WO2016101748A1

WO2016101748A1 - Method and device for caching network connection

Info

Publication number: WO2016101748A1
Application number: PCT/CN2015/095455
Authority: WO
Inventors: 任勇全; 赵安安; 陈磊
Original assignee: 北京奇虎科技有限公司; 奇智软件（北京）有限公司
Priority date: 2014-12-27
Filing date: 2015-11-24
Publication date: 2016-06-30
Also published as: CN104580435B; CN104580435A

Abstract

Provided in an embodiment of the present invention are a method and device for caching a network connection, the method specifically comprising: after completing one server access, determining a most recent selection probability of each network connection in a cache according to characteristics of a server load balancing mode, the most recent selection probability representing the probability that the server corresponding to the network connection will be selected next on the basis of the load balancing mode; ordering the network connections in the cache according to the most recent selection probabilities; and replacing the network connection having the smallest most recent selection probability in the cache, and preserving the network connection having the largest most recent selection probability in the cache. The embodiment of the present invention can avoid the problem of cache hit failure caused by replacing the network connection having the largest most recent selection probability in the cache, thus improving a network connection cache hit ratio.

Description

Method and device for buffering network connection

Technical field

The present invention relates to the field of communications technologies, and in particular, to a method and apparatus for caching a network connection.

Background technique

The ordinary web access process is a typical client and server model. As shown in FIG. 1, the user makes a request by using a program on the client 101, such as a browser, and the web server 103 responds to the request and provides corresponding data, and the proxy server 102 is The request and data are forwarded between the client 101 and the web server 103, and the request can be evenly distributed on the respective web servers 103 by load balancing of the web server 103. Assuming that load balancing is implemented on four Web servers by means of polling, each client's request can be assigned to different Web servers one by one in the order of Web server 1, Web server 2, Web server 3, and Web server 4.

When the number of concurrent accesses is large, the proxy server maintains a TCP (Transmission Control Protocol) connection with the Web server without consuming system resources, and may use a preset number of cache areas to cache the Web server. The TCP connection, here, the preset number is usually smaller than the number of web servers, such as when the number of web servers is 4, the number of cache areas is 3, and so on.

Assume that the cache area 1, the cache area 2, and the cache area 3 are used to cache the TCP connection of the Web server 1, the Web server 2, the Web server 3, and the Web server 4. In the current state, the cache area 1, the buffer area 2, and the buffer area 3 are respectively cached. The TCP connection between the Web server 1, the Web server 2, and the Web server 3, then, in the existing scheme, when the cache replacement is performed in combination with the above polling mode, the cache area is accessed after the Web server 1 is accessed by using the content in the buffer area 1. The replacement of the content in 2 causes the cache hit to fail when accessing the web server 2, and the connection to the web server 2 needs to be re-established; and after accessing the web server 2, the content in the cache area 3 is replaced, resulting in access to the web server. 3 o'clock cache hit failed, need to re-establish to the web service The connection of the device 3. It can be seen that the existing scheme has a problem that the cache hit rate is low.

Summary of the invention

In view of the above problems, the present invention has been made in order to provide a method and apparatus for caching a network connection that overcomes the above problems or at least partially solves the above problems.

According to an aspect of the present invention, a method for caching a network connection is provided, including:

After the server access is completed, the latest selection probability of each network connection in the cache is determined according to the characteristics of the server load balancing mode; wherein the latest selection probability is used to indicate the probability that the network connection corresponding server is selected next time according to the load balancing mode;

Sorting each network connection in the cache according to the latest selection probability;

Replacing the network connection with the least recent selection probability in the cache, and retaining the network connection with the most recent selection probability in the cache.

According to another aspect of the present invention, there is provided a computer program comprising computer readable code, when said computer readable code is run on a computing device, causing said computing device to perform a network connection as described above Cache method.

According to still another aspect of the present invention, there is provided a computer readable medium storing a computer program as described above.

According to still another aspect of the present invention, a cache device for network connection is provided, including:

The probability determining module is configured to determine, according to a characteristic of the server load balancing manner, a latest selection probability of each network connection in the cache after completing a server access; wherein the latest selection probability is configured to indicate that the network connection corresponding server is under load balancing mode Probability of secondary selection;

a probability ordering module configured to sort network connections in the cache according to the latest selection probability;

And a permutation retention module configured to replace a network connection with the least recent selection probability in the cache, and to retain a network connection with the most recent selection probability in the cache.

According to an embodiment of the present invention, a method and apparatus for caching a network connection, after completing a server access, determining a latest selection probability of each network connection in the cache according to characteristics of a server load balancing manner, according to the latest selection probability, Sort each network connection in the cache, and juxtapose Replacing the network connection with the least recent selection probability in the cache, and retaining the network connection with the most recent selection probability in the cache; since the latest selection probability is used to indicate that the network connection corresponding server is selected next time according to the load balancing mode Probability, therefore, the embodiment of the present invention only replaces the network connection with the smallest selection probability in the cache and the network connection that retains the most recent selection probability in the cache, and can avoid the network connection with the most recent selection probability in the cache. The problem of cache hit failure caused by being replaced can improve the cache hit rate of the network connection.

The above description is only an overview of the technical solutions of the present invention, and the above-described and other objects, features and advantages of the present invention can be more clearly understood. Specific embodiments of the invention are set forth below.

DRAWINGS

Various other advantages and benefits will become apparent to those skilled in the art from a The drawings are only for the purpose of illustrating alternative embodiments and are not to be considered as limiting. Throughout the drawings, the same reference numerals are used to refer to the same parts. In the drawing:

FIG. 1 is a schematic structural diagram of an HTTP access system;

2 is a flow chart showing the steps of a method for caching a network connection according to an example of the present invention;

FIG. 3 is a schematic structural diagram of a cache device for network connection according to an embodiment of the present invention; FIG.

Figure 4 shows schematically a block diagram of a computing device for performing the method according to the invention;

Fig. 5 schematically shows a storage unit for holding or carrying program code implementing the method according to the invention.

detailed description

Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While the embodiments of the present invention have been shown in the drawings, the embodiments Rather, these embodiments are provided to provide a more thorough The disclosure is to be understood, and the scope of the present disclosure can be fully conveyed to those skilled in the art.

2 is a schematic flow chart of a method for caching a network connection according to an embodiment of the present invention, which may specifically include the following steps:

Step 201: After completing a server access, determine a latest selection probability of each network connection in the cache according to the characteristics of the server load balancing mode; wherein the latest selection probability is used to indicate that the network connection corresponding server is selected next time according to the load balancing mode The probability;

Step 202: Sort each network connection in the cache according to the latest selection probability; and

Step 203: Replace the network connection with the smallest latest selection probability in the cache, and retain the network connection with the most recent selection probability in the cache.

The embodiment of the present invention can be applied to various proxy servers, which are between the client and the server, can be used for forwarding requests and data between the client and the server, and for implementing load balancing on the server. The request will be evenly distributed on each server, and can also be used to cache the network connection with the server to increase the communication speed between the client and the server without consuming system resources.

The load balancing process of the server is mainly: after completing a server access, determining how to select the next server according to the characteristics of the load balancing mode, and forwarding a new access request to it; wherein, one server access process forwards the access request to The process of selecting the server.

The condition that the cache hit succeeds in the load balancing process of the server is specifically as follows: the server selected according to the characteristics of the load balancing mode is successfully matched with the network connection in the cache, that is, the network connection corresponding to the selected server exists in the cache.

In order to save system resources, the maximum number of network connections that the cache can store is usually smaller than the number of servers. Thus, if the network connection stored in the cache does not change, there is always a server network connection that is not used, so that load balancing cannot be achieved. purpose. Cache replacement can achieve load balancing by replacing one or more network connections in the cache with other network connections.

In order to solve the problem that the existing solution has a low cache hit rate, the embodiment of the present invention determines the network connections in the cache according to the characteristics of the server load balancing mode after completing a server access. The latest selection probability, according to the latest selection probability, sorting each network connection in the cache, and replacing the network connection with the smallest latest selection probability in the cache, and retaining the network connection with the most recent selection probability in the cache Because the latest selection probability is used to indicate the probability that the network connection corresponding server is selected next time in the load balancing mode, the embodiment of the present invention only replaces the network connection with the smallest latest selection probability in the cache and retains the cache. The latest selection of the most probable network connection means avoids the problem of cache hit failure caused by the replacement of the network connection with the most recent selection probability in the cache, thereby improving the cache hit rate of the network connection.

The embodiment of the present invention may provide a solution for determining the latest selection probability of each network connection in the cache according to the characteristics of the server load balancing mode:

Option One,

In the first solution, the load balancing mode may be a polling mode, and the step of determining the latest selection probability of each network connection in the cache according to the characteristics of the server load balancing mode may specifically include: according to the latest network connection in the cache. The usage time is used to determine the latest selection probability of each network connection in the cache, wherein the latest selection probability of the network connection closest to the current time is the largest, and the latest selection probability of the network connection most recently from the current time is the smallest.

Specifically, the polling mode may be configured to send new requests to the next server in a load balancing process, so that the servers are successively and repeatedly restarted, that is, each server is alternately selected in an equal position.

According to the characteristics of the foregoing polling mode, it can be known that if a network connection corresponding server in the cache has just been used, it can be used again after a polling period, and therefore, according to the latest usage time of each network connection in the cache. Determining the latest selection probability of each network connection in the cache. Generally, the newer the latest usage time of the network connection in the cache is, the smaller the corresponding latest selection probability is, so that the cache is obtained according to the latest selection probability. The sorting of each network connection may be: sorting each network connection in the cache according to the latest usage time.

In an application example 1 of the present invention, it is assumed that the buffer area 1, the buffer area 2, and the buffer area 3 are used to cache the TCP connection of the web server 1, the web server 2, the web server 3, and the web server 4, and the buffer area in the current state. 1. Cache area 2 and cache area 3 respectively cache web services 1, web server 2 and web server 3 TCP connection.

Then, when the cache is replaced by the polling mode, the embodiment of the present invention can determine the network connections in the cache according to the latest usage time of each network connection in the cache after accessing the web server 1 in the cache area 1. The order of the latest selection probability is: buffer area 1 <cache area 2 < cache area 3, therefore, the content in the buffer area 1 can be replaced, and the contents in the buffer area 2 and the buffer area 3 can be reserved, thereby ensuring access to the web server. At 2 o'clock, the cache hit succeeds, and there is no need to re-establish the connection to the web server 2; then, after accessing the web server 2, the order of determining the latest selection probabilities of the network connections in the cache is: buffer 2 < cache area 1 < cache area 3 Therefore, the content in the buffer area 2 can be replaced, so that the cache hit is successful when accessing the web server 3, and there is no need to re-establish the connection to the web server 3. It can be seen that the present invention greatly improves the cache hit rate.

Option II,

In the second solution, the load balancing mode may be a hash mode, and the step of determining the latest selection probability of each network connection in the cache according to the characteristics of the server load balancing mode may include: The selection probability of one time determines the latest selection probability of each network connection in the cache, wherein the latest selection probability of the network connection selected last time is the smallest, and the latest selection probability of the network connection with the smallest selection probability is the largest.

The Hash method can send a request to the server according to a certain rule by a single-shot irreversible hash function, which usually has the following characteristics:

1. Balance; balance means that the results of Hash should be evenly distributed to each server to solve the load balancing problem;

2, monotonic; monotonic means that when adding or deleting a server, the same key (key) access value is always the same;

3. Dispersibility; Dispersibility means that data should be distributed and stored on each server.

For example, in an application embodiment of the present invention, the hash mode may be dispersed according to the modulo result of the number of servers, that is, the integer hash value of the key is first obtained, and then the integer hash value is performed on the number of servers. The modulo operation is performed, and the server is selected according to the corresponding modulo result.

According to the characteristics of the above hash mode, each network connection is in the last selection probability and the next time The selection probability is usually reversed, that is, the previous selection probability is greater than the next selection probability. Therefore, the latest selection probability of each network connection in the cache can be determined according to the previous selection probability of each network connection in the cache.

For the above application example 1, in the embodiment of the present invention, when the cache replacement is performed in combination with the hash mode, after the content of the network in the cache area 1 is accessed, the latest selection probability of the network connection in the buffer area 1 is 100. %, it can be determined that the next latest selection probability of the buffer area 1 is approximately 0. Therefore, the content in the buffer area 1 can be replaced, and the contents in the buffer area 2 and the buffer area 3 can be reserved, thereby ensuring access to the web server 2 When the cache hit succeeds, there is no need to re-establish the connection to the Web server 2; then, after accessing the Web server 2, it is determined that the latest selection probability of the network connection in the cache 2 is approximately 0, so the content in the buffer 2 can be replaced, thereby It is guaranteed that the cache hit is successful when accessing the web server 3, and there is no need to re-establish the connection to the web server 3. It can be seen that the present invention greatly improves the cache hit rate.

third solution,

In the third solution, the load balancing mode may be the lowest missing mode, and the step of determining the latest selection probability of each network connection in the cache according to the characteristics of the server load balancing mode may specifically include: according to each network connection in the cache Corresponding to the historical processing access number of the server, determining the latest selection probability of each network connection in the cache, wherein the server with the largest number of historical processing accesses has the latest selection probability of the network connection, and the server with the least number of historical processing accesses the latest selection of the network connection. The probability is the biggest.

The lowest missing method balances the request records to each server and sends the next request to the server with the least number of historical processing accesses. Therefore, scheme 3 may determine the latest selection probability of each network connection in the cache according to the historical processing access number of the corresponding server in each cache connection in the cache. Specifically, for the server with more historical processing access, the corresponding network connection The latest selection probability is smaller.

For the above application example 1, in the embodiment of the present invention, when the cache replacement is performed in combination with the lowest missing mode, after the content of the cache server 1 is used to access the web server 1, the current web server 1, the web server 2, the web server 3, and the web server are assumed. The number of historical processing accesses of 4 is 100w, 80w, 110w, and 90w respectively. Here, w represents 10,000, then, each of the caches can be determined. The order of the latest selection probability of the network connection is: buffer area 3 <buffer area 1 < cache area 2, therefore, the content in the buffer area 3 can be replaced, and the contents in the buffer area 1 and the buffer area 2 can be reserved, thereby being able to guarantee When the web server 2 is accessed, the cache hit is successful, and there is no need to re-establish the connection to the web server 2. It can be seen that the present invention can improve the cache hit rate.

Option 4

In the fourth solution, the load balancing mode may be the fastest response mode, and the step of determining the latest selection probability of each network connection in the cache according to the characteristics of the server load balancing mode may specifically include: according to each network in the cache The response time of the corresponding server is connected to determine the latest selection probability of each network connection in the cache. The server with the shortest response time has the highest latest selection probability corresponding to the network connection, and the server with the longest response time has the lowest latest selection probability of the network connection.

The fastest response method can record the network response time of each server and assign the next incoming request to the server with the shortest response time. In this way, scheme 4 may determine the latest selection probability of each network connection in the cache according to the response time of each network connection corresponding server in the cache. Specifically, the shorter the response time of the server, the greater the latest selection probability of the corresponding network connection. .

For the above application example 1, in the embodiment of the present invention, when the cache replacement is performed in combination with the lowest missing mode, after the content of the cache server 1 is used to access the web server 1, the current web server 1, the web server 2, the web server 3, and the web server are assumed. The response time of 4 is 10ms, 20s, 25ms and 30ms respectively. Then, the order of the latest selection probability of each network connection in the cache can be determined as: buffer 3<buffer 2<buffer 1 and therefore, buffer 3 can be The content is replaced, and the contents of the buffer area 1 and the buffer area 2 are reserved, so that the cache hit is successful when accessing the web server 1, and there is no need to re-establish the connection to the web server 1. It can be seen that the present invention can improve the cache hit rate.

The foregoing describes the characteristics of the polling mode, the hash mode, the minimum missing mode, and the fastest response mode, and the corresponding scheme for determining the latest selection probability of each network connection in the cache. It should be noted that those skilled in the art Depending on the actual situation, any one of the foregoing solutions may be used, or the other load balancing mode may be adopted, and the latest selection probability of each network connection in the cache may be determined according to the characteristics of the other load balancing modes. And its corresponding The scheme for determining the latest selection probability of each network connection in the cache according to the characteristics of other load balancing methods is not limited.

In summary, the embodiment of the present invention only replaces the network connection with the smallest recent selection probability in the cache and the network connection that retains the most recent selection probability in the cache, and can avoid the network connection with the most recent selection probability in the cache being The problem of cache hit failure caused by replacement can improve the cache hit rate of network connections.

For the method embodiments, for the sake of simple description, they are all expressed as a series of action combinations, but those skilled in the art should understand that the embodiments of the present invention are not limited by the described action sequence, because the embodiment according to the present invention Some steps can be performed in other orders or at the same time. In the following, those skilled in the art should also understand that the embodiments described in the specification are optional embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

Referring to FIG. 3, a schematic structural diagram of a cache device for network connection according to an embodiment of the present invention is shown, which may specifically include the following modules:

The probability determining module 301 is configured to determine, according to the characteristics of the server load balancing manner, the latest selection probability of each network connection in the cache after completing a server access; wherein the latest selection probability is configured to indicate that the server is connected according to the load balancing mode The probability of being selected next time;

The probability ranking module 302 is configured to sort the network connections in the cache according to the latest selection probability;

The permutation retention module 303 is configured to replace the network connection with the least recent selection probability in the cache, and to retain the network connection with the most recent selection probability in the cache.

In an optional embodiment of the present invention, the load balancing mode is a polling mode, and the probability determining module 301 may further include:

The first probability determining submodule is configured to determine a latest selection probability of each network connection in the cache according to a latest usage time of each network connection in the cache.

In another preferred embodiment of the present invention, the load balancing mode is a hash mode, and the probability determining module 301 may further include:

The second probability determining submodule is configured to determine a latest selection probability of each network connection in the cache according to a previous selection probability of each network connection in the cache.

In a further preferred embodiment of the present invention, the load balancing mode is the lowest missing mode, and the probability determining module 301 may further include:

The third probability determining submodule is configured to determine the latest selection probability of each network connection in the cache according to the historical processing access number of the corresponding server in each cache connection in the cache.

In another preferred embodiment of the present invention, the load balancing mode is the fastest response mode, and the probability determining module 301 may further include:

The fourth probability determining submodule is configured to determine, according to the response time of the corresponding server in each cache connection in the cache, the latest selection probability of each network connection in the cache, wherein the server with the shortest response time has the highest latest selection probability of the network connection. The server with the longest response time has the lowest probability of the latest selection of network connections.

For the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method embodiment.

The various component embodiments of the present invention may be implemented in hardware, or in a software module running on one or more processors, or in a combination thereof. Those skilled in the art will appreciate that a microprocessor or digital signal processor (DSP) may be used in practice to implement some or all of the functionality of some or all of the components of the method and apparatus for caching network connections in accordance with embodiments of the present invention. . The invention can also be implemented as a device or device program (e.g., a computer program and a computer program product) for performing some or all of the methods described herein. Such a program implementing the invention may be stored on a computer readable medium or may be in the form of one or more signals. Such signals may be downloaded from an internet platform, provided on a carrier signal, or provided in any other form.

For example, Figure 4 illustrates a computing device, such as a search engine server, that can implement the above described method in accordance with the present invention. The computing device conventionally includes a processor 410 and a computer program product or computer readable medium in the form of a memory 430. The memory 430 may be an electronic memory such as a flash memory, an EEPROM (Electrically Erasable Programmable Read Only Memory), an EPROM, a hard disk, or a ROM. The memory 430 has a storage space 450 that stores program code 451 for performing any of the method steps described above. For example, storage space 450 storing program code may include various program code 451 for implementing various steps in the above methods, respectively. The program code can be read from or written to one or more computer program products. These computer program products include program code carriers such as hard disks, compact disks (CDs), memory cards or floppy disks. Such computer program products are typically portable or fixed storage units such as those shown in FIG. The storage unit may have storage segments, storage spaces, and the like that are similarly arranged to memory 430 in the computing device of FIG. The program code can be compressed, for example, in an appropriate form. Typically, the storage unit comprises computer readable code 451' for performing the steps of the method according to the invention, ie code that can be read by a processor, such as 410, which, when executed by the server, causes the server to execute Each of the steps in the method described above.

"an embodiment," or "an embodiment," or "an embodiment," In addition, it is noted that the phrase "in one embodiment" is not necessarily referring to the same embodiment.

In the description provided herein, numerous specific details are set forth. However, it is understood that the embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques are not shown in detail so as not to obscure the understanding of the description.

It is to be noted that the above-described embodiments are illustrative of the invention and are not intended to be limiting, and that the invention may be devised without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as a limitation. The word "comprising" does not exclude the presence of the elements or steps that are not recited in the claims. The word "a" or "an" The invention can be implemented by means of hardware comprising several distinct elements and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means can be embodied by the same hardware item. The use of the words first, second, and third does not indicate any order. These words can be interpreted as names.

In addition, it should also be noted that the language used in this specification is primarily for readability and teaching. It is chosen for the purpose, and is not chosen to explain or define the subject matter of the invention. Therefore, many modifications and changes will be apparent to those skilled in the art without departing from the scope of the invention. The disclosure of the present invention is intended to be illustrative, and not restrictive, and the scope of the invention is defined by the appended claims.

Claims

A method of caching a network connection, comprising:

After the server access is completed, the latest selection probability of each network connection in the cache is determined according to the characteristics of the server load balancing mode; wherein the latest selection probability is used to indicate the probability that the network connection corresponding server is selected next time according to the load balancing mode;

Sorting each network connection in the cache according to the latest selection probability;

Replacing the network connection with the least recent selection probability in the cache, and retaining the network connection with the most recent selection probability in the cache.
The method of claim 1, wherein the load balancing mode is a polling mode, and the step of determining the latest selection probability of each network connection in the cache according to the characteristics of the server load balancing mode includes:

The latest selection probability of each network connection in the cache is determined according to the latest usage time of each network connection in the cache, wherein the latest connection probability of the network connection closest to the current time is the largest, and the network with the latest usage time is the farthest from the current time. The latest selection probability of the connection is minimal.
The method of claim 1, wherein the load balancing mode is a hash mode, and the step of determining the latest selection probability of each network connection in the cache according to the characteristics of the server load balancing mode comprises:

According to the last selection probability of each network connection in the cache, the latest selection probability of each network connection in the cache is determined, wherein the latest selection probability of the last selected network connection is the smallest, and the latest network connection with the lowest selection probability is the latest. The selection probability is the largest.
The method of claim 1, wherein the load balancing mode is the lowest missing mode, and the step of determining the latest selection probability of each network connection in the cache according to the characteristics of the server load balancing mode includes:

Determining the latest selection probability of each network connection in the cache according to the historical processing access number of the corresponding server in the cache, wherein the server with the largest number of historical processing accesses has the lowest selection probability of the network connection and the least number of historical processing accesses. The server has the highest probability of the latest selection of network connections.
The method of claim 1, wherein the load balancing mode is the fastest response mode, and the determining the latest selection of each network connection in the cache according to the characteristics of the server load balancing mode The steps to choose a probability include:

Determining, according to the response time of each network connection corresponding server in the cache, determining the latest selection probability of each network connection in the cache, wherein the server with the shortest response time has the highest latest selection probability corresponding to the network connection, and the server with the longest response time corresponds to the network connection. The latest selection probability is the smallest.
A computer program comprising computer readable code causing the computing device to perform a method of caching a network connection according to any one of claims 1 to 5 when the computer readable code is run on a computing device .
A computer readable medium storing the computer program of claim 6.
A cache device for network connection, comprising:

The probability determining module is configured to determine, according to a characteristic of the server load balancing manner, a latest selection probability of each network connection in the cache after completing a server access; wherein the latest selection probability is configured to indicate that the network connection corresponding server is under load balancing mode Probability of secondary selection;

a probability ordering module configured to sort network connections in the cache according to the latest selection probability;

And a permutation retention module configured to replace a network connection with the least recent selection probability in the cache, and to retain a network connection with the most recent selection probability in the cache.
The device of claim 8, wherein the load balancing mode is a polling mode, and the probability determining module comprises:

The first probability determining submodule is configured to determine, according to the latest usage time of each network connection in the cache, a latest selection probability of each network connection in the cache, wherein the most recent selection probability of the network connection closest to the current time is the most recent, most recently The latest selection probability of using a network connection that is farthest from the current time is the smallest.
The apparatus according to claim 8, wherein the load balancing mode is a hash mode, and the probability determining module comprises:

The second probability determining submodule is configured to determine, according to the last selection probability of each network connection in the cache, the latest selection probability of each network connection in the cache, wherein the latest selection probability of the last selected network connection is the smallest, the last time The latest selection probability of the network connection with the least probability of selection maximum.
The apparatus according to claim 8, wherein the load balancing mode is a minimum missing mode, and the probability determining module comprises:

The third probability determining sub-module is configured to determine the latest selection probability of each network connection in the cache according to the historical processing access number of each network connection corresponding server in the cache, wherein the server with the largest number of historical processing accesses corresponds to the latest network connection The server with the lowest selection probability and the least number of historical processing accesses has the highest probability of the latest selection of the network connection.
The device of claim 8, wherein the load balancing mode is the fastest response mode, and the probability determining module comprises:

The fourth probability determining submodule is configured to determine, according to the response time of the corresponding server in each cache connection in the cache, the latest selection probability of each network connection in the cache, wherein the server with the shortest response time has the highest latest selection probability of the network connection. The server with the longest response time has the lowest probability of the latest selection of network connections.