TWI535249B - Automate the way to extend the network cache system - Google Patents

Description

Method of automating the expansion of the network cache system

The present invention relates to a network cache system, and more particularly to a method for automatically detecting the current network topology of each cache host without requiring the user to individually set it.

According to the rapid development of the online world, various kinds of network equipment are continuously developed and widely used in various living and working environments by various industries. This development trend not only accelerates the information. The speed and efficiency of circulation also bring great convenience to people in life and work. Among them, the connection technology between the computer and the Internet has progressed very rapidly. The bandwidth has been developed from the original 14.4K to the current 10M or even 100M bandwidth, and the connection method is also a single telephone dialing method. As it has evolved into a variety of wired and wireless connection methods, terminal devices have also begun to develop toward mobile communication devices. Therefore, people today are accustomed to obtaining various news information through the World Wild Web anytime and anywhere.

In fact, with the rapid growth of the Internet population, issues such as network bandwidth, server load capacity and browsing speed, etc. have become important considerations for related companies. After all, the above issues affect people's experience in network use. good. In general, in order to reduce the waste of network bandwidth and speed up users' browsing of web pages, the industry usually uses the Network Cache System, which mainly stores commonly used web pages in the network cache system. In the cache or on the hard drive, to speed up access, after all, for the same area, people in the area usually have a higher chance to browse the same web page. According to the location and function of the network cache system, it can be roughly divided into two categories. The first category is "Forward Cache", which is usually close to the user end. It is to store the content of the webpage that the user has watched. At present, in order to save the bandwidth of the export bandwidth, filter inappropriate information access or block the virus, the general enterprise, school or telecom operator usually adopts the "transfer cache" method.

The second type is "Reverse Cache", which usually stores specific The content of the server is mainly used to distribute the operation load of the network server, that is, the IP address queried by the device on the user side first points to the IP address of the cache host in the network cache system, if When the webpage content required by the user terminal is not stored in the cache host, the device on the client side is connected to the web server having the content of the webpage. Among them, "Content Delivery Network (CDN)" is an application of "Reverse Cache", which mainly distributes the burden of the network server, increases the speed and makes the website speed. Faster, and CDN servers are usually placed in many different places to increase the burden on them. At the same time, they will use temporary storage technology to temporarily store the contents of web pages originally stored on the network. On the CDN server, the bandwidth of the network server can be effectively reduced, and the user directly accesses the content on the CDN server when browsing the webpage. In addition, the CDN belongs to The application of the intranet of the network server is iCDN (Internetworking of Content Delivery Networks), which is commonly used to deploy streaming media services.

However, no matter which network cache system is used, it usually requires the user to change the network architecture or the complicated setting procedure to increase the number of cached hosts to improve the capacity. Scalability and normal operation, for example, changes in network routing involve related network device settings, network IP and DNS settings, etc., which require professional network planning capabilities, which will inevitably increase the use. The inconvenience, and the user can not easily expand the number of cache hosts required, so how to effectively solve the above problems is an important goal that many Internet service providers are striving to develop and hope to achieve.

In view of the existing network cache system, the installation has high complexity, causing users to spend a lot of configuration and debugging time. Therefore, after long-term efforts and experiments, the inventors finally developed and designed the present invention. A method of automatically expanding a network cache system, with the expectation that the present invention can effectively solve the aforementioned problems.

It is an object of the present invention to provide a method for automatically expanding a network cache system so that a user can easily increase the number of cache hosts without complicated settings, and the method is applied to a network cache. The network cache system includes a user agent (UA), at least one cache host, and a server, wherein the cache hosts are respectively wired to the user agent. Device or the server, or between each other The interconnecting method is configured to enable each of the cached hosts to execute a detection cache architecture program after booting, wherein the detection cache architecture program causes the cache host to send detections from all connections of the cache. Packets are sent to adjacent user agent devices, servers, or other cache hosts. The cache host will determine its own and adjacent user agent devices and servos based on the reply packets sent from the ports. The connection between the device or other cache host to obtain the connection mode of the network route is one-to-one type or multiple input type and multiple output type, and the cache host still determines according to the content of each reply packet. Whether it supports the Link Aggregation Control Protocol (LACP) architecture between itself and the adjacent user agent device, server or other cache host; if so, it and itself and the adjacent user agent device, All bandwidths between servers or other cache hosts are combined; if not, multi-way bridges are used between themselves and adjacent user agent devices, servers or other cache hosts (multip Le bridge) protocol structure; in this way, the user can detect the current network topology by itself after connecting the required number of cache hosts between the user agent device and the server. The situation, to avoid the need for the user to set each cache host, greatly improving the convenience of use.

Another object of the present invention is that when the network cache system is provided with a plurality of cache hosts, each cache host performs a load balancing program, wherein the load balancing program can operate according to each cache host. Ability to adjust the network traffic burden between each cache host to improve the efficiency and speed of the network cache system. In addition, since the foregoing procedures are started after each cache host is started (for example, rebooting) If the line is interrupted and the connection is restored, it will be executed automatically. Therefore, the user can easily expand the network cache system only after increasing the number of cached hosts and then starting the cache host.

For your convenience, the review committee can make a further understanding and understanding of the purpose, technical features and effects of the present invention. The embodiments are combined with the drawings, and the details are as follows:

[study]

no

〔this invention〕

1‧‧‧Network cache system

11‧‧‧User agent device

12A, 12B, 12C, 14A, 14B, 17A, 17B, 17C, 18A, 18B, 18C, 19A, 19B, 21A, 21B, 21C, 22A, 22B, 23A, 23B, 24A, 24B, 25A, 25B, 25C‧ ‧‧Network route

13, 16‧‧‧ cache host

15‧‧‧Server

1A is a first schematic diagram of a network cache system of the present invention having a cache host; FIG. 1B is a second schematic diagram of a network cache system of the present invention having a cache host; 1C is a third schematic diagram of a network cache system of the present invention having a cache host; FIG. 2A is a first schematic diagram of a network cache system of the present invention having a plurality of cache hosts; 2B is a second schematic diagram of a network cache system of the present invention having a plurality of cache hosts; and FIG. 2C is a third schematic diagram of a network cache system of the present invention having a plurality of cache hosts; 3A is a fourth schematic diagram of a network cache system of the present invention with a plurality of cache hosts; FIG. 3B is a schematic diagram of a state of use of FIG. 3A; and FIG. 3C is a schematic diagram of another state of use of FIG. And Figure 4 is a timing diagram of the operation of the network cache system of the present invention.

The present invention is a method for automatically expanding a network cache system. The method is applied to a Network Cache System. Referring to FIG. 1A, in an embodiment, the network is fast. The system 1 includes a user agent device (UA), at least one cache host 13 and a server 15, wherein each of the cache hosts 13 is wired to the user agent device 11 or The server 15, in addition, when a plurality of cache masters 13 are provided, the cache masters 13 can be connected to each other according to different connection modes (for example, series mode, parallel mode or series-parallel mode). In addition, in actual implementation, the user agent device 11 and the cache host 13 can be directly connected through a switch or other network device in addition to being directly connected.

For details, please refer to FIG. 1A. For example, the network cache system 1 is provided with a cache host 13 as an example. The cache host 13 supports various commonly used network communication protocols, for example: The IEEE 802.1ab Link Layer Discovery Protocol (LLDP), which organizes the information of itself (that is, the cache host 13) into TLV (Type/Length/Value, Type/Length/Value). And encapsulated in a Link Layer Discovery Protocol Data Unit (LLDPDU) message, and sent to The LLDPDUs that are directly connected to other devices are also saved in the form of a Management Information Base (MIB). Thus, with LLDP, the cache host 13 can save and manage itself. And directly connected to other devices to automatically detect the network architecture and automatically set for different wiring combinations; Bridge mode (Bridge) and support IEEE 802.11d STP network Spanning Tree Protocol detection loop ( Loop) automatic blocking, and the IEEE 802.3ad Link Aggregation Control Protocol (LACP), which combines the bandwidth with the convergence control protocol; in addition, the cache host 13 can support the IEEE 802.11q virtual area network. Virtual LAN (VLAN) protocol, which can be used in a VLAN trunk environment, and can share cached content with each other for fast access; further, the cache host 13 is hardware. The aspect also has a hardware bypass function to achieve the purpose of continuous line service. However, the foregoing network communication protocol is only an embodiment. In other embodiments of the present invention, the cache host 13 can support more or less network communication protocols according to different usage requirements. Heming Chen Ming.

As shown in FIG. 1A, when the cache host 13 is started, it executes a detection cache architecture program, wherein the detection cache architecture program causes the cache host 13 to self. All the connections send the detection packet to the adjacent user agent device 11 and the server 15, and the cache host 13 determines the user and the user agent according to the reply packet sent from each connection port. The connection between the device 11 and the server 15 is such that the connection manner between the device and the user agent device 11 and the server 15 is one-to-one type or multiple input type and multiple output type, one to one. The type of the user means that there is only one network route between the user agent device 11 and the cache host 13, and only one network route is passed between the cache host 13 and the server 15. The multi-input and multi-out type refers to the user agent device 11 and The cache host 13 has a plurality of network routes, and the cache host 13 and the server 15 have a plurality of network routes. Moreover, when the connection manner of the network route between the user agent device 11, the cache host 13 and the server 15 is one-in-one type, as shown in FIG. 1A, the user agent device 11 and the cache host are 13 and the server 15 adopt a bridge mode (Bridge). In particular, the foregoing startup condition of the cache host 13 includes that the cache host 13 is rebooted, or the line of the cache host 13 is interrupted or abnormal, and then the connection is resumed, etc., which is first described.

In addition, as shown in FIG. 1B, when the user agent device 11, the cache host 13 and the server 15 are connected to the network route in a three-in-two-out type (ie, the user agent device 11 and the cache device) When there are three network routes between the host 13 and two network routes between the cache host 13 and the server 15, the cache host will determine the use of itself and the neighbor according to the contents of the reply packets. Whether the connection device between the proxy device 11 and the server 15 supports a Link Aggregation Control Protocol (LACP) architecture, and if so, the three network routes between the user agent device 11 and the cache host 13 Will be combined (as indicated by the dashed box) to achieve the effect of triple bandwidth, and the two network routes between the cache host 13 and the server 15 will also be combined (as indicated by the dashed box) to achieve Double the effect of bandwidth.

In addition, as shown in FIG. 1C, if the connection manner between the user agent device 11, the cache host 13 and the server 15 is three-in and two-out, and LACP is not supported, then the fast The host 13 will adopt a multiple bridge protocol architecture, that is, the first network route 12A between the user agent device 11 and the cache host 13, and the first between the cache host 13 and the server 15. A network route 14A forms a first group of bridge modes, a second network route 12B between the user agent device 11 and the cache host 13, and a second network route 14B between the cache host 13 and the server 15. A second group of bridge modes is formed, and the third network route 12C between the user agent device 11 and the cache host 13 can be used as a backup. At the same time, each group of bridge modes can support hardware bypass (hardware). Bypass) function to increase the fault tolerance of the network cache system 1. Thus, when the cache host 13 completes the aforementioned detection cache architecture program, the current network topology can be obtained.

In other embodiments of the present invention, when the number of cache hosts exceeds one, as shown in FIGS. 2A-2C, it includes two cache hosts 13, 16 and the cache hosts 13, 16 are In the serial mode, after the cache hosts 13 and 16 are powered on and execute the detection cache architecture program, in addition to transmitting the detection packets to the adjacent user agent device 11 or the server 15, respectively. The detection packets are transmitted to each other. In addition, the cache hosts 13 and 16 can also use LACP according to whether they support LACP. In the state of supporting LACP, the cache hosts 13 and 16 can use the bandwidth together. As shown in FIG. 2B, the virtual box indicates that the bandwidth of the network routes is used in combination; in the state where the LACP is not supported, the cache hosts 13, 16 adopt a multi-way bridge protocol structure, such as As shown in FIG. 2C, the user agent device 11, the cache host 13, The first network route 17A, 18A, 19A between the server 16 and the server 15 forms a first group of bridge modes; the second network route between the user agent device 11, the cache host 13, 16 and the server 15 17B, 18B, and 19B will form a second group of bridge modes; the remaining network routes 17C and 18C will be used as backup.

When the cache hosts 13 and 16 are respectively connected to the user agent device 11 and the server 15, the network routes are interconnected, and the cache hosts 13 and 16 are also interconnected with each other to form a complicated wiring manner. When, as shown in FIG. 3A, the cache masters 13 and 16 are preferentially connected in parallel, so that the cache hosts 13 and 16 can be mutually redundant and can adopt an active/active setting mode. , or Active/Standby setting mode, in order to ensure network stability and maximize the bandwidth, the cache hosts 13, 16 will form the 3B after executing the detection cache architecture program. The network topology of the figure, wherein the user agent device 11, the cache host 13, 16 and the server 15 form a four-group bridge mode, wherein the first group of bridge modes is the user agent device 11, the cache host 13 network route 21A, 22A with the server 15; the second group of bridge mode is the user agent device 11, the network route 21B, 22B between the cache host 13 and the server 15; the third group of bridge mode is the user agent Device 11, network route 23A, 2 between cache host 16 and server 15. 4A; the fourth group of bridge modes is the user agent device 11, the network route 23B, 24B between the cache host 16 and the server 15; the remaining network routes 21C, 25A, 25B, 25C are used as backup, and no data is transmitted. (As shown in Figure 3B, the X symbol). In addition, when the network between the cache host 13 and the server 15 is unstable or unreachable (such as the X symbol shown in FIG. 3C), the data between the cache host 13 and the server 15 will be cached. The host 16 transmits.

In particular, when the network cache system 1 has a plurality of cache hosts 13, 16, the cache masters 13, 16 still execute a load balancing program, wherein the load balancing program can be fast according to each Take the operation capability of the host 13, 16 (for example, the size of the cache space, the operation speed of the central processing unit (CPU), the size of the memory capacity or the current traffic load, etc.), and adjust each cache host 13, 16 The data capacity is used to distribute the network traffic burden between the cache hosts 13 and 16. In addition, the data cache mode of the network cache system 1 of the present invention is described. Referring to FIG. 4, first, the user agent device 11 directly performs data transmission with the server 15, such as: The user agent device 11 transmits a message to the server 15. A demand packet (as in a1 of Figure 4), the server will transmit a first response packet to the user agent device 11 (as in Figure 4 a2), 嗣, when the user agent device 11 and the servo After the two cache hosts 13 and 16 are connected in series, the cache hosts 13 and 16 execute the detection cache architecture program (that is, the cache hosts 13, 16 send detection packets from all the connections of the host to Adjacent user agent device 11, server 15 or other cache host 16, 13), 快, the cache hosts 13, 16 will determine the current network according to the reply packet sent by each connection port. After the road topology (such as b1, b2 in Figure 4) and/or after performing the load balancing procedure, they will be connected to each other and begin to provide services (as shown in Figure 4b3).

Referring to FIG. 4, when the user agent device 11 transmits a second request packet and is transparently directed to the cache host 13 (as shown in FIG. 4 c1), the cache host 13 checks itself. The content of the stored data, when the cache host 13 finds that it does not have the data content required by the second demand packet (such as c2 in FIG. 4), the cache host 13 will forward the second demand packet to the fast Taking the host 16 (as shown in Figure 4 c3), when the cache host 16 itself does not have the data content, it will forward the second demand packet to the server 15 (as shown in Figure 4, c4), when the cache After receiving the second response packet (such as c5 in FIG. 4) sent by the server 15, the host 16 stores the content of the data in the second response packet, and transmits the second response packet to the second response packet. The cache host 13 (such as c6 in FIG. 4), after receiving the second response packet, the cache host 16 stores the content of the data in the second response packet, and transmits the second response packet. To the user agent device 11 (as shown in Figure 4, c7). In addition, when the user agent device 11 transmits a third demand packet to the cache host 13 (as shown in FIG. 4, d1), the cache host 13 finds that it has stored the data required to meet the third requirement packet. After the content (such as d2 in FIG. 4), the cache host 13 directly transmits a third response packet to the user agent device 11 (as shown in FIG. 4, d3). Thus, the network cache is performed by the present invention. In the system 1, after the user connects the required number of cache hosts 13 and 16 between the user agent device 11 and the server 15, the cache hosts 13 and 16 can detect the current network by themselves. The topology of the road avoids the need for the user to set each cache host 13, 16 to greatly improve the convenience of use, and each time the user re-boots the cache host 13, 16 each time, The cache host 13, 16 will actively re-detect the current network topology so that the user can expand the network cache system 1.

According to the above, it is only a preferred embodiment of the present invention, but the present invention is The scope of the present invention is not limited thereto, and those skilled in the art, based on the technical content disclosed in the present invention, can easily think of equivalent changes, and should not fall within the protection scope of the present invention.

1‧‧‧Network cache system

11‧‧‧User agent device

13, 16‧‧‧ cache host

15‧‧‧Server

Claims (4)

A method for automatically expanding a network cache system, the method being applied to a network cache system, the network cache system comprising a user agent device, at least one cache host and a server, wherein the fast The host computers are respectively wired to the user agent device or the server, or are connected to each other. The method is such that after each cache host is started, the following steps can be performed: executing a detection cache architecture program, The detection cache architecture program causes the cache host to send a detection packet from all its connections to an adjacent user agent device, server or other cache host; according to the connection Responding to the packet and determining the connection between itself and the adjacent user agent device, server or other cache host to know that it is between the user and the adjacent user agent device, server or other cache host. The wiring mode of the network route is one-in-one-out type or multi-input and multi-out type; and in the state of multi-input and multi-out type, judging itself and neighboring according to the content of each reply packet Whether the proxy aggregation device, the server, or other cache host supports the bandwidth aggregation protocol architecture, and if so, combines all the bandwidths between itself and the adjacent user agent device, server, or other cache host. If not, the multi-way bridging protocol architecture is employed between itself and the adjacent user agent device, server or other cache host. According to the method of claim 1, in the state that the network cache system is provided with a plurality of cache hosts, each cache host further performs a load balancing program to distribute the cache between the cache hosts. Network traffic burden. According to the method of claim 1, in the state in which the network cache system is provided with a plurality of cache hosts, each of the cache hosts determines that they are respectively associated with the user agent device, the server, and others. If there is a network route interconnection between the main rooms, the parallel mode will be preferred. The method of any one of claims 1 to 3, wherein the cache host supports a network link detection protocol, a network extension tree protocol, and/or a virtual area network protocol.