JP2003178040A - Website configuration decision support method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a Web site constitution determination supporting method capable of supporting the proper enhancement or addition necessary for the Web site on the basis of the objective information. <P>SOLUTION: In this Web site constitution determination supporting method for supporting the constitution of each group of servers of the Web site 100 composed of a Web server group 101, an AP server group 102 and a DB server group 103, each processing time of each server group 101/102/103, CPU utilization ratio and memory utilization ratio are detected, and the number of addition of the servers forming the server group is calculated on the basis of the number of connected users and the number of CPU clocks in a state of being not over a threshold value, with respect to the server group over the threshold value, when the processing time, the CPU utilization or the memory utilization is over the predetermined threshold value by more than the predetermined number. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to a web.
The present invention relates to a website configuration determination support method that supports a configuration change of each server group of a website including a server group, an application (AP) server group, and a database (DB) server group.

[0002]

2. Description of the Related Art Generally, a website accessed by an unspecified number of users via the Internet is a group of web servers directly accessed by users, and an application connected to the group of web servers for storing and executing application software. (AP) server group, and a database (D) that stores various data that is accessed from the web server group and the application (AP) server group.
B) Server group. 1 for each server group
It is composed of one or more servers.

In a website constructed in this way, the load on each server group varies according to the connection status from an unspecified number of users, so it is difficult to determine the number of times of peak access, and an unspecified number of users. Due to the nature of the Web, which is accessible by users, the number of accesses often changes dramatically due to factors such as new product announcements.

For this reason, the actual number of visits to the website fluctuates with respect to the number of visits predicted at startup,
When the load increases, the response to the user may become extremely slow, so it is necessary to constantly consider and review whether the website server configuration is sufficient. This review includes "enhancement" for improving performance in the server such as memory expansion and disk capacity expansion for a certain server, and "expansion" for newly adding the number of servers.

Conventionally, such a review of the structure of a website is currently carried out by a system administrator by monitoring changes in the number of accesses to the website and variations in response time and using human intuition and experience. is there.

[0006]

The revision of the website structure according to the above-mentioned prior art is largely due to human intuition and experience, so it is not a sufficient scale, but conversely, it is necessary to review more than necessary. There was a problem that a lot of resources were wasted due to the operation.

Further, since the conventional review is carried out manually, it is impossible to reinforce or add it at an appropriate time, so that a business opportunity is missed and a profit to be raised cannot be raised.

[0008] The present invention is to eliminate the above-mentioned problems caused by the prior art, and to assist in the suitable configuration or expansion of a website required for the website based on objective information. Is to provide a method.

[0009]

In order to achieve the above object, the present invention provides a website for supporting the configuration of each server group of a website including a web server group, an AP server group and a DB server group. In the method for supporting configuration determination, the processing time, the CPU usage rate, and the memory usage rate of the web server group, the AP server group, and the DB server group are detected, and the processing time, the CPU usage rate, or the memory usage rate is a predetermined threshold value. If the number exceeds the predetermined number, the number of additional servers is calculated based on the number of connected users and the number of CPU clocks for the web server group, AP server group and DB server group that have exceeded the threshold value so as not to exceed the threshold value. This is the first feature.

Further, in the second aspect of the present invention, in the above-mentioned method for determining the composition of a website, the CPU processing speed and the memory capacity of the server to be added when the number of additional servers is calculated are used as the coefficients. In the website configuration determination support method of any one of the features, the processing time, the CPU usage rate, and the memory usage rate of the web server group, the AP server group, and the DB server group in the calculated additional number of servers are detected, Monitor whether the processing time or CPU usage rate or memory usage rate exceeds a predetermined threshold value or more,
The third feature is that the data is used at the time of the next review.

[0011]

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, one embodiment of a website configuration determination support method according to the present invention will be described with reference to the drawings. FIG. 1 is a conceptual diagram of a system configuration that realizes a website configuration determination support method according to the present embodiment, and FIG. 2 is a system configuration diagram of an embodiment of a monitoring / analysis machine group that monitors and analyzes a web server. is there.

First, as shown in FIG. 1, the system configuration to which the method for supporting the configuration determination of the website according to the present embodiment is applied, a plurality of user client PCs 104 used by general users, the user client PCs 104 and the public network. Website 100 connected via 108
And a monitoring / analysis machine 109 that is connected to the website 100, monitors the usage status of the website 100, which is a feature of the present embodiment, and supports a suitable expansion or expansion required for the website 100. Composed of.

The website 100 is directly accessed from the user client PC 104 and comprises a plurality of web servers 101 for providing web pages to the client PC 104.
And an application (AP) server 102 that is accessed from the web server group 101 and includes a plurality of servers that store and operate application software, and a plurality of applications that store data accessed from the web 101 and the AP server 102. And a DB server group 103 including a database (DB) server of the above. Although it has been described that the server group includes a plurality of servers, the number of servers may be one.

The monitoring / analyzing machine 109 has the above-mentioned 3
It stores a processing time (response time to access, etc.) of one server group 101 to 103, a monitoring server PC 106 for monitoring the CPU usage rate and memory usage rate thereof, information monitored by the monitoring server PC 106, and the like. And a monitoring information storage database (DB) 107 for monitoring, and an analysis client for monitoring the usage rate of the server groups 101 to 103 based on the monitoring result by the monitoring server PC 106 and for supporting system expansion and expansion. P
And C105.

Next, the function of the monitoring / analysis server 109 will be described with reference to FIG. First, the monitoring server PC
As shown in FIG. 2, 106 is a monitoring information collecting function 21 that collects the processing time (response time to access, etc.) of each of the server groups 101 to 103 collected by the monitoring server 106, the CPU usage rate, and the memory usage rate. , A threshold crossing determination function 22 for determining whether or not a threshold such as a processing time described later has been exceeded
With. The monitoring information is often provided as a function of the OS. For example, in the case of Windows (registered trademark) NT, the function of the performance monitor corresponds to that, and when it is acquired by another dedicated monitoring application. There is also.

Next, the analysis client PC 105
Sets a generally acceptable threshold value of the processing speed / CPU usage rate / memory usage rate of each server 101 to 103 (for example, processing time is 10 seconds, CPU usage rate is 70%, memory usage rate is 70%). A threshold value information setting function 31 and a performance difference information setting function 32 for setting a performance difference such as a processing speed of the equipment to be added (speed difference or drive clock number due to model difference for CPU, drive clock number or capacity for memory) The monitoring information storage DB 107 includes a server monitoring information 43 that stores the monitoring information collected by the monitoring information collecting function 21, and a threshold crossing determination function that determines whether the monitoring value (detection value) exceeds a predetermined threshold value. 22, the monitoring information storage DB server 107 is a server that is monitoring information (processing time, CPU usage rate, memory usage rate) collected by the monitoring server PC 106. Visual information 43 and the determination function 22
The monitoring information 41 including the threshold crossing information 44 that stores the number of times that the detected value has exceeded the threshold, the threshold setting information 45 that is the threshold set by the threshold information setting function 31 of the analysis client PC 105, and the performance. The setting information 42 including the performance difference information 46 set by the difference information setting function 32 is stored.

Monitoring / analyzing machine 1 configured in this way
In 09, the analysis client PC 105 has each server group 1
The threshold information setting function 3 is used to set a threshold value such as a processing speed to 01 to 103.
1 is input and is stored as the threshold setting information 45 of the monitoring information storage DB 107, the monitoring server PC 106 uses the monitoring information collection function 21 to set the server group 1
The processing times 01 to 103 / CPU usage rate / memory usage rate are collected at predetermined time intervals and stored in the server monitoring information 43 of the monitoring information DB 107, and the stored monitoring information is judged by the threshold crossing judging function 22, for example, The number of times the threshold value has been exceeded is determined, and the monitoring information storage D is used as the threshold value excess information 44.
It operates so as to be stored in B107.

Next, the detailed operation of the website composition determination support method configured as described above will be described with reference to FIG. 3 and subsequent figures. FIG. 3 is a flow chart showing the procedure for analyzing the current state of the website configuration determination support method according to the present embodiment, FIG. 4 is a diagram for explaining the procedure for supporting configuration determination at the time of expansion / expansion, and FIG. 5 is the current state analysis. FIG. 6 is a diagram showing an example of an analysis table created at this time, and FIG. 6 is a diagram showing an example of website performance difference information at the time of past expansion / expansion.

As shown in FIG. 3, the website configuration determination support method according to the present embodiment is used when performing a current state analysis which is a processing procedure for determining whether or not the current website needs to be augmented or expanded with a server. , A specific monitoring item (processing time or CPU usage rate or memory usage rate) is set (step 101), and the analysis range thereof, for example, an analysis period of 24 hours / week / month is set (step 10).
2) The server monitoring information 4 indicating the number of times a predetermined threshold is exceeded and the number of measurements in the set monitoring items and analysis period.
Get from 3.

The number of times the thresholds are exceeded and the number of measurements are, for example, three DB servers and four measurements per hour (15
The number of times the threshold value is exceeded is 50
In the case of times, 3 (units) x 4 (times / hour) x 24 (hours) =
288 times is the number of measurements per day. Monitoring server PC1
06 acquires these values from the server monitoring information 43, and calculates a threshold value excess ratio value based on the number of times the threshold has been exceeded and the number of measurements (step 103). This value is 50 (times) / 288 when the number of times the threshold is exceeded and the number of measurements are taken.
(Times) = 17%.

Then, the monitoring server PC 106 determines whether or not the threshold value exceeding ratio value is a level at which predetermined enhancement / expansion must be performed (referred to as an enhancement reference level) (step 104). For example, the standard level of enhancement / expansion is 1.
It is set to 0%, and it is determined whether or not the threshold crossing ratio value exceeds this level. In this case, the measured value is 17% with respect to the strengthening reference level of 10%, so there is a need for strengthening / extension. It is determined that there is, and in this case, the number of times the threshold is exceeded / the number of times of measurement / the threshold excess ratio value is stored in the monitoring information complete payment DB 107 as the threshold excess information 44, and marking is performed.

The processing of these steps 101 to 105 is as follows.
It is also executed in other setting items (CPU usage rate or memory usage rate) and other analysis period (1 week or 1 month).
These analysis results are created as a current status analysis table (steps 106 to 108).

This current state analysis table is shown in FIG. This current status analysis table is for each server group (web server, AP server, DB
Servers) and websites show how much each monitoring item (response / processing time, CPU usage rate, memory usage rate) exceeds the threshold within the past 24 hours, 1 week, and 1 month. Here, the guideline for expansion / expansion (enhancement / expansion standard level) is divided into two stages, and "Caution"
Is set to "5%" and "essential" is set to "10%".

In the present status analysis table, for example, although the uppermost web site as a whole has fallen below the threshold values for all monitored items in the past month and week, the past 24
Focusing on the time, it was found that the response processing time was 0.13, which exceeded the extension reference level 0.10. Focusing on each server group below this stage, the reason why the processing time exceeded was the past of the DB server group. The processing time of 24 hours is 0.
It turns out that the main cause is 33 (33%). That is, by creating the current situation analysis table shown in FIG. 5, where on the website is the weak point,
That is, it is possible to determine what should be strengthened.

This judgment is made by the monitoring / analysis machine 109.
However, the decision is made in accordance with the procedure for supporting the configuration at the time of expansion / expansion shown in FIG. This procedure is a processing procedure for studying the scale of how many servers and how many servers should be introduced at the time of server expansion / expansion. As a premise, the performance difference information necessary for estimating the scale is shown in FIG. It will be described with reference to FIG.

This performance difference information indicates when and what was added / expanded, and as a result, how much performance difference occurred in the website. Item 1 (No. 1) is 2001/02/2001.
There are 2 web / AP / DB servers each at the time of launching the 01 website, and the CPU of these servers is 600
With a Pentum III (registered trademark) of MHz, each memory has a configuration of 386 MB. With this configuration, when the number of user connections is 400 users, the processing time of the web server group is 3 seconds / CPU.
Usage rate is 60% / Memory usage rate is 50%, AP server group processing time is 4 seconds / CPU usage rate is 70% / Memory usage rate is 60%, DB server group processing time is 5 seconds / CPU usage rate Indicates that the memory usage rate is 80% / memory usage rate is 70%. Similarly, in other items, the number of user connections / processing time / CPU usage rate / memory usage rate is as described in FIG. 6, respectively. Is shown.

Further, "front and back" at the right end of Fig. 6 indicates the values before and after the configuration set for the item, for example, from item 3 (No. 3) to item 4 (No. 4) in Fig. 6. The expansion / addition of
A new AP server was added on April 20, 2001.
P server group processing time from 12 seconds to 7 seconds, CPU usage from 60% to 30%, memory usage from 40% to 2
It means that each is improved to 0%.

Next, a procedure for supporting the determination of the configuration at the time of augmentation / addition will be described with reference to FIG. Here, the example of the current situation analysis result is taken over and the item 3 of 2001/15 /
An example of increasing and expanding a DB server from 05:00 to 4/20/2001 of item 4 will be described. First, this processing searches and retrieves information (item 1 and item 2) related to the DB server enhancement and expansion performed in the past from the performance difference information (step 2).
01).

When the site shown in FIG. 6 is opened (item 1
Focusing on the value of (hour), when the number of user connections is 400 with two DB servers, the processing time of the DB server is 5 seconds. If it is assumed that there is only one DB server, the processing time of the DB server is considered to be double, so 1
It can be estimated to be 0 seconds. Further, assuming that there are three DB servers, the processing time is considered to be ⅓ of that of one server, and therefore 10 = 1/3 can be estimated to be 3.3 seconds. The above is D
This is a performance reference value when considering the expansion / addition of B server (step 202). In this example, since the processing time is the most bottleneck, this is the reference value, but other monitoring items (CPU usage rate, memory usage rate) may be used as the reference value.

Next, it is confirmed whether or not this performance standard value is acceptable if it is different (step 203). In this confirmation, when comparing the performance difference of the DB server at the time of item 1 and the time of item 3 in FIG. 6, the number of user connections is 800, which is double that at the time of opening, and the processing time before expansion / addition is also increased. It's doubled at 10 seconds, so it's a good fit. If the threshold values do not match, the previous reference value is regarded as invalid and this value is set as a new reference value (steps 203 to 202).

Next, in this processing, paying attention to the value after the expansion / addition, the expansion / addition condition is calculated (step 204). For example, comparing item 1 and item 3 time points, the number of user connections is 4
If one DB server is used at 00:00, the processing time will be 10 seconds, so when the number of user connections reaches 800, 2
It is estimated to take 0 seconds. If we go back to the previous logic,
Assuming that there are 3 DB servers, the processing time is 20 × 1 /
3 = 6.6 seconds.

However, since the performance of the DB server added in item 4 is higher than that at the time of site opening, it is necessary to take into consideration the performance difference. Here, paying attention to the CPU performance difference, 800 MHz / 600 MHz≈ We think that there is a 1.3 times difference in performance. This is set as the performance coefficient of the augmentation / expansion server (step 205). Considering this value and considering the entire DB server group, D when the site is opened
It is estimated that the performance is improved by the following amount as compared with three B servers. (1.0 + 1.0 + 1.3) / (1.0 + 1.0 + 1.
0) = 3.3 / 3.0 = 1.1 times.

Therefore, taking the above coefficient of performance into consideration,
The estimated processing time is as follows. 20 × (1 / (3 × 1.1)) = 6.1 seconds The above calculation process of the coefficient of performance is calculated for all the servers that have been added / enhanced in the past (step 206), and then Determine if it is adaptable to the extension plan. According to this determination, the number of user connections before the expansion at the time of item 4 is 1000 from the monitoring information, and the average processing time is about 7.5 seconds. In this case, the estimated processing time for one DB server is calculated. 10 × 1000/400 = 25 seconds

Next, the performance coefficient of the DB server to be introduced this time is specified. Here, the drive frequency is 1 GHz for the CPU.
A DB server of z class will be added. 1 GHz / 600 MHz≈1.7 times, the estimated processing time after introduction is calculated based on these (step 207). 25 × 1 / (1.0 + 1.0 + 1.3 + 1.7) = 5 seconds

In this trial calculation, it is estimated that the processing time after the addition of the item 4 of the AP server group is shortened to 5 seconds, and if it is considered that the performance after the introduction is satisfied, the processing is terminated, If not satisfied, a higher-class server with a higher drive frequency is specified, or the number of additional units is increased and recalculation is performed (step 208). That is, in steps 207 to 208, the processing time when one server is added is calculated in consideration of the type of CPU and the driving frequency, and it is determined whether or not the predetermined processing time is satisfied. If not, another server is added. The maximum processing time is calculated, and this processing is repeated until a predetermined processing time is satisfied. It should be noted that although the processing time monitoring has been described in this example, the accuracy can be improved by performing similar analysis on other monitoring items (CPU usage rate, memory usage rate, etc.).

Based on the above analysis results, an expansion / expansion plan is prepared to determine how much performance a server should be installed at the time of expansion / expansion of a server on the website (step 2
09), and what is the expected value regarding the performance of the site after the introduction, it is possible to support the administrator at the time of the expansion. It should be noted that although the server expansion is described in the present embodiment, for example, when the memory usage rate increases, the memory expansion is performed, and when the CPU usage rate increases, the CPU is configured in a dual configuration. You may go.

As described above, in this embodiment, since the objective and quantitative evaluation of the current website is performed, the server usage status (CPU usage rate, memory usage, user connection count, processing time, etc.) is maintained. Monitoring is performed, thresholds are set for each information, and whether or not the values are exceeded is constantly checked, and if they are exceeded, recorded each time (this number is a factor to determine the time of server expansion / expansion). In addition, the amount of change in the information value of the server usage status before and after the expansion or addition performed so far, that is, how much performance improvement is expected, is also recorded as information. (This value will be a factor for estimating the scale at the time of expansion / expansion). By keeping this information in a database as a knowledge base and using it effectively, It can serve as a strong-expansion plan.

[0038]

As described above, according to the present invention,
The current status of the website can be objectively grasped, and the information can be used for an appropriate site expansion / expansion plan, so that an effective and lean website can be constructed and strengthened.

[Brief description of drawings]

FIG. 1 is a conceptual diagram of a system configuration that realizes a website configuration determination support method according to an embodiment of the present invention.

FIG. 2 is a system configuration diagram of an embodiment of a monitoring / analysis machine group that monitors and analyzes a web server according to the present embodiment.

FIG. 3 is a flowchart showing a procedure for analyzing the current state of the website configuration determination support method according to the present embodiment.

FIG. 4 is a diagram for explaining a procedure for supporting determination of a configuration at the time of augmentation / addition.

FIG. 5 is a diagram showing an example of an analysis table created at the time of current state analysis.

FIG. 6 is a diagram showing an example of performance difference information of websites at the time of past expansion / expansion.

[Simple explanation of symbols]

101 to 103: monitoring target server group, 33: information extracting unit, 11: monitoring target information function, 106: monitoring server P
C, 107: monitoring information storage DB, 21: monitoring information collection function, 41: monitoring information, 22: threshold crossing determination function, 4
2: Setting information, 43: Server monitoring information, 105: Analysis client PC, 44: Threshold crossing information, 31: Threshold information setting function, 45: Threshold setting information, 32: Performance difference information setting function, 46: Performance difference information .

Claims (3)

[Claims] 1. A web server group, an AP server group, and a DB.
A website configuration determination support method for supporting the configuration of each server group of a website including a server group, comprising: processing time and CPU usage rate of the web server group, AP server group, and DB server group. When the memory usage rate is detected and the processing time or the CPU usage rate or the memory usage rate exceeds a predetermined threshold value by a predetermined number or more, for the web server group, the AP server group and the DB server group that have exceeded the threshold value, A method for supporting the determination of a website configuration, which comprises calculating an additional number of servers that does not exceed the threshold value based on the number of connected users and the number of CPU clocks. 2. The website configuration determination support method according to claim 1, wherein the CPU processing speed and the memory capacity of the server added when calculating the number of additional servers are used as coefficients. 3. The processing time, the CPU usage rate and the memory usage rate of the web server group, the AP server group and the DB server group in the calculated additional number of servers are detected, and the processing time or the CPU usage rate or the memory usage rate is detected. 3. The website composition determination support method according to claim 1, wherein whether or not the rate exceeds a predetermined threshold value by a predetermined number or more is monitored, and the data is used when supporting the composition determination of the next website.