JP2004110667A - Cache server control system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To enable stabilization of traffic in a network and effective use of a band in a network system having a plurality of cache servers connected through a node. <P>SOLUTION: This system has a traffic monitoring device to monitor traffic information in the node and give an instruction of communication control to a cache server. In the cache server, the communication in the cache server is controlled so as to reduce the traffic in the node when the load factor of traffic in the node is high and so as to increase the traffic in the node when the load factor of traffic in the node is low. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a cache server control system, and is suitable for use in a large-scale network in which a plurality of cache servers are connected by nodes, and measures traffic at each node to control communication settings of the cache server. The present invention relates to a cache server control system that stabilizes traffic in a network and enables effective use of a communication band.
[0002]
[Prior art]
With the spread of the Internet technology, access to contents on a Web server has become enormous. A cache server is installed on the network to reduce the network bandwidth and load on the Web server and to reduce the response time to the Web user. The temporarily stored Web contents are stored as temporary data in the cache server, and a plurality of Web users The method of sharing in is taken.
[0003]
As a technique using such a cache server, Patent Literature 1 discloses an example in which a cache server is applied to homepage data. Patent Literature 2 discloses a technique in which a plurality of cache servers communicate with each other to adjust cache data to be held. Examples are disclosed.
[0004]
By the way, the cache server is a server provided to store communication data as cache data and to speed up access to Web pages and the like, but has various functions for controlling communication.
[0005]
As a function for adjusting the contents in relation to the contents of the Web page, an HTML (HyperText Makeup Language) requested by a Web user is analyzed, a prefetch function for prefetching, and a URL (Uniform Resource Locators) registered in advance are downloaded. There is a schedule download function that always keeps the contents in a cache, and a contents distribution function that distributes Web contents to a server registered in advance.
[0006]
In addition, functions for adjusting whether or not to pass a designated packet include a bandwidth control for limiting a bandwidth to be allocated according to a type of a protocol passing through a cache server, a URL requested for a Web request from a Web user, a requested URL, and the like. There are a filtering function for performing access control according to the server IP address, and an access control function for restricting Web access to users who can access the cache server from the IP address.
[0007]
The function of adjusting the cache data of the cache server includes a function of adjusting the time for keeping the cache data as valid data in the cache, and a function of refreshing the cached data to maintain the freshness of the cached Web contents. is there.
[0008]
Further, for a TCP / IP session, there is a function of adjusting a TCP / IP session holding time and a function of adjusting a TCP / IP data reading timeout time.
[0009]
The operating conditions and settings of these functions are independent of the load status of the nodes on the network. The node on the network here is a gateway device of a gate network segment such as a router.
[0010]
Therefore, the setting of the function of the cache server as described above is not appropriate, and the load factor of the network node may be increased, or conversely, the bandwidth of the network may not be used effectively.
[0011]
[Patent Document 1]
JP 2000-315171 A [Patent Document 2]
JP 2000-155712 A
[Problems to be solved by the invention]
As described above, conventionally, the operating conditions such as the prefetch function, the schedule download function, the content distribution function, and the cache data refresh function of the cache server are independent of the load factor of the node. If it is not appropriate, the load factor of the nodes on the network will be increased, leading to a response delay to the user browsing the Web, or a delay in data flowing through other networks, or conversely, There is a problem that the band is not used effectively.
[0013]
Further, setting of bandwidth control, setting of TCP / IP session holding time, setting of TCP / IP data reading timeout time, setting of filter condition of filtering function, setting of access filter condition of access control function, setting of cache data holding time Are statically set independently of the load factor of the node, so those settings may not be appropriate for the load factor of the node. In the bandwidth control, for example, HyperText Transfer Protocol (HTTP), File Transfer Protocol (FTP), Real Time Streaming Protocol (RTSP), and a protocol that can use a Real-time Transport Protocol (RTP) that can use a Real-time Transport Protocol (RTP) can be used. If they are allocated in a practical manner, there is a waste or shortage for each allocated band in an actual use situation. In setting the TCP / IP session holding time and the data reading timeout time, if the bandwidth is busy, it takes a long time until the end of data transfer, so it is necessary to take a longer time. It becomes useless. In the filtering function, in the related art, the filter condition is changed for each time period. However, the filter condition is changed in cooperation with the load factor of the node to limit the number of Web sites that can be referred to, thereby limiting the Web request. The number cannot be adjusted. In the access control function, the prior art has a function of assigning a transfer priority by an IP address using QoS (Quality of Service), but the access control access by the IP address is performed in cooperation with the load factor of the node. It is not possible to change the filter conditions and adjust the number of Web requests.
[0014]
The present invention has been made to solve the above problems, and an object of the present invention is to monitor the load status of a node on a network in a network system in which a plurality of cache servers are connected by the node, and A cache server control system that stabilizes traffic in the network and makes effective use of bandwidth by dynamically controlling the operation of cache server functions and changing settings in cooperation with the load status of nodes. Is to provide.
[0015]
[Means for Solving the Problems]
According to the present invention, a traffic monitoring device is provided, monitors traffic in a node, and transmits a control command to a cache server to be controlled according to a traffic load factor in the node.
[0016]
The cache server receives the control command from the traffic monitoring device, and when the traffic load factor at the node is high, the traffic load at the node is low in the direction of decreasing the traffic at the node. In order to increase the traffic at the node, the communication is controlled. The control of the communication includes the setting change and execution of the prefetch function, the setting change and execution of the schedule download function, the setting change and execution of the content distribution function, the bandwidth Change control setting, change cache data retention time, change cache data refresh function setting, change filtering function setting, change access control function setting by IP address of packet issuer, set TCP / IP session retention time There is change, change of TCP / IP of the data read-out time.
[0017]
For the same purpose, a traffic monitoring unit may be provided in the cache server.
[0018]
The setting information of the traffic monitoring device is stored in a local storage device or a storage device of a device connected to a network such as a file server. Then, the setting information of the other traffic monitoring device can be loaded, and when the other traffic monitoring device fails and goes down, the setting information of the failed monitoring device is taken over and control of the cache server is performed. To be able to replace the directive.
[0019]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to FIGS. 1 and 2. FIG.
[0020]
[Embodiment 1]
Hereinafter, a first embodiment according to the present invention will be described with reference to FIG.
FIG. 1 is a block diagram showing the configuration of the cache server control system according to the first embodiment of the present invention.
[0021]
In the cache server control system shown in FIG. 1, a traffic monitoring device A 110 and cache servers 130 and 140 to be controlled are respectively connected to nodes A, B and C by a network such as a LAN.
[0022]
The traffic monitoring device A is connected in a LAN (Local Area Network) or the like in FIG. 1, but may be provided in the node A.
[0023]
The traffic monitoring device A includes a traffic monitoring unit 111, a trap signal transmitting unit 112, and setting information 121.
[0024]
The traffic monitoring unit monitors traffic at a node to be monitored on the network. Traffic at a node is represented by the amount of packets passing through that node.
[0025]
The trap signal transmitting unit 112 transmits a trap signal to the cache server. The trap signal mentioned here is a communication control command transmitted to the cache server to be controlled.
[0026]
The setting information 121 is setting information of the traffic monitoring server such as node information to be monitored, cache server information to be controlled, and trap signal information.
[0027]
This setting information may be stored in a local storage device of the traffic monitoring device A110, or may be stored in a storage device such as a file server or a NAS (Network Attached Storage) connected to a network. .
[0028]
The cache server A 130 to be controlled includes a control unit 131, a trap signal receiving unit 132, and setting information 133.
[0029]
The control unit 131 performs various communication controls with reference to the setting information 133. The trap signal receiving unit 132 receives the trap signal from the traffic monitoring device A 130 and transmits it to the control unit 131. The setting information 133 is information for performing communication control of the cache server.
[0030]
Here, consider a case where the traffic monitoring device A110 controls the cache server A130 to be controlled in cooperation with the load factor of the node A150 to be monitored.
[0031]
The traffic monitoring unit 111 acquires the traffic information of the node A150 and calculates the load factor of the node A150. The load factor in a node is the ratio of the amount of packets actually passing through the gateway device to the maximum value of the packet processing performance of the node.
[0032]
Then, the trap signal transmitting unit 112 transmits a trap signal to the trap signal receiving unit 131 of the cache server A 130 according to the setting content of the setting information 121 according to the state of the load factor of the node.
[0033]
In response to the trap signal, the communication control in the cache server A130 reduces the traffic of the node A when the load factor of the node A is high, that is, the communication in the node A becomes a bottleneck in the network system. When the load factor of the node A is low, the control is performed so as to increase the traffic of the node A, that is, to effectively use the communication band of the node A.
[0034]
Specifically, the function of the cache server is controlled as described below.
(1) Setting change and execution of the prefetch function The prefetch function is a function of prefetching a URL (link destination) described in an HTML file acquired by a Web user request even without a Web user request.
When the load ratio of the monitored node is high:
Turn off the prefetch function.
When the load ratio of the monitored node is low:
Turn on the prefetch function.
(2) The setting change of the schedule download function and the execution schedule download function are functions of accessing a specific URL at a preset time and caching Web contents.
When the load ratio of the monitored node is high:
Either completely turn off the scheduled download function or temporarily stop the download and retry after a certain time.
When the load ratio of the monitored node is low:
Turn on the schedule download function and complete the download.
(3) Change and execution of setting of content distribution function This function distributes specific Web contents to a specific cache server at a preset time.
When the load ratio of the monitored node is high:
Either turn off the content distribution function and do not distribute the content, or reduce the traffic volume due to the distribution of the content.
When the load ratio of the monitored node is low:
Turn on the content distribution function and increase the traffic volume due to content distribution as needed.
(4) Change of bandwidth control setting The bandwidth referred to here is the allocation of the amount of transmission on the network. As the bandwidth control, priority control for controlling the priority by a protocol (HTTP, FTP, RTSP, RTP, etc.) There is band control for reducing the band of the network and controlling the band.
When the load ratio of the monitored node is high:
Priority control: The bandwidth of RTSP or RTP that uses a lot of bandwidth is reduced, or the bandwidth of a protocol with a low priority in business is reduced.
-Own bandwidth control: When the load rate of the monitored node that controls to reduce the bandwidth of the own network and reduce the bandwidth is low:
Priority control: The bandwidth of RTSP and RTP that use a lot of bandwidth, and the bandwidth of a protocol with a low priority in business are expanded.
Own bandwidth control: expands its own network bandwidth and controls so as to make effective use of the bandwidth (5) setting change of filtering function The filtering function is to define destination information such as URL, IP address and subnet mask. , A source ID such as a user ID, an IP address, and a subnet mask. Based on a definition that combines the destination information definition and the source information definition (hereinafter, referred to as a “filtering definition”), a Web request from a Web user is When the header information matches the definition, this function enables or disables the Web request.
When the load ratio of the monitored node is high:
Strict filtering definitions to reduce the amount of packets passed.
When the load ratio of the monitored node is low:
Relax the filtering definition to increase the amount of packets passed.
(6) Changing the setting of the access control function using the IP address of the packet issuer The access control function using the IP address of the packet issuer refers to the Web user based on the source information definition such as the IP address of the packet issuer and the subnet mask. This function enables or disables the Web request when the header information of the Web request from the Web server matches the definition.
When the load ratio of the monitored node is high:
Strict access control definitions to reduce the amount of packets passed.
When the load ratio of the monitored node is low:
Loosen the access control definition to increase the amount of packets passed.
(7) Change of Cache Data Retention Time The cache data retention time is a time for storing cache data such as Web contents as valid data.
When the load ratio of the monitored node is high:
When the load ratio of the monitored node, which prioritizes faster response time even if the data is kept up-to-date by increasing the cache data retention time, is low:
(8) Refresh function of cache data The refresh function of cache data is to maintain the freshness of cache data after a certain period of time. This function checks the latestness of the header information by accessing the original server and acquiring and comparing the header information. If it is the latest, it continues to cache, and if it is old, the cache data is discarded and updated to the latest data .
When the load ratio of the monitored node is high:
Turn off the cache data refresh function.
When the load ratio of the monitored node is low:
Turn on the cache data refresh function.
(9) Change of TCP / IP session holding time setting A TCP / IP session is a term that means from the start to the end of a connection in the TCP / IP protocol. Is a time for keeping a TCP / IP session between the Web client PC and the cache server, and between the cache server and the Web server.
When the load ratio of the monitored node is high:
Since session retention consumes the resources of each network device, when the load factor of the monitored node that controls to shorten the session retention time after data transmission / reception ends is low:
The session retention consumes the resources of each network device, but if the session is maintained, the next data transmission / reception can be performed quickly, so the session retention time is maintained for a while after the data transmission / reception ends. (10) Change of TCP / IP data read timeout setting TCP / IP data read timeout is allowed to perform one data transmission / reception in the TCP / IP protocol. Timeout time.
When the load ratio of the monitored node is high:
If the network traffic is high, it takes time to read the data, so if the load ratio of the monitored node that controls to increase the timeout value is low:
As described above, the case where the traffic monitoring device A110 gives a control instruction to the cache server A130 with the node A as a monitoring target has been described as long as the control is performed so as to shorten the timeout value, but the monitoring target node is configured to monitor the nodes A150, B160 and C170. Even if there are a plurality of such servers, the traffic can be monitored to control the communication with the cache server.
[0035]
Further, even when there are a plurality of cache servers to be controlled, the traffic monitoring device A110 can control all the cache servers to be controlled.
[0036]
Further, in this example, one traffic monitoring device has been described, but a plurality of traffic monitoring devices may be provided. Then, the operating states are mutually monitored. Then, when a failure occurs in a certain traffic monitoring device, another traffic monitoring device may take over the setting information 121 so that it can be replaced.
[0037]
For example, the monitoring target of the traffic monitoring device A110 is the node A, the control target is the cache server A130, the monitoring target of the traffic monitoring device m (not shown) is the node n, and the control target is the cache server n140. Sometimes, when a failure occurs in the traffic monitoring device m and the traffic monitoring device m goes down, the traffic monitoring device A110 sets the nodes to be monitored to the nodes A and n, and the control targets to the cache server A130 and the cache server n140.
[0038]
In this way, the reliability of the system can be improved.
[0039]
[Embodiment 2]
Hereinafter, a second embodiment according to the present invention will be described with reference to FIG.
FIG. 2 is a block diagram showing a configuration of the cache server control system according to the second embodiment of the present invention.
[0040]
In the first embodiment, the means for monitoring traffic and the means for giving a command to the cache server are provided separately from the cache server. In the present embodiment, however, they are provided in the cache server. .
[0041]
The cache server A230 according to the present embodiment includes a control unit 231, a trap signal receiving unit 232, a trap signal transmitting unit 233, a traffic monitoring unit 234, and setting information 235, and is connected to a network by a node n. .
[0042]
The control unit 231, the trap signal receiving unit 232, the trap signal transmitting unit 233, and the traffic monitoring unit 234 have the same functions as those in the first embodiment, and the setting information 235 includes node information to be monitored and a cache server to be controlled. Information such as information and trap signal information.
[0043]
In this example, for example, consider a case where the cache server A230 is controlled in cooperation with the load factor of the node A250.
[0044]
In this case, the traffic monitoring unit 234 acquires the traffic information of the monitoring target node A250 and calculates the load factor of the node A250.
[0045]
Then, the trap signal transmitting unit 233 transmits a trap signal to the trap signal receiving unit 232 of the cache server A 130 according to the setting content of the setting information 235 according to the state of the load factor of the node.
[0046]
In response to the trap signal, the communication control in the cache server A130 prevents the traffic of the node A from increasing when the load factor of the node A is high, that is, the communication in the node A becomes a bottleneck in the network system. When the load factor of the node A is low, the control is performed so that the traffic of the node A is not reduced, that is, the communication band of the node A is effectively used.
[0047]
The specific contents of the communication control are the same as in the first embodiment.
[0048]
Further, similarly to the first embodiment, a plurality of nodes may be monitored.
[0049]
In this example, the traffic monitoring unit 234 and the trap signal transmitting unit 233 are provided in the cache server A 130 to be controlled, but another cache server n (not shown) may be controlled. In that case, the trap signal transmitting unit 233 in the cache server A 130 transmits the trap signal to the trap signal receiving unit 232 in the cache server n. Of course, similarly to the first embodiment, a plurality of cache servers may be controlled.
[0050]
A mechanism for monitoring traffic and a mechanism for transmitting a trap signal may be provided in a plurality of cache servers.
[0051]
At this time, it is also possible to improve the reliability of the system by monitoring the operation states of each other and, when a failure occurs in a certain traffic monitoring mechanism, by taking over the setting information 235. Same as 1.
[0052]
【The invention's effect】
According to the present invention, in a network system in which a plurality of cache servers are connected by a node, the load status of the node on the network is monitored, and the operation of the function of the cache server is coordinated with the load status of the node on the network. A cache server control system capable of stabilizing traffic in a network and effectively utilizing a band can be provided by dynamically performing control and changing setting contents.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a configuration of a cache server control system according to a first embodiment of the present invention.
FIG. 2 is a block diagram illustrating a configuration of a cache server control system according to a second embodiment of the present invention.
[Explanation of symbols]
110 ... Traffic monitoring device A
111 traffic monitoring unit 112 trap signal transmitting unit 121 setting information 130 cache server A
131 control unit 132 trap signal receiving unit 133 setting information 140 cache server B
141 control unit 142 trap signal receiving unit 143 setting information 150 node A
160 ... Node B
170 ... node n
230: Cache server A
231, control unit 232, trap signal receiving unit 233, trap signal transmitting unit 234, traffic monitoring unit 235, setting information 250, node A
260 node n

Claims (8)

In a cache server control system in which one or more cache servers are connected to a network via a node,
Equipped with one or more traffic monitoring devices,
This traffic monitoring device
Setting information of the cache server to be controlled;
Means for acquiring and analyzing traffic information at the node to be monitored;
Means for transmitting a control command to the cache server according to a traffic load factor in the node,
The cache server receives a control command from the traffic monitoring device,
When the load rate of the traffic at the node is high, the cache server may reduce the traffic at the node, and when the load rate of the traffic at the node is low, increase the traffic at the node. A cache server control system for controlling the communication of the cache server. As a type of control of communication of the cache server,
Change and execute the prefetch function settings,
Change and execute the schedule download function,
Change and execute the content distribution function settings,
Change bandwidth control settings,
Change the settings of the filtering function,
Change the setting of the access control function by the IP address of the packet issuer,
Change cache data retention time,
Change the setting of the cache data refresh function,
Change TCP / IP session retention time settings,
2. The cache server control system according to claim 1, wherein a change of a TCP / IP data read timeout time is included. In a cache server control system in which one or more cache servers are connected to a network via a node,
The cache server,
Equipped with a traffic monitoring unit,
This traffic monitor is
Means for acquiring and analyzing traffic information;
According to the load rate of the traffic in the node, itself, or, to another cache server, having means for transmitting a control command to the cache server,
When the load rate of the traffic at the node is high, the cache server may reduce the traffic at the node, and when the load rate of the traffic at the node is low, increase the traffic at the node. A cache server control system for controlling the communication of the cache server. As a type of control of communication of the cache server,
Change and execute the prefetch function settings,
Change and execute the schedule download function,
Change and execute the content distribution function settings,
Change bandwidth control settings,
Change the settings of the filtering function,
Change the setting of the access control function by the IP address of the packet issuer,
Change cache data retention time,
Change the setting of the cache data refresh function,
Change TCP / IP session retention time settings,
3. The cache server control system according to claim 2, wherein a change of a TCP / IP data read timeout time is included. The setting information of the traffic monitoring device connected to the network is stored in a local storage device of the traffic monitoring device, or a storage device of a device connected to the network,
2. The cache server control system according to claim 1, wherein the traffic monitoring device can load and use the stored setting information of itself or another traffic monitoring device. The traffic monitoring device has means for monitoring an operation state of another traffic monitoring device,
5. The cache server control system according to claim 4, wherein when a failure occurs in another traffic monitoring device, the setting information of the failed traffic monitoring device is loaded to control communication of the cache server. . The setting information of the traffic monitoring unit in the cache server connected to the network is stored in a local storage device of the cache server or a storage device of a device connected to the network,
4. The cache server control system according to claim 3, wherein the traffic monitoring unit can use the information stored therein by itself or by taking over the setting information of the traffic monitoring unit in another cache server. The traffic monitoring unit has means for monitoring an operation state of a traffic monitoring device in another cache server,
7. The cache according to claim 6, wherein when a failure occurs in a traffic monitoring unit in another cache server, communication of the cache server can be controlled by inheriting the setting information of the failed traffic monitoring unit. Server control system.

Images