 127, a service contents modifier 128, and a help message

providing analyzer 129.

The NPASS receives, in a predetermined log format, network

performance test information from a plurality of client systems 100 and

contents service performance information from corresponding media, analyzes

each information set, stores the same, and provides each stored analysis

information set to a network performance analysis request system through a

web server 130.

The network performance analysis request system may also receive a

predetermined request, and access a predetermined medium through

predetermined hops at a predetermined time period to thereby provide

customized network performance analysis information. In particular, the format of various types of network performance analysis information is obviously

processed to be other types so that the user may easily check the

corresponding analysis information.

The respective analyzers of FIG. 1 1 are conceptually separated so as

to describe functional operations of the database server, and they are not

separated in the hardwired manner.

As described above, the performance of the international and global

Internet and network infrastructures may be measured and managed by using

the Internet performance measuring program.

Also, services such as consulting may be provided so that effective

contents services may be provided through information on various types of

indices used for receiving contents on the web via the Internet and the

infrastructures.

Further, by measuring and managing far-reaching Internet service

performance and web using states, the Internet performance may be enhanced

through providing equipment services and contents services.

While this invention has been described in connection with what is

presently considered to be the most practical and preferred embodiment, it is to

be understood that the invention is not limited to the disclosed embodiments,

but, on the contrary, is intended to cover various modifications and equivalent

arrangements included within the spirit and scope of the appended claims.

Claims

WHAT IS CLAIMED IS:

1 . A method for collecting network performance information comprising:

(a) accessing a network performance analysis server according to

program starting, and downloading schedule information including a starting

time, a measuring target medium, a URL (uniform resource location), and a test

location used for testing the network performance;

(b) analyzing service response times for respective network intervals

and disconnection intervals on the basis of the schedule information to test

network or Internet devices;

(c) accessing a corresponding service to test homepage and file

downloading; and

(d) generating a log file on the basis of respective test results of (b) and

(c), and transmitting the log file to the network performance analysis server.

2. The method of claim 1 , wherein (b) further comprises attempting to

access a corresponding medium on the basis of the schedule information.

3. The method of claim 2, wherein (d) further comprises canceling the

access to the corresponding medium when the test of (c) is finished, and going

to a step of attempting to access the corresponding medium when the

schedule information includes a unit test process.

4. The method of claim 1 , wherein (c) comprises:

(b-1 ) requesting an HTTP connection from the corresponding medium,

and receiving an HTTP connection permission message according to its

response; and (b-2) requesting a test file from the measured server according to

receipt of the HTTP connection permission message, and receiving an ACK

message according to its response.

5. The method of claim 1 , wherein in (b), a trace-route function is used

to calculate an address of a hop provided on a corresponding path, a time

delay factor between the hops, and a packet loss rate.

6. The method of claim 1 or 5, wherein in (b), Internet control message

protocol items including a TTL (time to live), an Internet control message

protocol packet length, a number of Internet control message protocol packets,

and a time-out are used.

7. The method of claim 1 , wherein (d) comprises:

(d-1) requesting a test file header from the corresponding medium, and

receiving file size information including corresponding header data together

with an ACK message in response to it; and

(d-2) requesting a test file from the measured server according to

receipt of the file size information, and receiving an ACK message according to

its response and the actual file data.

8. A computer readable recording medium with a network performance

information collecting program comprising:

a schedule module, including at least one test from among a

predetermined time range and a number established by a system manager of a

measuring side, for managing schedule information on performance

measurement; an ICMP (Internet control message protocol) module having a trace-

route function to transmit an ICMP echo message to at least one network or

Internet device provided on a path of a measuring target URL on the basis of

the schedule information, and checking its response to check an existence

state and its delay time of the network or Internet device on the path;

a Winlnet module for measuring performance of CPs (contents

providers) or ISPs (Internet service providers) under an environment identical

with that of normal Internet use; and

an analysis module for calculating measuring results on the basis of

data measured by the ICMP module and the Winlnet module.

9. The recording medium of claim 8, wherein the trace-route function

includes:

at least one address of a hop provided on a path;

a time delay factor between the hops; and

a packet loss rate.

10. The recording medium of claim 8, wherein at least one Winlnet

function of the Internet Explorer installed in the Windows driving system is used

so as to measure the performance.

11. The recording medium of claim 8, further comprising a manager

program including a TCP client module for transmitting the analysis log

information using a predetermined protocol.

12. The recording medium of claim 1 1 , wherein the manager program

controls generation and removal of at least one agent program including an ICMP module, a Winlnet module, an analysis module, a TCP client module,

and a schedule module.

13. A network performance analysis system comprising:

a plurality of client systems for using a client program installed in a

client PC (personal computer) or a server-level computer of a plurality of

Internet users, accessing a measured system through the Internet or the

network, receiving network performance test information on web sites and

network performance, and outputting the network performance test information;

and

a network performance analysis server system, including a log server

and a database, for providing predetermined schedule instructions for a

network performance analysis to clients which are locally scattered and

installed on the system, receiving data collected according to its response, and

providing various categories of statistical data on the basis of the data.

14. The system of claim 13, wherein the network performance test

information comprises at least one of delay measuring information, packet loss

measuring information, disorder rate information, and transmission rate

measuring information.

15. The system of claim 14, wherein the delay measuring information

comprises at least one of delay factor test information in a client system, delay

factor test information to the measured system on the network, and delay factor

test information in the measured system.

16. The system of claim 13, wherein the client program comprises: a schedule module, including at least one test from among a

predetermined time interval and a number set by a system manager of the

measuring side, for managing schedule information on performance

measurement;

an ICMP module, having a trace-route function, for transmitting an

ICMP echo message to at least one network or Internet device provided on a

path of a measuring target URL on the basis of the schedule information,

checking its response, and checking an existence state and a delay time of the

network or Internet device on the path;

a Winlnet module for measuring performance of CPs or ISPs under an

environment identical with that of general Internet use; and

an analysis module for calculating measuring results on the basis of

data measured by the ICMP module and the Winlnet module.

17. The system of claim 13, further comprising a manager program

including a TCP client module for using a predetermined protocol to transmit

analysis log information generated according to measuring network

performance to a predetermined management server.

18. The system of claim 17, wherein the manager program controls

generation and removal of at least one agent program including an ICMP

module, a Winlnet module, an analysis module, a TCP client module, and a

schedule module.

19. A network performance analysis method using a plurality of client

systems for collecting network performance test information on the basis of predetermined schedule information and outputting the information and a

network performance analysis server system for providing various types of

statistical information on the basis of the information, the method comprising:

(a) downloading the schedule information of items for measuring the

network performance to the agent program when a task number and user

information are input to the manager program of the client system and at least

one agent program is generated;

(b) checking whether a format of log information is a previously

established data format when the log information of test data collected and

provided on the basis of the downloaded schedule information is transmitted;

(c) outputting an error message and going to (b) when the data format

is not previously established;

(d) storing the data format in a predetermined database when the data

format is previously established; and

(e) transmitting an ACK message to the manager program and going to

(b) when the data format is stored in the database.

20. The method of claim 19, wherein (b) comprises:

(1 ) attempting to access a corresponding medium on the basis of the

schedule information;

(2) testing a network or Internet device through a service response time

for each network interval and an analysis of a disconnection interval;

(3) accessing a corresponding service and testing downloading of a

homepage and a file; (4) generating a log file on the basis of each test result; and

(5) disconnecting the access to the corresponding medium when the

test of downloading the homepage and the file is finished, and going to (1)

when a unit test process is provided in the schedule information.

21. The method of claim 19, wherein (b) comprises:

(1 ) testing a network or Internet device through a service response time

for each network interval and an analysis of a disconnection interval on the

basis of the schedule information;

(2) accessing a corresponding service and testing downloading of a

homepage and a file;

(3) generating a log file on the basis of each test result; and

(4) going to (1 ) when a unit test process is provided in the schedule

information after the test of downloading the homepage and the file is finished.

22. The method of one of claims 19 through 21 , wherein the schedule

information includes a start time, a measuring target medium, a URL, and test

location information used for a network performance test.

23. The method of claim 20 or 21 , wherein the step of testing

downloading of the homepage and the file comprises:

requesting an HTTP connection from a corresponding medium, and

providing an HTTP connection permission message according to its response;

and

requesting a test file from the measured server according to receipt of

the HTTP connection permission message, and receiving an ACK message according to its response.

24. The method of claim 20 or 21 , wherein the step of testing the

network or Internet device comprises calculating an address of a hop provided

on a corresponding path, a time delay factor between hops, and a packet loss

rate by using a trace-route function.

25. The method of claim 20 or 21 , wherein the step of testing the

network or Internet device comprises using Internet control message protocol

items including a TTL, an Internet control message protocol packet length, an

Internet control message protocol packet number, and a time-out.

26. The method of claim 20 or 21 , wherein the step of testing

downloading of the homepage and the file comprises:

requesting a test file header from the corresponding medium, and

receiving file size information including corresponding header data together

with an ACK message according to its response; and

requesting a test file from the measured server according to receipt of

the file size information, and receiving actual file data together with an ACK

message according to its response.

27. The method of claim 19, further comprising (f) providing various

types of statistical information processed so that the stored data format may be

applied to a network performance analysis requester.