KR20170030903A - Fast Network Switching Method using the MAC switch - Google Patents

Fast Network Switching Method using the MAC switch Download PDF

Info

Publication number
KR20170030903A
KR20170030903A KR1020150128400A KR20150128400A KR20170030903A KR 20170030903 A KR20170030903 A KR 20170030903A KR 1020150128400 A KR1020150128400 A KR 1020150128400A KR 20150128400 A KR20150128400 A KR 20150128400A KR 20170030903 A KR20170030903 A KR 20170030903A
Authority
KR
South Korea
Prior art keywords
platform
low
performance
client
request
Prior art date
Application number
KR1020150128400A
Other languages
Korean (ko)
Inventor
윤정미
이상학
Original Assignee
전자부품연구원
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 전자부품연구원 filed Critical 전자부품연구원
Priority to KR1020150128400A priority Critical patent/KR20170030903A/en
Publication of KR20170030903A publication Critical patent/KR20170030903A/en

Links

Images

Classifications

    • 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
    • H04L41/08—Configuration management of networks or network elements
    • H04L41/0803—Configuration setting
    • H04L41/0823—Configuration setting characterised by the purposes of a change of settings, e.g. optimising configuration for enhancing reliability
    • H04L41/0833—Configuration setting characterised by the purposes of a change of settings, e.g. optimising configuration for enhancing reliability for reduction of network energy consumption
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/02—Protocols based on web technology, e.g. hypertext transfer protocol [HTTP]

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Power Sources (AREA)

Abstract

A fast network switching method utilizing a MAC switch is provided. A hybrid server system according to an embodiment of the present invention includes a low performance / low power system and a high performance / high power system for processing requests of clients, a low performance / low power system and a high performance / high power system are set to the same server address, A MAC address of either a low performance / low power system and high performance / high power is mapped to the server address. This enables a high-performance / high-power platform and a low-performance / low-power platform to provide web services without problems to a single web site for external clients, dramatically reducing the amount of energy consumed.

Description

{Fast Network Switching Method using the MAC switch}

The present invention relates to power reduction techniques in ICT equipment, and more particularly to an energy management platform system for power reduction in a web server.

One of the most important things when designing ICT equipment is energy efficiency. In other words, it is important to increase energy efficiency for servers, client computers, and network equipment, and to reduce CO 2 emissions, which are costly to operate these equipment and destroy the environment.

Much of the cost of running a data center goes to electricity billing, often more than the sum of hardware and software costs and labor costs.

Technology about energy consumed by servers installed in the data center started with the technology to power or shut down the integrated servers on demand in the early 2000s.

In the late 2000s, virtualization has become a key technology that can reduce the amount of power consumed by data center servers by enabling integration between services. With virtualization technology, it is possible to integrate physical servers in a form that is integrated on demand while maintaining a single image for client computers.

While cloud computing technology, which companies increasingly delegate server hosting to data centers, will become a market trend, companies will still install and operate their own computers in the enterprise for many business and technical reasons.

Also, it is more common for SMEs to build ICT equipment within the enterprise rather than in the data center.

The energy consumption of ICT equipment used in enterprises is mainly driven by client computers (such as desktop or notebook PCs), network infrastructures (access points, switches and routers) and servers.

The way to reduce energy usage by controlling individual PCs is now "a problem that has already been solved", as can be seen from many commercial products on the market today (like Verdiem's Surveyor products).

However, the way to reduce energy use has not yet been solved for servers in the enterprise.

SUMMARY OF THE INVENTION The present invention has been made in order to solve the above problems, and it is an object of the present invention to provide a high performance / high power platform (system) and a low performance / low power platform ), But they can be made to appear as a single web site to an external client.

According to an aspect of the present invention, there is provided a hybrid server system including a first system for processing a request of a client; And a second system for processing the request of the client, wherein the first system and the second system are set to the same server address, and the MAC address of either the first system or the second system is It is mapped to the server address.

If the request rate of the client is less than the threshold value, the MAC address of the first system is mapped to the server address. If the request rate of the client exceeds the threshold value, Address. ≪ / RTI >

The first system may switch to a standby mode if the request rate of the client is greater than or equal to a threshold value and the first system may switch to the standby mode if the request rate of the client exceeds a threshold value. have.

The first system and the second system may respond to the request of the client before switching to the idle mode.

In addition, the first system may have lower performance and lower power consumption than the second system.

According to another aspect of the present invention, there is provided a method of providing a service of a hybrid server system, the method comprising: mapping a MAC address of a first system and a second system, which are set to the same server address, to the server address; And a system in which the MAC address is mapped to the server address, processing the request of the client.

As described above, according to embodiments of the present invention, a high performance / high power platform and a low performance / low power platform can provide web service without problem to one single website for an external client, It is possible to drastically reduce the amount.

FIG. 1 is a graph showing the results of an experiment on a Web server,
Fig. 2 shows a log analysis result,
FIG. 3 is a graph showing the rate of occurrence of a service request below the threshold per minute,
4 is a SWEEP system configuration diagram,
Figure 5 shows the FSM,
Figure 6 shows the FSM of the Master platform,
7 is a diagram showing a prediction code.

Hereinafter, the present invention will be described in detail with reference to the drawings.

1. Hybrid Web Server System Overview

In the embodiment of the present invention, in order to reduce the amount of energy consumed by the web server, a low performance / low power platform (system) is operated together with a high performance / high power platform (system) We propose a hybrid web server system that makes a single web site appear.

Low-performance / low-power platforms are used as Web server proxy computers for high-performance / high-power platforms. A low-performance / low-power platform is capable of putting a high-performance / high-power platform in standby mode at low demand (demand) levels, thus saving energy.

Further, in the case of a high performance / high power flaton mode, it is more efficient to place the low performance / low power platform in standby mode.

To do this, it is necessary to establish a switching mechanism between two platforms and a switching method (switching timing decision policy). In order to build an effective policy, the demand for the web server should be accurately predicted.

2. Web server log properties

In most cases, utilization of servers and ICT equipment is very low. This is true for data centers as well. In order to find out when the utilization rate of the SME server is low (the request rate is also low), a result of analyzing a log of HTTP requests for one month for the main server installed in a specific institution is shown in FIG.

From the analysis results shown in FIG. 1, the request type, the interval between requests, and the request rate per minute can be known. This shows that 1) dynamic page requests account for less than 5% of the total requests, and 2) the fluctuation of the request rate is very large (very low request rate for more than half of the total time) 10 or less per minute).

2 is a graph showing a change in the daily request rate. It can be seen from Figure 2 that there is a high demand for the server during the day time, while there is less demand during the evening or weekend.

As a result of analyzing the traffic volume of the server, it can be seen that there is a time zone in which the request rate is considerably low.

3 is a table showing times with a request rate below a threshold for a particular period. As shown in FIG. 3, it can be seen that the times when less than 50 requests per minute are occupied account for more than 60% of the total time.

In an embodiment of the present invention, the high power platform maintains a standby mode for a time when the request rate is below a threshold, while at the same time the low power platform is able to handle incoming requests. This means that in most cases, the low-power platform replaces (or proxies) incoming requests on behalf of the high-power platform.

3. SWEEP system

In the embodiment of the present invention, an SME hybrid web server system considering energy efficiency is referred to as a SWEEP (SME Web Energy Efficient Platform) system. The SWEEP system according to the embodiment of the present invention seamlessly switches between the two platforms without affecting the service connection of the client.

In the SWEEP system according to the embodiment of the present invention, two different platforms provide the same web server image (S / W, web content). Hereinafter, the high-performance platform is referred to as a "master platform" and the low-power platform is referred to as an "assistant platform". The configuration of the SWEEP system is as shown in FIG.

At a time when the client connection is active, the master platform 120 operates in the ON mode, while the assistant platform 110 enters the sleep mode. In a time when the client connection is small, the master platform 120 enters the sleep mode, and the assistant platform 110 acts as a web service.

The assistant platform 110 can be implemented as a plug computer. The plug computer 110 is a low-cost platform with low processor, memory, and network performance, and low power consumption.

The master platform 120 is a high-performance, high-performance server computer with high power consumption.

The LAN switch 130 and the router 140 are network devices that connect the Assistant platform 110 and the Master platform 120 to the Internet to provide Web services to clients.

4. Switching Method (MAC Switching)

Hereinafter, the switching method will be described in detail. This is a method of switching switching between the assistant platform 110 and the master platform 120 based on the client's request rate.

When the request rate is low, the assistant platform 110 provides the service and the master platform 120 operates in the sleep state. On the contrary, when the request rate increases, the master wakes up and provides the service, and the assistant platform 110 operates in a sleep state.

Accordingly, switching between the assistant platform 110 and the master platform 120 is performed according to the request rate, and switching between the two platforms must be seamlessly performed. That is, losslessness of the request must be guaranteed.

To this end, the embodiment of the present invention introduces a MAC switching method. The MAC switching method is a method in which the assistant platform 110 and the master platform 120 share one server IP address connectable from the outside and additionally one local IP address. The local IP address is used to exchange messages between the two platforms.

On the other hand, the server IP address is mapped to the MAC address of the platform providing the service among the assistant platform 110 and the master platform 120, that is, the platform operating in the On mode. Mapping between server IP and MAC addresses is done via GARP (Gratuitous ARP). By using this method, switching can be performed without loss of request.

5 is an FSM of an assistant platform 110 for MAC switching. As shown in FIG. 5, when the Request Rate is higher than the defined threshold and the GoToSleep message is received, the assistant platform 110 switches to the WaitQEmpty state, and the client sends a request to the Queue to resend the request to the Master platform 120 Send an HTTP Redirect message to all stored requests (transition 1).

Thereafter, when the request queue becomes empty, the assistant platform 110 sends a sleep state notification message or a StartServing command to the master platform 120 and blocks the address mapping between its own MAC address and the server IP (transition 2) . The requests of all clients are then forwarded to the Master platform 120.

On the other hand, when the Request Rate is lower than the defined threshold and the Magic packet is received from the Master platform 120, the Assistant platform 110 switches to the OnNotSsrving state and transmits a status notification message to the Master platform 120 (transition 3).

Upon receiving the StartServing message from the master platform 120, the assistant platform 110 transmits a GARP packet to map its MAC address to the server IP address, and starts providing the web service to the client (transition 4) . After that, requests of all clients are forwarded to the assistant platform 110.

6 is the FSM of the Master platform 120 for MAC switching. As shown in FIG. 6, when the Request Rate is lower than the defined threshold, the Master Platform 120 switches to the WaitQEmpty state when the GoToSleep message is received. When the client platform 120 retransmits the request to the Assistant Platform 110, Send an HTTP Redirect message to all stored requests (transition 1).

Thereafter, when the request queue becomes empty, the master platform 120 transmits a sleep state notification message or a StartServing command to the assistant platform 110 and blocks the address mapping between its own MAC address and the server IP (transition 2) . After that, requests of all clients are forwarded to the assistant platform 110.

On the other hand, when the request rate is lower than the defined threshold and the Magic packet is received, the Master platform 120 switches to the OnNotSsrving state and transmits a status notification message to the assistant platform 110 (transition 3).

Then, upon receiving the StartServing message from the assistant platform 110, the master platform 120 transmits a GARP packet, maps its MAC address to the server IP address, and starts providing the web service to the client (transition 4) . The requests of all clients are then forwarded to the Master platform 120.

5. Request rate prediction

Assessor platform 110 wakes up Master platform 120 when a request rate that is higher than the threshold that the assistant platform 110 can service is predicted. This request rate threshold is the maximum value that the assistant platform 110 can process.

The request prediction is to predict whether the request rate will exceed the threshold during the next sampling interval. This is implemented through the makePrediction () function, and the actual implementation code is as shown in FIG.

FIG. 7 is a code for a prediction function using EWMA, and the right function is a technique for predicting the present sampling result and past history information with weighting.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, It will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the spirit and scope of the present invention.

110: Plug computer (Assistant platform)
120: Server computer (Master platform)
130: LAN switch
140: Router

Claims (6)

A first system for processing a request of a client; And
And a second system for processing the request of the client,
Wherein the first system and the second system are set to the same server address,
Wherein the MAC address of either the first system or the second system is mapped to the server address.
The method according to claim 1,
The MAC address of the first system is mapped to the server address if the request rate of the client is less than or equal to a threshold,
Wherein the MAC address of the second system is mapped to the server address when the request rate of the client exceeds a threshold value.
The method of claim 2,
The first system comprises:
If the request rate of the client is greater than or equal to a threshold value,
The first system comprises:
And switches to the standby mode when the request rate of the client exceeds a threshold value.
The method of claim 3,
The first system and the second system comprising:
And responds a Redirect message to the request of the client before switching to the standby mode.
The method of claim 4,
The first system comprises:
Wherein the second system has lower performance and lower power consumption than the second system.
Mapping a MAC address of one of the first system and the second system to the server address set to the same server address; And
And a system in which a MAC address is mapped to a server address, processing a request of a client.
KR1020150128400A 2015-09-10 2015-09-10 Fast Network Switching Method using the MAC switch KR20170030903A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
KR1020150128400A KR20170030903A (en) 2015-09-10 2015-09-10 Fast Network Switching Method using the MAC switch

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
KR1020150128400A KR20170030903A (en) 2015-09-10 2015-09-10 Fast Network Switching Method using the MAC switch

Publications (1)

Publication Number Publication Date
KR20170030903A true KR20170030903A (en) 2017-03-20

Family

ID=58502770

Family Applications (1)

Application Number Title Priority Date Filing Date
KR1020150128400A KR20170030903A (en) 2015-09-10 2015-09-10 Fast Network Switching Method using the MAC switch

Country Status (1)

Country Link
KR (1) KR20170030903A (en)

Similar Documents

Publication Publication Date Title
US8898493B2 (en) Architecture to enable energy savings in networked computers
US10693975B2 (en) Capacity scaling of network resources
Gunaratne et al. Managing energy consumption costs in desktop PCs and LAN switches with proxying, split TCP connections, and scaling of link speed
US7644159B2 (en) Load balancing for a server farm
WO2020199487A1 (en) Method, apparatus and device for responding to access request, and storage medium
US20090106571A1 (en) Systems and Methods to Adaptively Load Balance User Sessions to Reduce Energy Consumption
Anastasi et al. A bittorrent proxy for green internet file sharing: Design and experimental evaluation
US20100037225A1 (en) Workload routing based on greenness conditions
CN102246489A (en) Systems and methods for connection management for asynchronous messaging over http
Xu et al. A modified round-robin load-balancing algorithm for cluster-based web servers
Cheng et al. SEEDS: A solar-based energy-efficient distributed server farm
Soleimanzadeh et al. SD‐WLB: An SDN‐aided mechanism for web load balancing based on server statistics
Kuribayashi Improving quality of service and reducing power consumption with WAN accelerator in cloud computing environments
CN102137158A (en) Method for sharing information equipment resources by utilizing equipment agent system
Khan et al. Achieving energy saving through proxying applications on behalf of idle devices
Nordman et al. Greener PCs for the enterprise
Karthikeyan et al. Energy aware network scheduling for a data centre
Bolla et al. Saving energy by delegating network activity to home gateways
Park et al. Energy efficient multi-function home gateway in always-on home environment
Jegadeesan et al. Energy Consumption Power Aware Data Delivery in Wireless Network
KR20170030903A (en) Fast Network Switching Method using the MAC switch
Al‐Jarrah et al. Integrated network and hosts energy management for cloud data centers
Olson et al. Hybrid web server: Traffic analysis and prototype
KR20170030890A (en) SWEEP System for Web Server Energy Reduction
Fayyaz et al. Energy efficient resource scheduling through VM consolidation in cloud computing