WO2017107577A1

WO2017107577A1 - Node probing method and device, path selection method and device, and network system

Info

Publication number: WO2017107577A1
Application number: PCT/CN2016/098651
Authority: WO
Inventors: 洪珂; 程豪; 谢玉燕
Original assignee: 网宿科技股份有限公司
Priority date: 2015-12-25
Filing date: 2016-09-11
Publication date: 2017-06-29
Also published as: CN105515915A; CN105515915B

Abstract

Provided in the present invention are a node probing method and device, a path selection method and device, and a network system. The method comprises: instructing an acceleration node in a network to measure the network delay and the packet loss rate of transmissions from the acceleration node to other acceleration nodes in the network and to a source station at a first preset cycle; sending the network delay and the packet loss rate to a server for the server to compute an optimum path from the acceleration node in the network to the source station, and acquiring the optimum path computed by the server at a second preset cycle and saving the acquired optimum path locally, to ensure that when accessing the corresponding source station via the acceleration node in the network, a client can select the optimum path for the access at least using the network delay and the packet loss rate in combination as selection parameters for the optimum path, so that both transmission time and stability of the selected optimum path can better meet user requirements.

Description

Node detection method and device, path selection method and device, and network system

Technical field

The present invention relates to a network communication system, and in particular, to a node detection method and apparatus, a path selection method and apparatus, and a network system.

Background technique

The application acceleration network is a network data acceleration system. It includes a series of acceleration nodes deployed by different operators in different geographical locations. Users of different areas access different source stations, and select some acceleration nodes as data transmission paths. Accelerate data transfer by selecting a path with a small network latency.

With the continuous increase of network bandwidth and the continuous development of Internet applications, there are more and more network applications such as live video, online education, remote desktop, and massive data transmission. These applications have the following characteristics:

(1) It is necessary to maintain a long transmission time: for example, an online course takes 45 minutes or more, and a 1T file needs to be transmitted in a normal network environment for several hours;

(2) Transmission distance is often long, and transmission stability is poor: for example, multinational enterprises transmit data, or foreign foreign teachers teach domestic students, they need to carry out transnational data transmission, and the transmission stability of multinational networks is often low;

(3) The transmission stability requirement is high. The video transmission interruption needs to re-learn or refresh the page, which affects the user experience. The file transmission may need to be transmitted from the beginning when the breakpoint retransmission is not supported, and the remote desktop may interrupt the ongoing operation. .

Then, the above-mentioned application acceleration network routing method can not meet the requirements of the above application, because the measurement factor for measuring the distance between two nodes in the application acceleration network is the delay time between them, and it cannot measure two nodes. The stability of the connection between the best paths selected in this way may be the fastest, but may not be the most stable, and cannot meet the requirements when the transmission time and transmission distance are long.

Summary of the invention

In view of the above-mentioned shortcomings of the prior art, an object of the present invention is to provide a node detecting method and device, a path selecting method and device, and a network system, which are used to solve the problem that the corresponding source station cannot be effectively selected in the prior art. The problem of the optimal path.

To achieve the above and other related objects, the present invention provides a node detection method, which is applied to a network acceleration node, including: detecting a network delay and a packet loss rate of a detected node and a source station by using a first preset period. Transmitting the network delay and the packet loss rate to a server, where the server calculates an optimal path of the network acceleration node to the source station; acquiring the server calculation by a second preset period The optimal path obtained and saved locally for The client of the network acceleration node communication selects the optimal path to access the source station.

In a specific embodiment of the present invention, the step of detecting the packet loss rate of the network acceleration node to the detected node and the source station by using the first preset period comprises: sending the detected node to the detected node by using the first preset period a specific data, and obtaining the number of transmitted bytes and the number of retransmitted bytes to calculate a packet loss rate between the network acceleration node and the detected node; or according to a client to accelerate the node to the corresponding Detecting data sent by the node to calculate a packet loss rate between the network acceleration node and the corresponding detected node; and calculating data according to data sent by the client to the source station by using the network acceleration node Accelerating the packet loss rate between the node and the source station.

In a specific embodiment of the present invention, the weight loss calculation between a detected node or a source station detected by detecting a preset time period is performed as the network acceleration node and the corresponding detected node or The packet loss rate between source stations.

In an embodiment of the present invention, the N packet loss ratios of a detected node or a source station are detected during the preset time period, and the weighting calculation method for the N packet loss rates is:

L=L _avg +L _var where,

And L _i represents the i-th packet loss rate, and when the network node and the source node or the detected station can not communicate, L _i is -1, l _'i is the corrected value of L _i, if L _i ≠-1, then l′ _i =l _i , when l _i =-1, l′ _i =l _avg *5, where l _avg is the packet loss rate with the _median of the N packet loss rates being -1 The arithmetic mean after.

In a specific embodiment of the present invention, the packet loss ratios obtained by weighting calculations in multiple time periods are separately obtained, and the weights of the packet loss rates in the corresponding time segments are respectively set, and according to the packet loss rate and Corresponding weights are calculated, and the weighted packet loss rate is calculated, so that the server calculates an optimal path of the network node to the source station according to the network delay and the weighted packet loss rate.

In an embodiment of the present invention, the method further includes: counting the number of active connections of the network acceleration node by using the first preset period, and transmitting, to determine whether the network acceleration node is selected to access the corresponding source station. .

In a specific embodiment of the present invention, the method further includes: collecting, by using the first preset period, a bandwidth usage rate of the network acceleration node, and transmitting, for sending the network acceleration node to the source station. The calculation of the excellent path is corrected.

In order to achieve the above object and other related objects, the present invention further provides a node detecting apparatus, which is configured to detect a node by using the node detecting method according to any one of the preceding claims, and at least includes: a detecting module, configured to use a first preset The network detects the network delay and the packet loss rate of the detected node and the source station; the sending module is configured to send the network delay and the packet loss rate to a server, where the server calculates the network acceleration An optimal path from the node to the source station; the acquiring module is configured to acquire the optimal path calculated by the server in a second preset period, and save the local path for local selection for communication with the network acceleration node The optimal path accesses the source station.

To achieve the above and other related objects, the present invention also provides a network acceleration node, including the above described section Point detection device.

To achieve the above and other related objects, the present invention also provides a path selection method, which is applied to an acceleration network having a plurality of network acceleration nodes as described above, for a client to access a corresponding source through the acceleration network. The station includes: receiving network delays and packet loss rates sent by the network acceleration nodes; and selecting, according to the received network delay and packet loss rate, each of the network nodes to the The optimal path of the source station.

In a specific embodiment of the present invention, the preset selection manner is one of the following: manner 1) calculating a path delay and a path loss ratio of all paths of each of the acceleration nodes to the source station, Selecting a preset number of to-be-selected paths in the order of the path delay from high to low, and selecting the path with the lowest path loss rate in the candidate path as the optimal speed of the corresponding network acceleration node to the source station Path 2) calculating a path delay and a path loss rate of all paths of each of the acceleration nodes to the source station, and selecting a preset number of candidate to be selected in descending order of the path loss rate a path, and selecting a path with the lowest path delay in the to-be-selected path as an optimal path from the corresponding network acceleration node to the source station; and (3) selecting each of the network acceleration nodes to other network acceleration nodes and the source station The network delay and the packet loss rate are weighted, and an optimal path of each of the network nodes to the source station is selected according to the weighting calculation result.

In a specific embodiment of the present invention, for the mode 1) and the second mode, for a path including N network acceleration nodes, the path has a N-1 segment, and the corresponding path loss rate calculation formula is used. for:

The l _i is the packet loss rate of the i th segment.

In a specific embodiment of the present invention, for the mode 3), the formula for weighting the network delay and the packet loss rate of each of the network acceleration nodes to other network acceleration nodes and the source station is:

The RTT indicates the network delay of the current network acceleration node to a other network acceleration node or source station, and L indicates the packet loss rate of the current network acceleration node to a other network acceleration node or source station, and f(RTT, L) indicates the front The weighting value of the network delay and packet loss rate of the network acceleration node to a network acceleration node or source station is weighted, and λ is a weighting factor, which can be adjusted according to actual conditions; wherein f(RTT, L) is equal to -1 , indicating that the path is unreachable.

In a specific embodiment of the present invention, the calculation of the optimal path of the network acceleration node to the source station is modified according to the following formula:

BWR1 and BWR2 respectively represent the bandwidth usage rate of two endpoints of a connection, and when the connection fails, the calculation result is -1.

In an embodiment of the present invention, the method further receives the number of active connections sent by the network acceleration node, and compares the number of active connections with a threshold of a first connection number and a second connection number, and When the number of active connections is greater than the threshold of the first connection, the network acceleration node is removed from the network acceleration network, and the number of active connections of the network acceleration node is continuously monitored, and the number of active connections is smaller than the number of active connections. When the number of connections is thresholded, the network acceleration node is added to the acceleration network.

In order to achieve the above and other related objects, the present invention further provides a path selection apparatus, which uses the path selection method according to any one of the preceding items to select an optimal path, and at least includes: a receiving module, configured to receive a plurality of the network accelerations a network delay and a packet loss rate sent by the node; a selection module, configured to select an optimal path of each of the network nodes to the source station by using a preset selection manner according to the received network delay and packet loss rate .

To achieve the above and other related objects, the present invention also provides a server comprising the path selection device as described above.

To achieve the above and other related objects, the present invention also provides a network system including a plurality of network acceleration nodes as described above, and a server as described above for enabling a client to pass the network system to optimize The path accesses the corresponding source station.

As described above, the node detecting method and device, the path selecting method and device, and the network system of the present invention enable the network acceleration node to detect network delays of other acceleration nodes and source stations in the network in a first preset period. And the packet loss rate; the network delay and the packet loss rate are sent to a server, so that the server calculates an optimal path of the network acceleration node to the source station, and obtains the second preset period. Describe the optimal path calculated by the server and save it to the local device to ensure that a client accesses the corresponding source station through the network acceleration node, and then selects the optimal path for access, and at least combines network delay and packet loss. The rate is used as the selection parameter of the optimal path, so that the transmission time and stability of the selected optimal path can better meet the user's needs.

DRAWINGS

FIG. 1 is a flow chart showing a node detecting method of the present invention in a specific embodiment.

2 is a flow chart showing a node detecting method of the present invention in a specific embodiment.

FIG. 3 is a flow chart showing a path selection method of the present invention in a specific embodiment.

4 is a block diagram showing a node detecting device and a path selecting device of the present invention in a specific embodiment.

FIG. 5 shows a block diagram of a network acceleration system of the present invention in a specific embodiment.

Component label description

A node detection device

A1 detection module

A2 send module

A3 acquisition module

B path selection device

B1 receiving module

B2 selection module

1 network acceleration system

11 network acceleration node

12 server

2 client

3 source station

S11~S22 steps

detailed description

The embodiments of the present invention are described below by way of specific examples, and those skilled in the art can readily understand other advantages and effects of the present invention from the disclosure of the present disclosure. The present invention may be embodied or applied in various other specific embodiments, and various modifications and changes can be made without departing from the spirit and scope of the invention. It should be noted that the features in the following embodiments and embodiments may be combined with each other without conflict.

It should be noted that the illustrations provided in the following embodiments merely illustrate the basic concept of the present invention in a schematic manner, and only the components related to the present invention are shown in the drawings, rather than the number and shape of components in actual implementation. Dimensional drawing, the actual type of implementation of each component's type, number and proportion can be a random change, and its component layout can be more complicated.

The application acceleration network is a network data acceleration system. It includes a series of acceleration nodes deployed by different operators in different geographical locations. Users of different areas access different source stations, and select some acceleration nodes as data transmission paths. Accelerate data transfer by selecting a path with a small network latency. In one embodiment, the manner in which the user selects the corresponding acceleration node is selected by the geographic location of the user and the operator. For example, the telecommunications user in Guangzhou selects the acceleration node of the telecommunications set in Guangzhou to access the corresponding source station.

In a specific embodiment, the routing process of the application acceleration network mainly includes the following steps:

(1) Uninterrupted detection of each other in all accelerating nodes, and all accelerating nodes detect the accelerating source station, The probe uses the way to establish a TCP connection, and the time to establish a TCP connection is taken as the distance factor between the two nodes.

(2) The application acceleration network includes a central server, and the acceleration node periodically uploads the detection results of the other acceleration nodes to the central server, and the central server periodically according to the distance factor between all nodes and the node to the source station (ie, the The time to establish a TCP connection), calculate the optimal path from any acceleration node to any source station.

(3) The node periodically obtains the optimal path of the node to all the acceleration source stations to the central server, and saves it locally. When accessed by the user through the node, the node selects the optimal path to the source station accessed by the user for data transmission.

However, in this embodiment, only the delay problem of data transmission is considered, and the problem of stability of data transmission between nodes is neglected. Therefore, the acceleration path selected by this embodiment may be the fastest, but not necessarily The most stable, the transmission time and transmission distance can not meet the corresponding requirements. Therefore, in the present invention, a reference factor for adding stability in the process of selecting an optimal path is also proposed.

Referring to FIG. 1, a flow chart of a node detecting method according to the present invention is shown in a specific embodiment. The network node detection method is applied to a network acceleration node, including:

S11: Detecting the network delay and the packet loss rate of the detected node and the source station by using a first preset period; ensuring the acquired network delay and the timeliness of the packet loss rate. The first preset period can be set according to the needs of the customer.

S12: Send the network delay and the packet loss rate to calculate an optimal path of the network acceleration node to the source station.

S13: Acquire the optimal path in a second preset period, and save it locally, so that the client that communicates with the network acceleration node selects the optimal path to access the source station, and ensures the obtained optimal path. Timeliness.

In a specific embodiment of the present invention, the step of detecting the network delay of the network acceleration node to the detected node and the source station by using a first preset period comprises: the network acceleration node being in the first preset period and The detected node and the source station establish a TCP connection, and acquire a network delay between the network acceleration node and the detected node and the source station according to a time when the TCP connection is established.

Referring to FIG. 2, when the network acceleration node initiates a TCP connection with a detected node or a source station, the time recorded at this time is the start time Tstart, and when the connection fails, the corresponding network delay time is set to -1 (indicating that the connection cannot be made), and setting the packet loss rate to 100%; and when the two connections are successful, the recording time is the end time Tend, that is, the network delay between the two is the end time minus At the beginning time, and sending a data to the corresponding detected node or source station, the system interface of the network acceleration node is called to obtain the data packet loss rate (retransmission rate).

In a specific embodiment of the present invention, the step of detecting the packet loss rate of the network acceleration node to the detected node and the source station by using the second preset period comprises: sending the detected node to the detected node by using the second preset period a specific data, and obtaining the number of transmitted bytes and the number of retransmitted bytes to calculate a packet loss rate between the network acceleration node and the detected node, wherein the special The fixed data is, for example, a specific piece of data stored in advance, or a specific data generated in real time according to a program, and the data of the specific data is relatively small to alleviate the network pressure of the packet loss rate detection, and the detected node receives the received data. The processing of the particular data includes discarding or returning an acknowledgment. Or in another specific embodiment, according to data that the client actually sends to the corresponding detected node through the network acceleration node, to calculate a packet loss rate between the network acceleration node and the corresponding detected node, Using the data sent by the user to the corresponding detected node to detect the packet loss rate between the network acceleration node and the corresponding probe node can replace a part of the active detection, which can reduce the bandwidth usage and reduce the processing load of the detected node. In a specific application, the client is, for example, an intelligent data processing device such as a mobile phone or a computer.

Since the source station cannot correctly process the data of the non-user, the source packet cannot be detected by the detection method of sending the specific data. Therefore, we use the packet loss rate of the actual data (that is, the data actually sent by the user) as the data packet loss rate from the node to the source station. However, in the case that there is no data actually sent by the user, 0 may be used as the data packet loss rate between the node and the source station, that is, the node to the source station is not lost.

In a specific embodiment of the present invention, the weight loss calculation between a detected node or a source station detected by detecting a preset time period is performed as the network acceleration node and the corresponding detected node or The packet loss rate between source stations. To more accurately reflect the stability of the data transmission of the segment network in the past period of time.

Preferably, in the preset time period, the N packet loss ratios of a detected node or a source station are detected, and the formula for weighting the N packet loss rates is:

L=L _avg +L _var , where

And L _i represents the i-th packet loss rate, and when the network node and the source node or the detected station can not communicate, L _i is -1, l _'i is the corrected value of L _i, if L _i ≠ -1, then the l _{_'i} = l _i, or _{_{l' i = l avg * 5}} , wherein the arithmetic l _avg removing the N -1 packet loss rate in the average packet loss rate.

Further, the packet loss ratios obtained by weighting calculations in multiple time segments are respectively obtained, and the weights of the packet loss ratios in the corresponding time segments are respectively set, and the weighted loss is calculated according to the packet loss rate and the corresponding weights. a packet rate, for the server to calculate an optimal path of the network node to the source station according to the network delay and a weighted packet loss rate, thereby reflecting a user's preference for stability requirements, for example, three identical The packet loss rate calculated by the weighting in the time period is calculated, and is based on the current time, and the three time periods are divided into the first time period, the second time period, and the first time from the near to the far time. In the three time period, the weight of the first time period is the highest, the second time period is second, and the weight of the third time period is the lowest.

In a specific embodiment of the present invention, the network acceleration node further detects the number of active connections of the network acceleration node by using the first preset period, and sends the information, for example, to a server for the The server determines whether the network acceleration node is selected to access the corresponding source station. Specifically, the server sets the number of active connections and the first connection threshold And comparing the value with the second connection threshold, and when the number of active connections is greater than the first connection threshold, removing the network acceleration node from the network acceleration network, and continuing to listen to the network acceleration node The number of active connections, and when the number of active connections is less than the second connection threshold, the network acceleration node is added to the acceleration network. Wherein the first connection number threshold is greater than the second connection number threshold, and the two thresholds are set to prevent a sudden change of a state in which the network node is added to or removed from the corresponding network, and the first connection number threshold and the number When the number of connections is between the thresholds, the state of the network acceleration node is maintained. The addition of the number of active connections prevents the network acceleration nodes with excessive active connections from accessing the corresponding source stations, that is, preventing the network from accelerating the overload of the node ports.

In a specific embodiment of the present invention, the method further includes: detecting, by using the first preset period, a bandwidth usage rate of the network acceleration node, and transmitting, for (for example, a server) to accelerate the node to the network The calculation of the optimal path of the source station is corrected. It is preferred to use network acceleration nodes with lower bandwidth usage to improve the stability and efficiency of network operation and prevent individual network acceleration nodes from overloading.

Please refer to FIG. 3, which is a flow chart showing the path selection method of the present invention in a specific embodiment.

The path selection method is applied to an acceleration network having a plurality of network acceleration nodes as described above, for example, in an acceleration network, for the client to access the corresponding source station through the acceleration network, including:

S21: Receive network delays and packet loss rates sent by multiple network acceleration nodes.

S22: Select, according to the received network delay and the packet loss rate, an optimal path of each of the network nodes to the source station by using a preset selection manner.

For the mode 1) and the mode 2), for the path including the N network acceleration nodes, the path has a N-1 segment, and the corresponding path loss rate calculation formula is:

The l _i is the packet loss rate of the i-th segment, that is, the average value of the packet loss rate of the N-1 segment on the path is calculated, and the average packet loss rate is used as the packet loss rate of the path. The specific derivation process is: for a segment of a path, the packet loss rate refers to the ratio of the amount of retransmitted data and the amount of transmitted data in the process of transmitting data. For a path composed of multiple segments, the amount of data to be sent is S. from the first node to retransmit the data amount of the second node S * l _1, is transmitted using a TCP protocol, the amount of data transmitted to the third node of the second node is still S, so that the second The amount of retransmitted data from the node to the third node is S*l2, and so on, and the amount of retransmitted data from the kth (0<k<N) nodes to the k+1th node is S*l _k . For the entire path, the ratio of the amount of retransmitted data to the amount of transmitted data is

Is the retransmission rate (packet loss rate) of the entire path.

In a specific embodiment of the invention, the value of λ is 0.6.

In an embodiment of the present invention, the method further includes: receiving a bandwidth usage rate sent by the network acceleration node, and performing calculation on an optimal path of the network acceleration node to the source station according to the bandwidth usage rate. The correction is that the network acceleration node has a bandwidth usage rate of 0% when no data is transmitted, and the bandwidth utilization rate when the network acceleration node is occupied by 100%.

In a specific embodiment of the invention, according to

The following formula modifies the calculation of the optimal path of the network acceleration node to the source station:

BWR1 and BWR2 respectively represent the bandwidth usage rate of two endpoints of a connection, and when the connection fails, the calculation result is -1. It is preferred to use network acceleration nodes with lower bandwidth usage to improve the stability and efficiency of network operation and prevent individual network acceleration nodes from overloading.

In an embodiment of the present invention, the method further receives the number of active connections sent by the network acceleration node, and compares the number of active connections with a threshold of a first connection number and a second connection number, and When the number of active connections is greater than the threshold of the first connection, the network acceleration node is removed from the network acceleration network, and the number of active connections of the network acceleration node is continuously monitored, and the number of active connections is smaller than the number of active connections. When the number of connections is thresholded, the network acceleration node is added Into the acceleration network. Specifically, the server compares the number of active connections with a first connection number threshold and a second connection number threshold, and when the number of active connections is greater than the first connection number threshold, the network acceleration node Removing from the network acceleration network, and continuing to monitor the number of active connections of the network acceleration node, and adding the network acceleration node to the acceleration network when the number of active connections is less than the second connection threshold. Wherein the first connection number threshold is greater than the second connection number threshold, and the two thresholds are set to prevent a sudden change of a state in which the network node is added to or removed from the corresponding network, and the first connection number threshold and the number When the number of connections is between the thresholds, the state of the network acceleration node is maintained. The addition of the number of active connections prevents the network acceleration nodes with excessive active connections from accessing the corresponding source stations, that is, preventing the network from accelerating the overload of the node ports.

In a specific embodiment of the present invention, a specific reference is made to FIG. 4, which is a schematic diagram of a node detecting device according to an embodiment of the present invention. The node detecting device A applies a node detecting method as shown in FIG. 1 to perform node detection. The detecting module A1 is configured to detect the network delay and the packet loss rate of the detected node and the source station by using a first preset period, and the sending module A2 is configured to delay the network and the packet loss rate. Sending to a server for the server to calculate an optimal path of the network acceleration node to the source station; acquiring module A3, configured to acquire an optimal path calculated by the server in a second preset period, and save Locally, the client for communicating with the network acceleration node selects the optimal path to access the source station.

The technical solution of the node detecting device A corresponds to the technical solution of the node detecting method, that is, all technical descriptions about the node detecting method can be applied to the detecting device of the node, and details are not described herein.

Referring to FIG. 4 again, the present invention further provides a path selection device B, which uses the path selection method shown in FIG. 3 to select an optimal path, and includes: a receiving module B1, configured to receive a plurality of nodes of the network acceleration node application. The network delay and the packet loss rate sent by the sending module A1 of the detecting module A; the selecting module B2 is configured to select each of the network nodes according to the received network delay and the packet loss rate by using a preset selection manner. Describe the optimal path of the source station. The technical solution of the path selection device B corresponds to the path selection method, and all the descriptions of the path selection method can be applied to the embodiment, which is not described herein.

In another embodiment, a network system as shown in FIG. 5 is further provided. In this embodiment, the network system is preferably a network acceleration system 1 for the user to accelerate through the network acceleration system 1. Access to the corresponding source station improves the efficiency of the network system operation and improves the user's access experience. The network acceleration system 1 includes a plurality of network acceleration nodes 11 and a server 12, wherein the network acceleration node 11 includes a node detection device A as shown in FIG. The server 12 includes a path selection device B as shown in FIG. 4 to select an optimal path (shortest path) for the corresponding network acceleration node 11 to access the corresponding source station 3. The client 2 may access according to the optimal path stored in the network acceleration node 11 when selecting a network acceleration node 11 to access the corresponding source station 3 to optimize the operation of the network system 1. OK, improve the user experience. In a specific application, the client 2 is, for example, an intelligent data processing device such as a mobile phone or a computer.

In summary, the node detecting method and device, the path selecting method and device, and the network system of the present invention enable the network acceleration node to detect the network of other acceleration nodes and the source station in the network in a first preset period. Deferring and packet loss rate; sending the network delay and the packet loss rate to a server, so that the server calculates an optimal path of the network acceleration node to the source station, and obtains the second preset period The optimal path calculated by the server is saved to the local device to ensure that when the client accesses the corresponding source station through the network acceleration node, the optimal path can be selected for access, and at least combined with network delay and loss The packet rate is used as the selection parameter of the optimal path, so that the transmission time and stability of the selected optimal path can better meet the user's needs. Therefore, the present invention effectively overcomes various shortcomings in the prior art and has high industrial utilization value.

The above-described embodiments are merely illustrative of the principles of the invention and its effects, and are not intended to limit the invention. Modifications or variations of the above-described embodiments may be made by those skilled in the art without departing from the spirit and scope of the invention. Therefore, all equivalent modifications or changes made by those skilled in the art without departing from the spirit and scope of the invention will be covered by the appended claims.

Claims

A node detection method is characterized in that it is applied to a network acceleration node, including:

Detecting the network delay and the packet loss rate of the detected node and the source station in a first preset period;

And sending the network delay and the packet loss rate to a server, where the server calculates an optimal path of the network acceleration node to the source station;

Acquiring the optimal path calculated by the server in a second preset period, and saving the local path, so that the client that communicates with the network acceleration node selects the optimal path to access the source station.
The node detecting method according to claim 1, wherein the step of detecting the packet loss rate of the network acceleration node to the detected node and the source station by using the first preset period comprises:

Transmitting, by the first preset period, a specific data to the detected node, and acquiring the number of transmitted bytes and the number of retransmitted bytes to calculate a packet loss rate between the network acceleration node and the detected node; or according to A client accelerates data sent by the node to the corresponding detected node through the network to calculate a packet loss rate between the network acceleration node and the corresponding detected node;

And calculating, according to data sent by the client to the source station by the network acceleration node, a packet loss rate between the network acceleration node and the source station.
The node detecting method according to claim 1, wherein weighting calculation is performed according to a packet loss rate detected between a detected node or a source station for detecting a preset time period, and the network acceleration node is correspondingly The rate of packet loss between the detected nodes or source stations.
The node detecting method according to claim 3, wherein: detecting N packet loss rates with a detected node or a source station during the preset time period, and weighting the N packet loss rates The calculated formula is:

L=L avg +L var where,
And L i represents the i-th packet loss rate, and when the network node and the source node or the detected station can not communicate, L i is -1, l 'i is the corrected value of L i, if L i ≠-1, then l′ i =l i , when l i =-1, l′ i =l avg *5, where l avg is the packet loss rate with the median of the N packet loss rates being -1 The arithmetic mean after.
The method for detecting a node according to claim 4, wherein the packet loss rate obtained by weighting in a plurality of time periods is separately acquired, and the weight of the packet loss rate in the corresponding time period is respectively set, and Determining a packet loss rate and a corresponding weight, and calculating a weighted packet loss rate, for the server to calculate the network according to the network delay and the weighted packet loss rate The optimal path from the node to the source station.
The node detection method according to claim 1, further comprising: counting the number of active connections of the network acceleration node by using the first preset period, and transmitting, for determining whether to select the network acceleration node. Access the appropriate source station.
The node detection method according to claim 1, further comprising: counting, by the first preset period, a bandwidth usage rate of the network acceleration node, and transmitting, for the network acceleration node to The calculation of the optimal path of the source station is corrected.
A node detecting apparatus, wherein the detecting of a node by using the node detecting method according to any one of claims 1 to 7 includes at least:

a detecting module, configured to detect a network delay and a packet loss rate of the detected node and the source station by using a first preset period;

a sending module, configured to send the network delay and the packet loss rate to a server, where the server calculates an optimal path of the network acceleration node to the source station;

And an acquiring module, configured to acquire the optimal path calculated by the server in a second preset period, and save the local path, so that the client that communicates with the network acceleration node selects the optimal path to access the source station.
A network acceleration node, comprising the node detecting device according to claim 8.
A path selection method, which is applied to an acceleration network having a plurality of network acceleration nodes according to claim 9, wherein a client accesses the corresponding source station through the acceleration network, including:

Receiving network delays and packet loss rates sent by multiple network acceleration nodes;

And selecting, according to the received network delay and the packet loss rate, an optimal path of each of the network nodes to the source station in a preset selection manner.
The path selection method according to claim 10, wherein the preset selection mode is one of the following:

Manner 1) calculating a path delay and a path loss ratio of all the paths of the acceleration node to the source station, and selecting a preset number of candidate paths in the order of the path delay from high to low, and Selecting a path with the lowest path loss rate in the to-be-selected path as an optimal path from the corresponding network acceleration node to the source station;

Method 2) calculating a path delay and a path loss rate of all paths of each of the acceleration nodes to the source station, to Selecting a preset number of to-be-selected paths, and selecting a path with the lowest path delay in the to-be-selected path is an optimal path from the corresponding network acceleration node to the source station. ;

Manner 3) performing weighting calculation on network delay and packet loss rate of each network acceleration node to other network acceleration nodes and source stations, and selecting each of the network nodes to the source station according to the weighting calculation result Excellent path.
The path selection method according to claim 11, wherein for the mode 1) and the mode 2), for the path including the N network acceleration nodes, the path has N-1 segments, and the corresponding path is lost. The formula for calculating the packet rate is:

The l i is the packet loss rate of the i th segment.
The path selection method according to claim 11, wherein for the mode 3), a formula for weighting calculation of network delay and packet loss rate of each of the network acceleration nodes to other network acceleration nodes and source stations is performed. for:

The RTT indicates the network delay of the current network acceleration node to a other network acceleration node or source station, and L indicates the packet loss rate of the current network acceleration node to a other network acceleration node or source station, and f(RTT, L) indicates the front The weighting value of the network delay and packet loss rate of the network acceleration node to a network acceleration node or source station is weighted, and λ is a weighting factor, which can be adjusted according to actual conditions; wherein f(RTT, L) is equal to -1 , indicating that the path is unreachable.
The path selection method according to claim 13, wherein the calculation of the optimal path of the network acceleration node to the source station is modified according to the following formula:

BWR1 and BWR2 respectively represent the bandwidth usage rate of two endpoints of a connection, and when the connection fails, the calculation result is -1.
The path selection method according to claim 10, further configured to: receive the number of active connections sent by the network acceleration node, and perform the number of active connections with the first connection number threshold and the second connection number threshold Comparing, and when the number of active connections is greater than the first connection number threshold, the network acceleration node is from the network acceleration network Removing, and continuing to monitor the number of active connections of the network acceleration node, and adding the network acceleration node to the acceleration network when the number of active connections is less than the second connection threshold.
A path selection device, wherein the path selection method according to any one of claims 10 to 15 is used to select an optimal path, which at least includes:

a receiving module, configured to receive network delays and packet loss rates sent by multiple network acceleration nodes;

The selecting module is configured to select an optimal path of each of the network nodes to the source station in a preset selection manner according to the received network delay and packet loss rate.
A server comprising the path selection device of claim 16.
A network system, comprising: a plurality of network acceleration nodes according to claim 9, and a server according to claim 17, wherein a client accesses the corresponding path through the network system with an optimal path Source station.