KR20100113383A - Electric power management system for server and method thereof - Google Patents

Electric power management system for server and method thereof Download PDF

Info

Publication number
KR20100113383A
KR20100113383A KR1020090031935A KR20090031935A KR20100113383A KR 20100113383 A KR20100113383 A KR 20100113383A KR 1020090031935 A KR1020090031935 A KR 1020090031935A KR 20090031935 A KR20090031935 A KR 20090031935A KR 20100113383 A KR20100113383 A KR 20100113383A
Authority
KR
South Korea
Prior art keywords
server
load
switch
servers
traffic
Prior art date
Application number
KR1020090031935A
Other languages
Korean (ko)
Inventor
안태형
Original Assignee
주식회사 엔씨소프트
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 주식회사 엔씨소프트 filed Critical 주식회사 엔씨소프트
Priority to KR1020090031935A priority Critical patent/KR20100113383A/en
Publication of KR20100113383A publication Critical patent/KR20100113383A/en

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L12/00—Data switching networks
    • H04L12/02—Details
    • H04L12/12—Arrangements for remote connection or disconnection of substations or of equipment thereof
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/02—Standardisation; Integration
    • H04L41/0213—Standardised network management protocols, e.g. simple network management protocol [SNMP]
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00—Arrangements for monitoring or testing data switching networks
    • H04L43/04—Processing captured monitoring data, e.g. for logfile generation
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00—Arrangements for monitoring or testing data switching networks
    • H04L43/10—Active monitoring, e.g. heartbeat, ping or trace-route
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00—Arrangements for monitoring or testing data switching networks
    • H04L43/16—Threshold monitoring

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Health & Medical Sciences (AREA)
  • Cardiology (AREA)
  • General Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Power Sources (AREA)

Abstract

PURPOSE: An electric power management system for a server and a method thereof are provided to reduce the power consumption through the connection with a server power management system without remarkably changing the setting of plural servers and an existing L4 switch. CONSTITUTION: A traffic monitoring module(110) monitors the load states of plural servers connected to an L4 switch. If the total load is reduced below a preset critical value, a traffic control module(120) stops the load distribution for one server according to a predetermined order among the servers as the result of the change in the setting of the L4 switch. If the traffic of the server which stops the load distribution is lower than the predetermined critical value, a server power control module(130) generates a server termination signal to the servers.

Description

Server power management system and its method {ELECTRIC POWER MANAGEMENT SYSTEM FOR SERVER AND METHOD THEREOF}

The present invention relates to a system and method for managing total power consumption by shutting down or starting a specific server belonging to a plurality of server groups according to traffic.

In general, a server group consisting of a plurality of servers is used for a web service or an online game service. The main point of the server group operation is to operate a stable service considering a failure of a server or a congestion of load. For this purpose, it is common to distribute traffic and use L4 switches for fail-over in connection with server groups.

However, since the power consumption has increased greatly as the performance of the hardware has developed, it is required to derive a method for reducing the power consumption in consideration of environmental problems.

However, reducing the number of servers or building low-power systems in order to reduce power consumption can cause serious problems in service quality, so it must be cautious from the standpoint of focusing on stable service operation.

Therefore, it is urgent to develop a methodology that can reduce power consumption without affecting the quality of service.

The present invention has been made in order to solve the above problems, it is not necessary to start the entire server at the same time when the maximum load time point, the maximum load time is not reached at a time, but gradually arrives with a time difference It is an object of the present invention to provide a server power management system and method for managing power of a server sequentially according to load.

Server power management system of the present invention for achieving the above object is a traffic monitoring module for monitoring the load of a plurality of servers connected to the L4 switch;

A traffic control module that prevents load distributing to any one server in a predetermined order among running servers by changing a setting of an L4 switch when the total load applied to the plurality of servers decreases below a preset threshold; And

And a server power control module configured to generate a server termination signal for the corresponding server when the traffic of the server whose load is blocked by the traffic control module drops below a preset threshold.

On the other hand, the server power management method of the present invention for achieving the above object comprises the step 110 of monitoring the load of a plurality of servers connected to the L4 switch;

Changing a setting of an L4 switch to prevent load distribution among any one of the servers in operation according to a predetermined order when the total load applied to the plurality of servers decreases below a preset threshold; And

And generating a server termination signal for the server when the traffic of the server whose load is blocked is dropped below a preset threshold.

On the other hand, the server power management method of the present invention for achieving the above object, step 210 of monitoring the load of a plurality of servers connected to the L4 switch; And

And a step 220 of generating a WOL signal for any one server according to a predetermined order when the total load applied to the plurality of servers increases above a preset threshold.

According to the present invention as described above it is possible to easily lower the overall power consumption simply by connecting the server power management system according to the present invention without significantly changing the configuration consisting of a conventional L4 switch and a plurality of servers.

On the other hand, by using SNMP, it is possible not only to monitor and control the traffic state of the server group even at a remote location, but also to ensure the quality of service by sequentially starting the servers even when the peak load point is reached.

Hereinafter, a configuration of a server power management system according to the present invention will be described in detail with reference to the accompanying drawings. FIG. 1 is a network diagram showing a connection relationship between an L4 switch, a plurality of servers, and a server power management system, and FIG. 2 is a functional block diagram showing a configuration of a server power management system.

First, as shown in FIG. 1, the server power management system 100 according to the present invention is connected to an L4 switch 2 and a plurality of servers 1 connected to the L4 switch. The L4 switch provides a load balancing function for distributing loads applied to a plurality of servers and a fail over function for ensuring integrity, and the plurality of servers 1 are clients connected via the Internet. Provide a variety of services.

On the other hand, the server power management system 100 is connected to the L4 switch 2 and the plurality of servers 1 to manage the power of the server as described below.

As shown in FIG. 2, the server power management system 100 includes a traffic monitoring module 110, a traffic control module 120, and a server power control module 130.

The traffic monitoring module 110 monitors the load of the L4 switch 2 and the plurality of servers 1 connected to the L4 switch 2. To this end, the traffic monitoring module 110 communicates with the L4 switch 2 and the plurality of servers 1 using the Simple Network Management Protocol (SNMP). SNMP is an application layer standard protocol defined on TCP / IP for the purpose of managing and monitoring network equipment. As a result, the traffic monitoring module 110 can remotely control the L4 switch (2) even if the expansion or other setting of the server is changed. And it is possible to easily monitor the load applied to the plurality of servers (1).

To this end, the traffic monitoring module 110 sends a query to each of the SNMP agents of the L4 switch 2 and the plurality of servers 1 using SNMP, and loads the load from the L4 switch 2 and each server 1. You can monitor the load on the server by receiving a reply.

Meanwhile, the traffic control module 120 changes the setting of the L4 switch 2 when the total load applied to the plurality of servers 1 decreases below a preset threshold as a result of monitoring by the traffic monitoring module 120. The load distribution is prevented with respect to any one server 1 according to a predetermined order among the servers in operation.

That is, when the total amount of the load applied to the entire server 1 that is currently running is lowered below a certain level, the L4 switch 2 causes the load on any one server 1 by changing the setting of the L4 switch 2. To prevent it from dispersing anymore.

In this case, the threshold value may be set or changed by an administrator in consideration of the server's processable capacity. Conceptually, the threshold is a figure for the total load and the total load to adjust the number of servers to be started, such that only 5 servers start up when the total load falls below 100, and 4 servers start up when the load falls below 80. to be.

The server power control module 130 generates a server termination signal for the server when the load of the server 1 whose load is blocked by the L4 switch 2 falls below a predetermined threshold as described above.

At this time, the server end signal is a signal that causes the server (1) receiving it to shut down the system and power down. Accordingly, when the server 1 receives the server termination signal from the server power control module 130, the server 1 terminates the system so that it does not consume any more power. However, as will be described later, a minimum power should be supplied to the network adapter for the Wake On Lan (WOL) function. Since the amount of power consumed is small enough to be negligible compared to the amount of power consumed during server startup, power consumption is greatly reduced by the termination of the server 1.

On the other hand, if the total load applied to the plurality of servers 1 currently running as described above decreases below the preset threshold, the traffic control module 120 loads the load on any one server 1 in a predetermined order. The load is further reduced even after the server 1, which is prevented from being distributed, is shut down under the control of the server power control module 130, and if the decrease is less than the next threshold, the L4 switch is changed. Prevent load distribution for the server 1 in the next order.

In other words, the server 1 is terminated one by one in order according to the decrease in load.

On the other hand, contrary to the above description, when the total load applied to the plurality of servers increases above the preset threshold, the previously terminated server 1 is started in a predetermined order. To this end, if the total load applied to the plurality of servers 1 increases as a result of the monitoring of the traffic monitoring module 110, the server power control module 130 for any one server in a predetermined order. Generate Wake On Lan (WOL) signal.

This WOL signal is a Magic Packet sent to the Mac Address of the server 1 to be started, and the server 1 which receives it starts booting automatically.

Magic packets are generally broadcast frames sent on any of ports 0, 7, and 9. They are sent using protocols such as User Datagram Protocol (UDP). On the other hand, such a magic packet may be, for example, the form of the hexadecimal FF FF FF FF FF FF followed by the MAC address of the server 1 to be started.

According to the control of the server power control module 130, when the server 1 is activated, the traffic monitoring module 110 confirms the activation of the server 1 to which the WOL signal is applied using SNMP. When booting is complete and the server 1 is in a state where service can be provided, the traffic control module 120 changes the setting of the L4 switch 2 to initiate load distribution to the started server 1. do.

That is, the newly started server 1 is to handle a part of the increased load amount.

Hereinafter, a process of managing server power when the total load is reduced in the power management system having the above configuration will be described with reference to the accompanying drawings. However, descriptions of matters overlapping with those described above will be omitted.

3 is a reference diagram illustrating a process of terminating a server in a predetermined order when the network total load is reduced, and FIG. 4 is a flowchart illustrating the process illustrated in FIG. 3 in time series.

First, the traffic monitoring module 110 monitors the load of a plurality of servers 1 connected to the L4 switch 2 using SNMP (S110). At this time, the traffic monitoring module 110 sends a query to the SNMP agent of each server 1, and receives a reply from the SNMP agent of each server to monitor the load applied to the server.

At this time, if the total load on the entire server 1 that is running decreases below a preset threshold, the traffic control module 120 may be configured to the server 1 according to a predetermined order among the servers 1 that are running. In order to prevent load distribution, the setting of the L4 switch is changed (S120).

Subsequently, when the traffic of the server 1 in which load distribution is blocked in step S120 falls below a preset threshold, the server power control module 130 generates a server termination signal for the server 1 (S130). At this time, the traffic of the server 1, that is, the threshold for the load is different from the threshold for the total load determined on the entire server 1 in operation S120, and the traffic of the server 1 in which load distribution is blocked is prevented. If the monitoring module 110 monitors and falls below the threshold, the server is shut down. Of course, by changing the setting of the L4 switch, traffic that has been applied to the server 2 to which the server termination signal has been transmitted should be handled by other servers 1 in operation.

On the other hand, the server 1 receiving the server end signal terminates the system and powers down (S140). Power off means that no more power is consumed by the shutdown of the system. Of course, minimal power is consumed for the WOL feature, but this is negligible.

Meanwhile, with reference to the accompanying drawings, a process of managing server power when the total load increases in the power management system having the above configuration will be described. However, descriptions of matters overlapping with those described above will be omitted.

5 is a reference diagram illustrating a process of starting a server in a predetermined order when the network total load increases, and FIG. 6 is a flowchart illustrating the process illustrated in FIG. 5 in time series.

First, the traffic monitoring module 110 monitors the load of the plurality of servers 1 connected to the L4 switch 2 (S210).

When the total load applied to the plurality of servers 1 that are running increases above a preset threshold, the server power control module 130 may be configured to perform one of the servers 1 in accordance with a predetermined order. Generate a WOL signal (S220). At this time, the generation of the WOL signal means that the server power supply control module 130 sends a magic packet to a Mac address of a server to be activated.

When the server 1 receiving the WOL signal starts booting by automatically booting the system, the traffic monitoring module 110 confirms the starting of the server to which the WOL signal is applied using SNMP (S230).

When the activation is confirmed, the traffic control module 120 sets the L4 switch to initiate load distribution to the activated server 1 (S240).

Although the present invention has been described with reference to several embodiments, the present invention is not limited to these embodiments and can be freely modified and practiced within the scope of the technical idea described in the claims.

1 is a network diagram illustrating a connection relationship between an L4 switch, a plurality of servers, and a server power management system according to the present invention;

2 is a functional block diagram showing the configuration of a server power management system according to the present invention,

3 is a reference diagram illustrating a process of terminating servers in order when network total load is reduced in the network configuration of FIG. 1;

4 is a flowchart illustrating a process of shutting down servers in order when network total load is reduced.

FIG. 5 is a reference diagram illustrating a process of starting servers in order when network total load increases in the network configuration diagram of FIG. 1.

6 is a flowchart for explaining a process of starting servers in order when network total load increases.

<Explanation of symbols for the main parts of the drawings>

110: traffic monitoring module 120: traffic control module

130: server power control module

Claims (16)

A power management system for multiple servers connected to an L4 switch, A traffic monitoring module for monitoring load of a plurality of servers connected to the L4 switch; A traffic control module that prevents load distributing to any one server in a predetermined order among running servers by changing a setting of an L4 switch when the total load applied to the plurality of servers decreases below a preset threshold; And And a server power control module configured to generate a server termination signal for the server when the traffic of the server whose load is blocked by the traffic control module drops below a preset threshold. The method of claim 1, And the server power control module generates a WOL signal for any one server in a predetermined order when the total load applied to the plurality of servers increases above a predetermined threshold. The method of claim 2, When the traffic monitoring module confirms the start of the server to which the WOL signal is applied using SNMP, the traffic control module initiates load distribution to the started server by changing the setting of the L4 switch. Server Power Management System. The method of claim 1, The traffic monitoring module sends a query to the SNMP agent of each server using SNMP, and monitors the load of the server by receiving a response to the load from each server. The method of claim 1, The traffic monitoring module is a server power management system, characterized in that for monitoring the load of the L4 switch using SNMP. The method of claim 1 The server end signal is a server power management system, characterized in that for the server receiving the signal to shut down and power down the system. The method of claim 2 The WOL signal is a server power management system, characterized in that the Magic Packet (Magic Packet) sent to the Mac Address (Mac Address) of the server to be activated. In the power management method for a plurality of servers connected to the L4 switch, Step 110 for monitoring load of a plurality of servers connected to the L4 switch; Changing a setting of an L4 switch to prevent load distribution among any one of the servers in operation according to a predetermined order when the total load applied to the plurality of servers decreases below a preset threshold; And And generating a server termination signal for the server when the traffic of the server whose load is blocked is dropped below a preset threshold. The method of claim 8, Step 110 is a step 111 of sending a query to the SNMP agent of each server; And And receiving a reply from an SNMP agent of each server and measuring load applied to the corresponding server (step 112). The method of claim 8, And after step 130, the server receiving the server end signal further comprises a step 140 in which the server terminates the system and turns off the power. In the power management method for a plurality of servers connected to the L4 switch, Monitoring a load of a plurality of servers connected to the L4 switch (210); And And generating a WOL signal for any one server according to a predetermined order when the total load applied to the plurality of servers increases above a predetermined threshold value. The method of claim 11, Step 210 may include sending a query to an SNMP agent of each server; And And receiving a reply from an SNMP agent of each server to measure a load applied to the corresponding server. 212. The method of claim 11, The step 220 is performed by sending a Magic Packet to a Mac Address of a server to be activated. The method of claim 11, And after step 220, step 221 of starting the system by the server receiving the WOL signal. The method of claim 11, And a step 230 of confirming the start of the server to which the WOL signal is applied by using SNMP after step 220. The method of claim 15, And after step 230, step 240 of changing a setting of the L4 switch to initiate load distribution to the started server.
KR1020090031935A 2009-04-13 2009-04-13 Electric power management system for server and method thereof KR20100113383A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
KR1020090031935A KR20100113383A (en) 2009-04-13 2009-04-13 Electric power management system for server and method thereof

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
KR1020090031935A KR20100113383A (en) 2009-04-13 2009-04-13 Electric power management system for server and method thereof

Publications (1)

Publication Number Publication Date
KR20100113383A true KR20100113383A (en) 2010-10-21

Family

ID=43132980

Family Applications (1)

Application Number Title Priority Date Filing Date
KR1020090031935A KR20100113383A (en) 2009-04-13 2009-04-13 Electric power management system for server and method thereof

Country Status (1)

Country Link
KR (1) KR20100113383A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101346188B1 (en) * 2011-12-27 2013-12-31 삼성중공업 주식회사 Method for processing data distributed and systems performing the same
CN104009862A (en) * 2013-02-27 2014-08-27 腾讯科技（深圳）有限公司 Equipment scheduling method and system
WO2017019035A1 (en) * 2015-07-28 2017-02-02 Hewlett Packard Enterprise Development Lp Monitoring network utilization
KR20180008236A (en) * 2016-07-14 2018-01-24 주식회사 팀스톤 Computer-executable method of monitoring server load, apparatus performing the same and medium storage storing the same

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101346188B1 (en) * 2011-12-27 2013-12-31 삼성중공업 주식회사 Method for processing data distributed and systems performing the same
CN104009862A (en) * 2013-02-27 2014-08-27 腾讯科技（深圳）有限公司 Equipment scheduling method and system
WO2017019035A1 (en) * 2015-07-28 2017-02-02 Hewlett Packard Enterprise Development Lp Monitoring network utilization
KR20180008236A (en) * 2016-07-14 2018-01-24 주식회사 팀스톤 Computer-executable method of monitoring server load, apparatus performing the same and medium storage storing the same

Similar Documents

Publication Publication Date Title
US10866624B2 (en) Power management method of a system made of devices powered over data cable
US8108705B2 (en) Power supplying device, power supply controlling method, power supply controlling program and network system
US7779282B2 (en) Maintaining network connectivity while operating in low power mode
US9173116B2 (en) Method, apparatus and system for controlling access points
US20090172163A1 (en) Systems and methods of on-demand waking of computers
KR100932908B1 (en) Power Saving Method of Optical Termination Device and Optical Line Terminal in Optical Access Network
US8516127B2 (en) Method and system for maintaining connections between a terminal and servers in a communication system
KR100812374B1 (en) System and method for managing protocol network failures in a cluster system
US9753514B2 (en) Apparatus and method for reducing power consumption in customer premises equipment
US20060155828A1 (en) Router setting method and router device
CN101656764A (en) Method, system and device for keeping session of DHCP user active
KR20100113383A (en) Electric power management system for server and method thereof
US20140280467A1 (en) Method of managing port dhcp server protocol addresses
US8214680B1 (en) PoE powered management circuitry using out-of-band data port
US8219691B1 (en) Methods and apparatus for remotely waking up a computer system on a computer network
US20150169033A1 (en) Systems and methods for power management in stackable switch
WO2016145858A1 (en) Power over ethernet control method and device
CN101453395A (en) Switching method, system and device for uplink mode
WO2021073367A1 (en) Data processing method, device, and system
CN107682231A (en) Monitor method, apparatus, computer equipment and the storage medium of PD equipment
JP7443800B2 (en) Power management device, information processing system, and power management device control method
CN115174415B (en) Information release control method, device and equipment for virtual tunnel endpoint
JP5793701B2 (en) Monitoring system, monitoring information transmitting apparatus and program, monitoring information receiving apparatus and program
JP2012216011A (en) Resource distribution system
CN113098721A (en) Multi-bandwidth backup and superposition strategy and method for router

Legal Events

Date Code Title Description
WITN Withdrawal due to no request for examination