Connection scheduling method and system for data transmission

By detecting abnormalities in the QUIC connection between the cache server and the user side and switching to a healthy backup server, the data transmission problem caused by the poor connection quality of the cache server and the user side is solved, and the stability and efficiency of data transmission are achieved.

WO2024244401A1PCT designated stage expired Publication Date: 2024-12-05CHINA TELECOM CLOUD TECH CO LTD
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
PCT/CN2023/139634
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2023-05-31
Filing Date
2023-12-18
Publication Date
2024-12-05

AI Technical Summary

Technical Problem

The poor quality of the connection between the cache server and the user side leads to a worsening of the data transmission effect, and the prior art is difficult to effectively solve this problem.

Method used

A connection scheduling method for data transmission is adopted to obtain and determine the status of the QUIC connection between the cache server and the user side, generate a migration signal, determine the backup cache server with a higher health value, and switch the connection to ensure the stability of data transmission. .

Benefits of technology

It realizes automatic detection and timely detection of abnormal connections, quickly switch to a healthy backup cache server, ensuring efficient and stable data transmission, and the connection switching operation is simple and does not affect the data reception of the user.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN2023139634_05122024_PF_FP_ABST
    Figure CN2023139634_05122024_PF_FP_ABST
Patent Text Reader

Abstract

The embodiments of the present description mainly relate to the technical field of connection scheduling, and specifically to a connection scheduling method and system for data transmission. The connection scheduling method for data transmission comprises the following steps: acquiring a first connection, the first connection being a connection between a first cache server and a user end, and the first connection being a QUIC connection; determining that the first connection is a normal connection or an abnormal connection; when the first connection is an abnormal connection, generating a first migration signal; after the first migration signal is acquired, determining a second cache server; when determining the second cache server is completed, generating a second migration signal; and after the second migration signal is acquired, cutting off the first connection and forming a second connection. The connection scheduling method and system in the embodiments of the present description can solve the problem in the prior art of poor data transmission effects caused by poor connection quality between cache servers and user ends.
Need to check novelty before this filing date? Find Prior Art

Description

A connection scheduling method and system for data transmission Technical Field

[0001] The embodiments of this specification mainly relate to the field of connection scheduling technology, and specifically to a connection scheduling method and system for data transmission. Background Art

[0002] The content distribution network includes multiple groups of cache servers distributed in various regions. Each group of cache servers stores a copy of the data. When a user has a data request, the data is provided to the user through a cache server in the cache server group closest to the user.

[0003] When a cache server provides data to users, it must connect to the client. The quality of the connection directly affects the effectiveness of data transmission. When the cache server malfunctions or experiences network jitter, the connection quality between the cache server and the client deteriorates, leading to poor data transmission.

[0004] Summary of the Invention

[0005] In response to the problems existing in the prior art, the embodiments of this specification propose a connection scheduling method and system for data transmission to solve the problem in the prior art that the data transmission effect deteriorates due to the deterioration of the connection quality between the cache server and the user end.

[0006] In a first aspect, an embodiment of the present disclosure provides a connection scheduling method for data transmission, comprising the following steps: obtaining a first connection, where the first connection is a connection between a first cache server and a client, and the first connection is a QUIC connection;

[0007] determining whether the first connection is a normal connection or an abnormal connection;

[0008] When the first connection is an abnormal connection, generating a first migration signal;

[0009] After obtaining the first migration signal, determining a second cache server;

[0010] When the second cache server determines that the migration is completed, generating a second migration signal;

[0011] After obtaining the second migration signal, the first connection is disconnected and a second connection is formed. The second connection is a connection between the second cache server and the user terminal, and the second connection is a QUIC connection.

[0012] Preferably, determining whether the first connection is a normal connection or an abnormal connection specifically includes the following steps:

[0013] calculating a data transmission rate of the first connection;

[0014] Calculating a data packet loss rate of the first connection;

[0015] When the data transmission rate of the first connection is less than a preset data transmission rate threshold and / or when the data packet loss rate of the first connection is greater than a preset data packet loss rate threshold, the first connection is determined to be an abnormal connection; otherwise, the first connection is determined to be a normal connection.

[0016] Preferably, determining the second cache server specifically includes the following steps:

[0017] Calculate the health values ​​of other cache servers in the cache server group;

[0018] A second cache server is determined according to the health value.

[0019] Preferably, the health value of each cache server is calculated using the following formula:

[0020] Among them, ioutil is the IO usage of the disk in the cache server, max The maximum IO usage of the disk in the cache server, ioutil min is the minimum IO usage of the disk in the cache server, x ioutil The first health value of the cache server;

[0021] Among them, iowait is the percentage of cache server CPU idle but with unfinished disk IO requests in a sampling period. max iowait is the maximum percentage of cache servers with idle CPU but unfinished disk IO requests within a sampling period. min The minimum percentage of cache server CPU idle but unfinished disk IO requests in a sampling period, x iowait The second health value of the cache server;

[0022] in, is the average CPU usage of the cache server, c is an empirical constant, and x cpu is the third health value of the cache server; X=max(x ioutil , x iowait , x cpu ) (4)

[0023] Where X is the health value of the cache server.

[0024] Preferably, disconnecting the first connection and establishing the second connection specifically includes the following steps:

[0025] Get the ID password and encryption method of the first connection;

[0026] The first connection is disconnected and the second connection is completed according to the ID password and encryption method.

[0027] In a second aspect, an embodiment of this specification provides a connection scheduling system for data transmission, including:

[0028] A first connection acquisition module, configured to acquire a first connection, where the first connection is a connection between a first cache server and a client, and the first connection is a QUIC connection;

[0029] A first connection determination module, configured to determine whether the first connection is a normal connection or an abnormal connection;

[0030] A first transition signal generating module, configured to generate a first transition signal when the first connection is an abnormal connection;

[0031] A first transition signal acquisition module, configured to acquire the first transition signal;

[0032] A second cache server determining module, configured to determine a second cache server after obtaining the first migration signal;

[0033] A second migration signal generating module, configured to generate a second migration signal when the second cache server determines that the migration is completed;

[0034] A second transition signal acquisition module, configured to acquire the second transition signal;

[0035] A connection disconnection formation module is configured to disconnect the first connection and form a second connection after obtaining the second migration signal, wherein the second connection is a connection between the second cache server and the user terminal, and the second connection is a QUIC connection. Preferably, the first connection determination module specifically includes:

[0036] a data transmission rate calculation unit, configured to calculate the data transmission rate of the first connection;

[0037] a data packet loss rate calculation unit, configured to calculate the data packet loss rate of the first connection;

[0038] The first connection determination unit is configured to determine that the first connection is an abnormal connection when the data transmission rate of the first connection is less than a preset data transmission rate threshold and / or when the data packet loss rate of the first connection is greater than a preset data packet loss rate threshold; otherwise, determine that the first connection is a normal connection.

[0039] Preferably, the second cache server determination module specifically includes:

[0040] A health value calculation unit, used to calculate the health values ​​of other cache servers in the cache server group;

[0041] The second cache server determining unit is configured to determine the second cache server according to the health value.

[0042] Preferably, the health value calculation unit specifically includes:

[0043] A first health value calculation subunit calculates a first health value using the following formula (1);

[0044] Among them, ioutil is the IO usage of the disk in the cache server, max The maximum IO usage of the disk in the cache server, ioutil min is the minimum IO usage of the disk in the cache server, x ioutil The first health value of the cache server;

[0045] The second health value calculation subunit calculates the second health value using the following formula (2);

[0046] Among them, iowait is the percentage of cache server CPU idle but with unfinished disk IO requests in a sampling period. max iowait is the maximum percentage of cache servers with idle CPU but unfinished disk IO requests within a sampling period. min The minimum percentage of cache server CPU idle but unfinished disk IO requests in a sampling period, x iowait The second health value of the cache server;

[0047] The third health value calculation subunit calculates the third health value using the following formula (3);

[0048] in, is the average CPU usage of the cache server, c is an empirical constant, and x cpu The third health value of the cache server;

[0049] The health value calculation subunit calculates the health value using the following formula (4): X = max (x ioutil , x iowait , x cpu ) (4)

[0050] Where X is the health value of the cache server.

[0051] Preferably, the disconnection forming module specifically includes:

[0052] A connection information acquisition unit, configured to acquire an ID password and encryption method for the first connection;

[0053] The connection disconnection forming unit is used to disconnect the first connection and complete the second connection according to the ID password and encryption method. Beneficial effects

[0054] The connection scheduling method and system of the embodiments of this specification can automatically detect and promptly discover abnormalities in the first connection between a first cache server and a user terminal. When an abnormality occurs in the first connection, a second cache server that establishes a second connection with the user terminal can be quickly and effectively found. The second cache server can continue to efficiently and stably transmit data to the user terminal through the second connection. The switching operation between the first and second connections is simple and convenient, and after the first connection is switched to the second connection, the connection itself remains unchanged and will not affect the user terminal's reception of data. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] FIG1 is a schematic diagram of a flow chart of a connection scheduling method for data transmission provided in an embodiment of this specification;

[0056] FIG2 is a schematic diagram of the structure of a connection scheduling system for data transmission provided by an embodiment of this specification. DETAILED DESCRIPTION

[0057] Hereinafter, embodiments of the present specification will be described in more detail with reference to the accompanying drawings.

[0058] Embodiment 1: A connection scheduling method for data transmission, as shown in FIG1 , includes the following steps:

[0059] L1. Obtain a first connection, where the first connection is a connection between a first cache server and a client, and the first connection is a QUIC connection.

[0060] Assume that the cache server group of this embodiment includes five cache servers, namely cache server 1, cache server 2, cache server 3, cache server 4, and cache server 5. Cache server 1 is connected to client 1, cache server 3 is connected to client 2, and the other three cache servers are not connected to any client.

[0061] Then, the first connection acquisition module of cache server No. 1 can acquire the first connection, the first connection acquisition module of cache server No. 3 can also acquire the first connection, and the other three cache servers cannot acquire the first connection.

[0062] L2. Determine whether the first connection is a normal connection or an abnormal connection.

[0063] Because the first connection acquisition module of cache server 1 has acquired the first connection, the first connection determination module of cache server 1 needs to determine whether the corresponding first connection is an abnormal connection. Similarly, because the first connection acquisition module of cache server 3 has acquired the first connection, the first connection determination module of cache server 3 also needs to determine whether the corresponding first connection is an abnormal connection. Since the other three cache servers did not acquire the first connection, they do not need to perform first connection determination.

[0064] In this embodiment, the specific steps of determining whether the first connection is a normal connection or an abnormal connection include:

[0065] L21. Calculate the data transmission rate of the first connection, that is, calculate the data transmission rate between the first cache server and the client.

[0066] Specifically, a certain number (for example, 100) of data transmission rate initial values ​​may be collected at a certain data collection frequency, and then the data transmission rate initial values ​​are grouped (for example, data transmission rate initial values ​​starting with 10, such as 10.0K / s, 10.9K / s, and 10.3K / s, are grouped together; data transmission rate initial values ​​starting with 15, such as 15.0K / s, 15.2K / s, and 15.7K / s, are grouped together; data transmission rate initial values ​​starting with 22, such as 22.1K / s, 22.2K / s, and 22.8K / s, are grouped together). Then, a group of data transmission rate initial values ​​with the largest number of data is obtained. For example, a group of data transmission rate initial values ​​has the largest number of data, including 8.0K / s, 8.9K / s, 8.3K / s, 8.1K / s, 8.2K / s, 8.3K / s, 8.8K / s, 8.9K / s, 8.3K / s, and 8.6K / s. Then, the average value of the initial data transmission rate values ​​is calculated to obtain 8.44 K / s. Finally, the average value of the initial data transmission rate values ​​is used as the data transmission rate of the first connection, that is, in this embodiment, 8.44 K / s is used as the data transmission rate of the first connection.

[0067] L22. Calculate the packet loss rate of the first connection, i.e., calculate the packet loss rate between the first cache server and the client. Data is transmitted in packets across the network. The packet loss rate is the ratio of the number of packets lost during transmission to the number of packets sent. This embodiment directly uses existing techniques to calculate the packet loss rate of the first connection.

[0068] L23. When the data transmission rate of the first connection is less than a preset data transmission rate threshold and / or when the data packet loss rate of the first connection is greater than a preset data packet loss rate threshold, the first connection is determined to be an abnormal connection; otherwise, the first connection is determined to be a normal connection.

[0069] In this embodiment, the preset data transmission rate threshold may be 10K / s, and the preset data packet loss rate threshold may be 10%. As long as the data transmission rate of the first connection is less than 10K / s or the data packet loss rate is greater than 10%, the first connection is determined to be an abnormal connection, which cannot efficiently and stably transmit data. Only when the data transmission rate of the first connection is greater than or equal to 10K / s and the data packet loss rate is less than or equal to 10%, is the first connection determined to be a normal connection.

[0070] Assume in this embodiment that the data transmission rate of the first connection of cache server No. 1 is 6K / s and the data packet loss rate is 11%, then the first connection of cache server No. 1 is determined to be an abnormal connection.

[0071] Assuming in this embodiment that the data transmission rate of the first connection of cache server No. 3 is 54 K / s and the data packet loss rate is 2%, the first connection of cache server No. 3 is determined to be a normal connection. Furthermore, after the first connection of cache server No. 3 is determined to be a normal connection, steps L1 and L2 need to be repeated continuously until the first connection is no longer available (i.e., cache server No. 3 and client 2 are disconnected), or the first connection of cache server No. 3 is determined to be an abnormal connection.

[0072] L3. When the first connection is an abnormal connection, generate a first transition signal.

[0073] When a first connection is determined to be an abnormal connection, the first migration signal generation module of the corresponding cache server generates a first migration signal. For example, in this embodiment, when the first connection determination module of cache server No. 1 determines that the first connection is an abnormal connection, the first migration signal generation module of cache server No. 1 generates a first migration signal and sends the first migration signal to the first migration signal acquisition module of the cache server group. The first migration signal can be a combination signal of a high level, a high level, a low level, and the corresponding cache server number.

[0074] L4. After obtaining the first migration signal, determine a second cache server.

[0075] When the cache server group's first migration signal acquisition module receives the first migration signal sent by cache server #1, it will know that there is a problem with the first connection between cache server #1 and client #1. Cache server #1 cannot efficiently and stably send data to client #1 via the first connection. At this time, it is necessary to find another cache server (i.e., a second cache server) and connect it to client #1 (i.e., form a second connection). The second cache server can then efficiently and stably send data to client #1 via the second connection.

[0076] In this embodiment, determining the second cache server specifically includes the following steps:

[0077] L41. Calculate the health values ​​of other cache servers in the cache server group.

[0078] In this embodiment, each cache server is provided with a health value calculation unit, which calculates the health value of the corresponding cache server using the following formula:

[0079] Among them, ioutil is the IO usage of the disk in the cache server, max The maximum IO usage of the disk in the cache server, ioutil min is the minimum IO usage of the disk in the cache server, x ioutil The first health value of the cache server.

[0080] ioutil max is an empirical constant, for example, it can be 0.9. min It is also an empirical constant, for example, it can be 0.1. Assuming that ioutil in this embodiment is 0.2 (corresponding to cache server No. 2) or 0.5 (corresponding to cache server No. 3) or 0.3 (corresponding to cache server No. 4) or 0.2 (corresponding to cache server No. 5), then correspondingly, x in this embodiment is ioutil (ie, the first health value) is 0.125 (corresponding to cache server No. 2) or 0.5 (corresponding to cache server No. 3) or 0.25 (corresponding to cache server No. 4) or 0.125 (corresponding to cache server No. 5).

[0081] Among them, iowait is the percentage of cache server CPU idle but with unfinished disk IO requests in a sampling period. max iowait is the maximum percentage of cache servers with idle CPU but unfinished disk IO requests within a sampling period. min The minimum percentage of cache server CPU idle but unfinished disk IO requests in a sampling period, xiowait This is the second health value of the cache server.

[0082] iowait max Is an empirical constant, for example, it can be 0.8. min It is also an empirical constant, for example, it can be 0.1. Assuming that iowait in this embodiment is 0.2 (corresponding to cache server No. 2) or 0.6 (corresponding to cache server No. 3) or 0.9 (corresponding to cache server No. 4) or 0.3 (corresponding to cache server No. 5), then correspondingly, x in this embodiment is iowait (ie, the second health value) is 0.143 (corresponding to cache server No. 2) or 0.714 (corresponding to cache server No. 3) or 1 (corresponding to cache server No. 4) or 0.286 (corresponding to cache server No. 5).

[0083] in, is the average CPU usage of the cache server, c is an empirical constant, and x cpu is the third health value of the cache server. Assume c is 0.1, is 0.3 (corresponding to cache server No. 2) or 0.5 (corresponding to cache server No. 3) or 0.4 (corresponding to cache server No. 4) or 0.2 (corresponding to cache server No. 5), then correspondingly, x in this embodiment cpu (ie, the third health value) is 0.4 (corresponding to cache server No. 2) or 0.6 (corresponding to cache server No. 3) or 0.5 (corresponding to cache server No. 4) or 0.3 (corresponding to cache server No. 5).

[0084] X=max(x ioutil , x iowait , x cpu ) (4)

[0085] Where X is the cache server's health value. The cache server's final health value is the maximum of the first, second, and third health values. The larger the cache server's health value, the less healthy it is.

[0086] Since the first health value of cache server 2 is 0.125, the second health value is 0.143, and the third health value is 0.4, the final health value of cache server 2 is 0.4. Since the first health value of cache server 3 is 0.5, the second health value is 0.714, and the third health value is 0.6, the final health value of cache server 3 is 0.714. Since the first health value of cache server 4 is 0.25, the second health value is 1, and the third health value is 0.5, the final health value of cache server 4 is 1. Since the first health value of cache server 5 is 0.125, the second health value is 0.286, and the third health value is 0.3, the final health value of cache server 5 is 0.3.

[0087] L42. Determine a second cache server based on the health value.

[0088] The cache server group can obtain the current health value of each cache server. For example, the current health value of cache server 2 is 0.4, the health value of cache server 3 is 0.714, the health value of cache server 4 is 1, and the health value of cache server 5 is 0.3. The smaller the health value, the healthier the corresponding cache server. Therefore, in this embodiment, cache server 5 is determined to be the second cache server.

[0089] L5. When the second cache server determines that the migration is complete, it generates a second migration signal.

[0090] Once the second cache server is determined, the cache server group generates a second migration signal, indicating that the connection can be migrated. The second migration signal can be a combination of a low level, a high level, and a high level. The cache server group sends the second migration signal to the first cache server (e.g., cache server 1 in this embodiment) and the second cache server (e.g., cache server 5 in this embodiment).

[0091] L6. After obtaining the second migration signal, disconnect the first connection and form a second connection, where the second connection is a connection between the second cache server and the user terminal, and the second connection is a QUIC connection.

[0092] The first connection includes a connection body, a connection port 1 and a connection port 2, wherein the connection port 1 is connected to the user terminal, and the connection port 2 is connected to the first cache server.

[0093] The second connection also includes a connection body, connection port 1, and connection port 2. Connection port 1 is connected to the client, and connection port 2 is connected to the second cache server. The second connection is essentially the same as the first connection, except that connection port 2 of the second connection is connected to the second cache server.

[0094] When the second cache server receives the second migration signal, it uses the connection disconnection module to change the connection port 2 of the first connection to the second cache server. At this point, the second connection is established, and the first connection is automatically disconnected. This means that the same connection initially has its connection port 2 connected to the first cache server, which is called the first connection. Later, when its connection port 2 is changed to the second cache server, it is called the second connection.

[0095] The first connection and the second connection in this embodiment are both QUIC connections. In the prior art, there are TCP connections based on the HTTP protocol and QUIC connections based on the UDP protocol. A TCP connection consists of four elements: source IP, source port, destination IP, and destination port. When one of these elements changes, the TCP connection will be disconnected, and a new TCP connection can only be re-established subsequently. The QUIC connection uses an ID password (a 64-bit random number) to identify the connection. As long as the ID password remains unchanged, even if the user end or the cache server end changes, the connection itself remains unchanged, allowing the newly formed second connection to continue to transmit data for the same data request. For example, the first connection was originally transmitting data for data request A, and 2 / 3 of the data had been transmitted. At this time, it was found that the first connection was abnormal, so it was changed to the second connection. The second connection can continuously transmit the remaining 1 / 3 of the data for data request A.

[0096] Therefore, in step L6, disconnecting the first connection and establishing the second connection specifically includes the following steps:

[0097] L61. Obtain the ID password and encryption method for the first connection. The cache server group simultaneously sends the second migration signal to the first cache server and the second cache server. Upon receiving the second migration signal, the first cache server extracts the ID password and encryption method for the first connection and sends it to the cache server group's connection information forwarding unit. The connection information forwarding unit then forwards the ID password and encryption method to the second cache server. Finally, the second cache server's connection information acquisition unit obtains the ID password and encryption method for the first connection.

[0098] L62. Disconnect the first connection and complete the second connection based on the ID password and encryption method. After the second cache server's connection information acquisition unit acquires the ID password and encryption method of the first connection, the second cache server's connection disconnection establishment unit can identify the first connection based on the ID password and encryption method and change the connection port 2 of the first connection to connect to the second cache server. At this point, the second connection is automatically established and the first connection is automatically disconnected.

[0099] After the second connection is established, the second cache server can continue to provide data to the client. The connection body of the second connection is actually the same as the first connection, and the connection body has not changed, so it will not affect the client's reception of data.

[0100] The connection scheduling method for data transmission of this embodiment can automatically detect and promptly discover abnormalities in the first connection between a first cache server and a user terminal. When an abnormality occurs in the first connection, a second cache server that establishes a second connection with the user terminal can be quickly and effectively found. The second cache server can continue to efficiently and stably transmit data to the user terminal through the second connection. The switching operation between the first and second connections is simple and convenient, and after the first connection is switched to the second connection, the connection itself remains unchanged, without affecting the user terminal's reception of data.

[0101] Example 2: A connection scheduling system for data transmission, used to execute the connection scheduling method described in Example 1, as shown in Figure 2. The connection scheduling system of this embodiment specifically includes: a first connection acquisition module, a first connection determination module, a first migration signal generation module, a first migration signal acquisition module, a second cache server determination module, a second migration signal generation module, a second migration signal acquisition module and a connection disconnection formation module.

[0102] The first connection acquisition module is used to acquire a first connection, wherein the first connection is a connection between a first cache server and a user terminal, and the first connection is a QUIC connection. The first connection determination module is used to determine whether the first connection is a normal connection or an abnormal connection. The first migration signal generation module is used to generate a first migration signal when the first connection is an abnormal connection; the first migration signal acquisition module is used to acquire the first migration signal. The second cache server determination module is used to determine the second cache server after acquiring the first migration signal. The second migration signal generation module is used to generate a second migration signal when the second cache server is determined to be completed. The second migration signal acquisition module is used to acquire the second migration signal. The connection disconnection formation module is used to disconnect the first connection and form a second connection after acquiring the second migration signal, wherein the second connection is a connection between the second cache server and the user terminal, and the second connection is a QUIC connection.

[0103] The first connection determination module specifically includes: a data transmission rate calculation unit, a data packet loss rate calculation unit, and a first connection determination unit. The data transmission rate calculation unit is configured to calculate the data transmission rate of the first connection. The data packet loss rate calculation unit is configured to calculate the data packet loss rate of the first connection. The first connection determination unit is configured to determine that the first connection is an abnormal connection when the data transmission rate of the first connection is less than a preset data transmission rate threshold and / or when the data packet loss rate of the first connection is greater than a preset data packet loss rate threshold; otherwise, determine that the first connection is a normal connection.

[0104] The second cache server determination module specifically includes: a health value calculation unit and a second cache server determination unit. The health value calculation unit is configured to calculate the health values ​​of other cache servers in the cache server group. The second cache server determination unit is configured to determine the second cache server based on the health values.

[0105] The health value calculation unit specifically includes: a first health value calculation subunit, a second health value calculation subunit, a third health value calculation subunit and a health value calculation subunit.

[0106] The first health value calculation subunit calculates the first health value using the following formula (1).

[0107] Among them, ioutil is the IO usage of the disk in the cache server, max The maximum IO usage of the disk in the cache server, ioutil min is the minimum IO usage of the disk in the cache server, x ioutil The first health value of the cache server.

[0108] The second health value calculation subunit calculates the second health value using the following formula (2).

[0109] Among them, iowait is the percentage of cache server CPU idle but with unfinished disk IO requests in a sampling period. max iowait is the maximum percentage of cache servers with idle CPU but unfinished disk IO requests within a sampling period. min The minimum percentage of cache server CPU idle but unfinished disk IO requests in a sampling period, x iowait This is the second health value of the cache server.

[0110] The third health value calculation subunit calculates the third health value using the following formula (3).

[0111] in, is the average CPU usage of the cache server, c is an empirical constant, and x cpu The third health value of the cache server.

[0112] The health value calculation subunit calculates the health value using the following formula (4): X = max (x ioutil , x iowait , x cpu ) (4)

[0113] Where X is the health value of the cache server.

[0114] The disconnection formation module specifically includes: a connection information acquisition unit and a disconnection formation unit. The connection information acquisition unit is used to obtain the ID password and encryption method of the first connection. The disconnection formation unit is used to disconnect the first connection and complete the second connection based on the ID password and encryption method.

[0115] The connection scheduling system for data transmission of this embodiment can automatically detect and promptly discover abnormalities in the first connection between a first cache server and a user terminal. When an abnormality occurs in the first connection, a second cache server that establishes a second connection with the user terminal can be quickly and effectively found. The second cache server can continue to efficiently and stably transmit data to the user terminal through the second connection. The switching operation between the first and second connections is simple and convenient, and after the first connection is switched to the second connection, the connection itself remains unchanged, without affecting the user terminal's reception of data.

[0116] Although certain embodiments of the present specification are shown in the accompanying drawings, it should be understood that the present specification may be implemented in various forms and should not be construed as limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the present specification. It should be understood that the accompanying drawings and embodiments of the present specification are for illustrative purposes only and are not intended to limit the scope of protection of the present specification.

Claims

1. A connection scheduling method for data transmission, comprising the following steps: Acquire a first connection, where the first connection is a connection between a first cache server and a client, and the first connection is a QUIC connection; determining whether the first connection is a normal connection or an abnormal connection; When the first connection is an abnormal connection, generating a first migration signal; After obtaining the first migration signal, determining a second cache server; When the second cache server determines that the migration is completed, generating a second migration signal; After obtaining the second migration signal, the first connection is disconnected and a second connection is formed, where the second connection is a connection between the second cache server and the user terminal, and the second connection is a QUIC connection.

2. According to the connection scheduling method of claim 1, determining whether the first connection is a normal connection or an abnormal connection specifically comprises the following steps: calculating a data transmission rate of the first connection; Calculating a data packet loss rate of the first connection; When the data transmission rate of the first connection is lower than a preset data transmission rate threshold and / or when the data packet loss rate of the first connection is higher than a preset data packet loss rate threshold, the first connection is determined to be an abnormal connection; otherwise, the first connection is determined to be a normal connection.

3. According to the connection scheduling method of claim 1, determining the second cache server specifically comprises the following steps: Calculate the health values ​​of other cache servers in the cache server group; A second cache server is determined according to the health value.

4. According to the connection scheduling method of claim 3, the health value of each cache server is calculated by the following formula: in, ioutil is the IO usage of the disk in the cache server. max The maximum IO usage of the disk in the cache server, ioutil min is the minimum IO usage of the disk in the cache server, x ioutil is the first health value of the cache server; Among them, iowait is the percentage of cache servers with idle CPU but unfinished disk IO requests in a sampling period. max The maximum proportion of the cache server's CPU being idle but with unfinished disk IO requests within a sampling period, iowait min is the minimum proportion of the cache server's CPU being idle but with unfinished disk IO requests within a sampling period, x iowait is the second health value of the cache server; in, is the average CPU usage of the cache server, c is an empirical constant, and x cpu is the third health value of the cache server; X=max(x ioutil ,x iowait ,x cpu ) (4) Where X is the health value of the cache server.

5. According to the connection scheduling method of claim 1, disconnecting the first connection and establishing the second connection specifically comprises the following steps: Obtain the ID password and encryption method of the first connection; The first connection is disconnected and the second connection is completed according to the ID password and encryption method.

6. A connection scheduling system for data transmission, comprising: A first connection acquisition module, used to acquire a first connection, where the first connection is a connection between a first cache server and a client, and the first connection is a QUIC connection; A first connection determination module, used to determine whether the first connection is a normal connection or an abnormal connection; A first migration signal generating module, configured to generate a first migration signal when the first connection is an abnormal connection; A first migration signal acquisition module, used to acquire the first migration signal; A second cache server determining module, configured to determine a second cache server after acquiring the first migration signal; A second migration signal generating module, configured to generate a second migration signal when the second cache server determines that the migration is completed; A second migration signal acquisition module, used for acquiring the second migration signal; The connection disconnection forming module is used to disconnect the first connection and form a second connection after obtaining the second migration signal, wherein the second connection is a connection between the second cache server and the user terminal, and the second connection is a QUIC connection.

7. The connection scheduling system according to claim 6, wherein the first connection determination module specifically comprises: a data transmission rate calculation unit, configured to calculate a data transmission rate of the first connection; a data packet loss rate calculation unit, configured to calculate the data packet loss rate of the first connection; The first connection determination unit is configured to determine that the first connection is an abnormal connection when the data transmission rate of the first connection is less than a preset data transmission rate threshold and / or when the data packet loss rate of the first connection is greater than a preset data packet loss rate threshold; otherwise, determine that the first connection is a normal connection.

8. The connection scheduling system according to claim 6, wherein the second cache server determination module specifically comprises: A health value calculation unit, used to calculate the health values ​​of other cache servers in the cache server group; The second cache server determining unit is configured to determine the second cache server according to the health value.

9. The connection scheduling system according to claim 8, wherein the health value calculation unit specifically comprises: A first health value calculation subunit, which calculates a first health value by the following formula (1); Among them, ioutil is the IO usage of the disk in the cache server. max The maximum IO usage of the disk in the cache server, ioutil min is the minimum IO usage of the disk in the cache server, x ioutil is the first health value of the cache server; The second health value calculation subunit calculates the second health value by the following formula (2); Among them, iowait is the percentage of cache servers with idle CPU but unfinished disk IO requests in a sampling period. max The maximum proportion of the cache server's CPU being idle but with unfinished disk IO requests within a sampling period, iowait min is the minimum proportion of the cache server's CPU being idle but with unfinished disk IO requests within a sampling period, x iowait is the second health value of the cache server; A third health value calculation subunit calculates a third health value using the following formula (3); in, is the average CPU usage of the cache server, c is an empirical constant, and x cpu is the third health value of the cache server; The health value calculation subunit calculates the health value using the following formula (4); X=max(x ioutil ,x iowait ,x cpu ) (4) Where X is the health value of the cache server.

10. The connection scheduling system according to claim 6, wherein the connection disconnection forming module specifically comprises: A connection information acquisition unit, used to acquire an ID password and an encryption method for the first connection; The connection disconnection forming unit is used to disconnect the first connection and complete the second connection according to the ID password and encryption method.

Citation Information

Patent Citations

  • Data transmission method and device, electronic equipment and storage medium

    CN115134553A

  • Load balancing system and data packet distribution method

    CN115225644A

  • Connection scheduling method and system for data transmission

    CN116366709A

  • Method of QUIC communication via multiple paths

    US20200120015A1

  • Methods and apparatuses for enabling multi-host multipath secure transport with quic

    US20230074838A1