WO2018152919A1 - 一种路径选取方法及系统、网络加速节点及网络加速系统 - Google Patents

一种路径选取方法及系统、网络加速节点及网络加速系统 Download PDF

Info

Publication number
WO2018152919A1
WO2018152919A1 PCT/CN2017/078398 CN2017078398W WO2018152919A1 WO 2018152919 A1 WO2018152919 A1 WO 2018152919A1 CN 2017078398 W CN2017078398 W CN 2017078398W WO 2018152919 A1 WO2018152919 A1 WO 2018152919A1
Authority
WO
WIPO (PCT)
Prior art keywords
network acceleration
path
acceleration node
source station
path selection
Prior art date
Application number
PCT/CN2017/078398
Other languages
English (en)
French (fr)
Inventor
陈凯林
柳小鹏
Original Assignee
网宿科技股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 网宿科技股份有限公司 filed Critical 网宿科技股份有限公司
Priority to US16/326,779 priority Critical patent/US10951733B2/en
Priority to EP17897430.9A priority patent/EP3528440B1/en
Publication of WO2018152919A1 publication Critical patent/WO2018152919A1/zh

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00Network arrangements or protocols for supporting network services or applications
    • H04L67/50Network services
    • H04L67/60Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network resources
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00Arrangements for monitoring or testing data switching networks
    • H04L43/08Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
    • H04L43/0805Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters by checking availability
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00Arrangements for monitoring or testing data switching networks
    • H04L43/08Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
    • H04L43/0852Delays
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00Routing or path finding of packets in data switching networks
    • H04L45/70Routing based on monitoring results
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00Network arrangements or protocols for supporting network services or applications
    • H04L67/01Protocols
    • H04L67/10Protocols in which an application is distributed across nodes in the network
    • H04L67/1001Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
    • H04L67/1029Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers using data related to the state of servers by a load balancer
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00Arrangements for monitoring or testing data switching networks
    • H04L43/08Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
    • H04L43/0805Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters by checking availability
    • H04L43/0811Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters by checking availability by checking connectivity
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00Arrangements for monitoring or testing data switching networks
    • H04L43/08Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
    • H04L43/0823Errors, e.g. transmission errors
    • H04L43/0829Packet loss
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00Arrangements for monitoring or testing data switching networks
    • H04L43/08Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
    • H04L43/0876Network utilisation, e.g. volume of load or congestion level
    • H04L43/0882Utilisation of link capacity
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00Arrangements for monitoring or testing data switching networks
    • H04L43/08Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
    • H04L43/0876Network utilisation, e.g. volume of load or congestion level
    • H04L43/0894Packet rate

Definitions

  • the present invention relates to the field of network acceleration technologies, and in particular, to a path selection method and system, a network acceleration node, and a network acceleration system.
  • the application acceleration network is a network data acceleration system for improving the data transmission rate between the user and the source station, and can select a better quality transmission path for data transmission for different source stations accessed by users in different areas. accelerate.
  • the application acceleration network mainly adopts the architecture mode of the central server and the network acceleration node. In this mode, the network acceleration node is responsible for detecting the link state of the link to the source station and the link state of the other connected network acceleration nodes, and reporting the detection result to the central server, and the central server detects the detection result from each network acceleration node.
  • the optimal path from the user to the source station is selected, and the selection result is returned to each network acceleration node, and the network acceleration node performs transmission and feedback of the user request according to the selection result.
  • the existing network acceleration mode has the following disadvantages: (1) When a network acceleration node receives a user request, each network acceleration node in the application acceleration network newly adds a link state of its own to the source station. As a task, the task of detecting the link state of other network acceleration nodes is detected. As a result, the number of probe tasks suddenly increases, and the probe traffic needs to occupy a large bandwidth and may even affect service traffic. (2) The huge number of detection tasks is reported to the central server, which causes the central server to have a particularly high computational pressure and the feedback speed is affected. As a result, the existing application acceleration network improves the data transmission speed and quality between the user and the source station to a certain extent, but it has its own drawbacks, which cannot meet the increasing user demand and cannot provide better. User experience.
  • the present invention provides a path selection method and system, a network acceleration node, and a network acceleration system.
  • the network acceleration node itself selects the optimal path to the source station, thereby reducing the number of path detection, optimizing the quality of the path selection, and speeding up the path selection efficiency.
  • an ingress network acceleration node detects a link delay of each egress network acceleration node, and determines whether each egress network acceleration node is available.
  • the ingress network acceleration node is configured to receive a user access request; the ingress network acceleration node obtains, from an available egress network acceleration node, a link delay of the egress network acceleration node to the source station IP; the ingress network acceleration node is configured according to The link delay of the available egress network acceleration node and the link delay of the available egress network acceleration node to the source station IP, and the optimal selection of the user access request to reach the source station IP path.
  • the determining, by the ingress network acceleration node, whether the egress network acceleration node is available includes the following steps: the ingress network acceleration node detects and acquires a delay of a link to the egress network acceleration node, Packet loss rate, load, and bandwidth usage rate; when one or more of the delay, packet loss rate, load, and bandwidth usage rate of the link to the egress network acceleration node are lower than a corresponding preset threshold The egress network acceleration node is available.
  • the ingress network acceleration node calculates a total link delay of the available egress network acceleration node and the available egress network acceleration node to the source station IP, and selects a total link. A path with the smallest delay is used as the optimal path.
  • a link delay of the ingress network acceleration node reaching the egress network acceleration node, and a link delay of the egress network acceleration node reaching the source station IP using an average method Calculated.
  • a path saving step is included to save the optimal path.
  • a query step is further included, and the saved optimal path is queried according to the user access request information. If the corresponding record exists, the path is directly selected as the current time. The path to the request.
  • the ingress network acceleration node classifies source IPs belonging to the same operator and the same geographical area into the same IP packet, and uses the path list to save the IP packet for each of the IP packets.
  • the portal network acceleration node generates a historical path list, where the historical path list is used to save a historical path of the user IP to the source station IP, and each of the historical paths has an effective time; In the valid time, the access request of the same source IP to the same source IP is used in the historical path list. User IP to the historical path of the source station IP, and reset the effective time.
  • the present invention further provides a path selection system, including a detection module, an analysis processing module, a transceiver module, and an operation processing module; the detection module is respectively connected to the analysis processing module, the transceiver module, and the operation processing module.
  • the transceiver module is respectively connected to the detection module, the analysis processing module, and the operation processing module; the detection module detects link delay and loss of the network acceleration node itself to other network acceleration nodes or source station IPs.
  • the analysis processing module receives the detection result from the detection module, Determining whether the egress network acceleration node is available based on the detection result, and sending the determination result to the transceiver module; the egress network acceleration node is a last hop network acceleration node connected to the source station; the operation processing module is The transceiver module acquires and or the network acceleration node itself of the transceiver module arrives The egress network accelerates the delay of the node link, and the available egress network accelerates the link delay of the node to the source station IP, and calculates the sum of the two, selects the path with the smallest delay as the path selection result, and sends the path selection result
  • the transceiver module receives a user access request, a detection result from the detection module, a determination result of the analysis processing module, and a path selection result of the operation processing module, and sends the path selection result.
  • the analyzing, by the analysis processing module, determining whether the egress network acceleration node is available includes the following steps: acquiring a link delay, a packet loss rate, and a load from the detecting module to the egress network acceleration node. And bandwidth usage rate; determining that the egress network acceleration node is available when one or more of the link delay, packet loss rate, load, and bandwidth usage rate are lower than a corresponding preset threshold.
  • the detecting module detects a link delay of the ingress network acceleration node reaching the egress network acceleration node and a delay of reaching the link of the source station IP, and calculates by the average method. .
  • a storage module is further included, and the storage module is connected to the transceiver module, configured to receive the path selection result from the transceiver module, and store the path selection result in a path selection list.
  • the query is first performed in the path selection list. If the same user IP accesses the record of the same source IP address, the corresponding path is selected for transmission and returning.
  • the storage module further includes a path list, where the source station IPs belonging to the same operator and the same geographical area are classified into the same IP group, and each of the IP packets is recorded.
  • a path list where the source station IPs belonging to the same operator and the same geographical area are classified into the same IP group, and each of the IP packets is recorded.
  • the method further includes a storage module, wherein the storage module stores a historical path list, and the historical road The path list is used to store the historical path of the user IP to the source station IP, and the effective time of each of the historical paths; if the same user IP is sent to the same source station IP within the validity time of the historical path The request adopts the historical path of the user IP saved in the historical path list to the source station IP, and resets the effective time.
  • the storage module stores a historical path list, and the historical road The path list is used to store the historical path of the user IP to the source station IP, and the effective time of each of the historical paths; if the same user IP is sent to the same source station IP within the validity time of the historical path The request adopts the historical path of the user IP saved in the historical path list to the source station IP, and resets the effective time.
  • the present invention further provides a network acceleration node, which uses the path selection system described in any of the above embodiments to select an optimal path of the source station IP.
  • the present invention further provides a network acceleration system, including a plurality of the foregoing network acceleration nodes, so that a user selects an optimal path through the network acceleration node to access a corresponding source station IP.
  • the path selection method and system, the network acceleration node, and the network acceleration system of the present invention have the following beneficial effects:
  • the network acceleration node itself calculates the optimal path to the source station, without the intervention of the central server, speeding up the efficiency of path selection;
  • the subsequent path selection can be performed based on the previous path selection result, which improves the path selection efficiency.
  • FIG. 1 is a flow chart of a path selection method according to a preferred embodiment of the present invention.
  • FIG. 2 is a schematic diagram of a path selection method according to an embodiment of the present invention.
  • FIG. 3 is a schematic structural diagram of a path selection system according to a preferred embodiment of the present invention.
  • FIG. 4 is a schematic structural diagram of a network acceleration system according to an embodiment of the present invention.
  • Source station a website where Internet content providers provide users with access to resources, deployed on servers in different geographical locations and different Internet operators.
  • Source station IP the IP address corresponding to the source station.
  • the same source station has different IP addresses assigned due to the different servers.
  • the network acceleration node which is deployed in different geographical locations and different carrier networks, has a connection with each other to form a network, which connects the user and the source station.
  • the ingress network acceleration node receives the network acceleration node of the user access request.
  • the intermediate network acceleration node is respectively connected to the network acceleration node of the ingress network acceleration node and the egress network acceleration node.
  • Delay refers to the time it takes for a message or packet to be transmitted from one end of a network to the other.
  • Packet loss rate refers to the ratio of the number of lost packets to the transmitted data group.
  • Bandwidth usage the ratio of used bandwidth to total link bandwidth.
  • FIG. 1 is a flow chart of a path selection method according to a preferred embodiment of the present invention. As shown in FIG. 1 , when the ingress network acceleration node in the present invention receives the user access request, the path selection method provided by the embodiment is used to select the optimal path, which specifically includes the following steps:
  • Step S1 The ingress network acceleration node detects the link delay (RTT1) of each egress network acceleration node, and determines whether each egress network acceleration node is available.
  • RTT1 link delay
  • the ingress network acceleration node determining whether the egress network acceleration node is available includes the following steps:
  • the ingress network acceleration node detects and acquires the delay, packet loss rate, load and bandwidth usage rate of the link to the egress network acceleration node;
  • the link delay, packet loss rate, load, and bandwidth usage rate have corresponding preset thresholds.
  • the preset threshold may be preset according to the actual state and actual demand of each network acceleration node, and the preset threshold is saved in the program of each network acceleration node, may be saved in an unmodifiable form, or may be in a modifiable form. Save it.
  • link delay, packet loss rate, the load, and the bandwidth usage rate can each have a corresponding priority.
  • the parameter of a certain priority is lower than its preset threshold, it can be determined that the egress network acceleration node is available. For example, the load and bandwidth usage rate is the first priority, the delay is the second priority, and the packet loss rate is the third priority.
  • the egress network acceleration node When the load and/or bandwidth usage is lower than the corresponding preset threshold, it may be determined that the egress network acceleration node is available. Alternatively, when the parameters of the second priority level and above are lower than their corresponding preset thresholds, that is, the load, the bandwidth usage rate, and the delay are both lower than their corresponding preset thresholds, the egress network acceleration node may be determined to be available.
  • the ingress network acceleration node is also an egress network acceleration node, that is, the network acceleration node not only accepts user access requests, but is also the last hop network acceleration node connected to the source station.
  • the ingress network acceleration node detects the link delay, the packet loss rate, the load, and the bandwidth usage rate of each egress network acceleration node, the link delay of detecting the arrival of the ingress network acceleration node itself is included. , packet loss rate, load and bandwidth usage.
  • the ingress network acceleration node and the egress network acceleration node may be directly connected, or may be connected by one or more other network acceleration nodes. That is to say, the link of the ingress network acceleration node to each egress network acceleration node may be connected to one or more intermediate network acceleration nodes, or may be a direct connection without connecting other network acceleration nodes.
  • the delay, the packet loss rate, the load, and the bandwidth usage rate are sent to the ingress network acceleration node, and the ingress network acceleration node is available to the next hop node of the intermediate network acceleration node according to the foregoing determining method, that is, in this embodiment.
  • the availability judgment of the intermediate network acceleration node and the egress network acceleration node is completed by the ingress network acceleration node; and the second method: the ingress network acceleration node or the intermediate network acceleration node first delays the link acceleration of the network acceleration node that reaches the next hop, The packet loss rate, the load, and the bandwidth usage rate are detected, and the next hop network acceleration node is judged by the above method, and if available, a probe request is sent to the available network acceleration node, and the available network acceleration node continues to One-hop network acceleration node performs detection and judgment until the egress network accelerates the node, so
  • the delay of the network acceleration node reaching the egress network acceleration node is the sum of the delays of the intermediate link links, and the availability judgment tasks of the relevant nodes are also dispersed in the intermediate nodes, thereby reducing the task amount of the ingress network acceleration node and implementing the load. balanced.
  • Step S2 The ingress network acceleration node acquires, from the available egress network acceleration node, a link delay (RTT2) of the egress network acceleration node to the source station IP.
  • RTT2 link delay
  • step S2 includes the following steps:
  • the ingress network acceleration node initiates a probe request for the source station IP only to the available egress network acceleration node;
  • the ingress network acceleration node receives the link delay (RTT2) of the egress network acceleration node sent by the egress network acceleration node to the source station IP.
  • RTT2 link delay
  • the egress network acceleration node When the egress network acceleration node receives the probe request for the source station IP, it detects the link delay (RTT2) from itself to the source station IP, and returns the detected link delay (RTT2) to the ingress network acceleration. node. That is to say, the egress network acceleration node detects its own link delay to the source station IP only after receiving the probe request.
  • RTT2 link delay
  • RTT2 link delay
  • Step S3 The ingress network acceleration node accelerates the link delay of the node according to the egress network that is available to the egress network, and the available egress network accelerates the link delay of the node to the source IP, and selects the egress network acceleration node to reach the source IP address. Excellent path.
  • the ingress network acceleration node calculates the total link delay of the available egress network acceleration node and the available egress network acceleration node to the source station IP, and selects one path with the smallest total link delay as the optimal path.
  • the link delay of the ingress network acceleration node to the egress network acceleration node and the link delay of the egress network acceleration node to the source station IP are obtained, and are calculated by the average method for the last few times or a period of time.
  • Link The delay is used to calculate the average value as the delay data of the link. In this way, the influence of abnormal data generated by the network link jitter on the path selection can be reduced.
  • FIG. 2 is a schematic diagram of a path selection method according to an embodiment of the present invention.
  • A is an ingress network acceleration node, and directly receives an access request of the user U;
  • B, C, and D are an egress network acceleration with only one hop from the source station X. node.
  • A When A selects a path for access request to user U, it first detects link delays RTTB1, RTTC1, and RTTD1 that arrive at B, C, and D, respectively, and according to the obtained link delays of arrival B, C, and D, The packet loss rate, load, bandwidth usage, and corresponding preset thresholds determine whether B, C, and D are available egress network acceleration nodes. If the load of B is higher than the preset threshold, it is judged that B is unavailable. Therefore, A only sends probe requests to C and D. After receiving the probe request from A, C and D detect the link delays RTTC2 and RTTD2 that they reach the source station X, and return the detection result to A.
  • A calculates the sum of RTTC1 and RTTC2, and the sum of RTTD1 and RTTD2, and finally selects the path AC corresponding to the smaller sum of RTTC1 and RTTC2 as the result of path detection, that is, selects the ingress network acceleration node A-egress network acceleration node C-
  • the source station X serves as the optimal path for the user U to access the source station X.
  • the acquired RTTB1, RTTC1, RTTC2, RTTD1, and RTTD2 are calculated by an average value method, that is, an average value of the corresponding link delays in the most recent or one time period is calculated as the link. Delay data.
  • the method further includes the step of saving the optimal path obtained according to the path selection method.
  • the saved optimal path is first queried. Whether there is a corresponding record in the middle, if any, the corresponding path is directly selected as the access path. If there is no corresponding record, the optimal path is detected according to the above method.
  • the saved optimal path record retention time is one detection period, and the current detection period is cleared after the end of the current detection period, and continues to be stored in the new detection period.
  • the clearing method may be directly recording. Deletion may also be to hide it, and the invention is not limited.
  • the optimal path is the optimal path when the same user sends an access request to the same source station IP in the current detection period, that is, in the same detection period, when the same user accesses the source station again.
  • IP the access request will be transmitted from the ingress network acceleration node to the source station IP according to the previously detected optimal path, and the source station IP will return to the source according to the original path, thereby ensuring that the user request can obtain the fastest transmission and feedback.
  • the new cycle begins, it needs to be re-detected.
  • the ingress network acceleration node groups the source station IP according to the network operator and the geographic location, and classifies the source station IPs belonging to the same carrier and the same geographical area into the same IP group, and for each IP.
  • the packet uses a path list to store each path used by the source station IP in the IP packet and its number of uses.
  • the paths are sorted in descending order of the number of uses of each path used by the IP packet. Every When the network acceleration node completes the path selection for a source station IP, it adds 1 to the number of times the selected optimal path is used in the path list of the IP packet to which the source station IP belongs.
  • the ingress network acceleration node When the ingress network acceleration node receives the user's access request for a certain source station IP, if the path selection has not been performed for the source station IP (that is, the corresponding record cannot be queried in the saved optimal path record), then the selection is performed.
  • the path with the highest number of paths used in the path list corresponding to the IP address of the source station IP is used as the path of the current access request, and the path selection method is used for the current access request according to the path selection method, and the detection result is saved for subsequent users. The same access request is used.
  • the current access path is transmitted and returned according to the preset default path, and according to the path selection method, the access request is requested. Perform path selection to save the probe results.
  • the ingress network acceleration node can select a preferred path for transmission and returning according to the path list or the default path in time, thereby ensuring user access request transmission and feedback.
  • the stability and to some extent, guarantees the quality of the transmission path.
  • a history path list is generated on the ingress network acceleration node, where the history path list is used to save the historical path of the user IP to the source station IP, and configure an effective time for each history path.
  • the historical path of the user IP saved in the historical path list to the source IP is used, and the effective time is reset.
  • the source station requires that the requested user IP may not change during the same session, which requires the same user IP to reach the source IP path during this session. It cannot be changed internally, otherwise the source station may request to log in again, so that a valid connection cannot be established.
  • a historical path list is generated on the network acceleration node for storing a historical path of a user IP to a source station IP, and each history path is set with a valid time. If the request of the same user IP to the same source IP address is received within the validity time of the historical path, the historical path of the user IP saved in the historical path list to the source IP address is used, and the reset is valid. time.
  • the path selection method provided by the present invention implements path selection through the ingress network acceleration node, and the egress network acceleration node only needs to detect the source station IP when receiving the probe request. In this way, not only the original acceleration network architecture is simplified, but also the number of detection tasks of the network acceleration node and the final calculation selection amount are reduced, thereby releasing the broadband occupied by the excessive detection amount. Moreover, the path selection method provided by the present invention comprehensively considers the comprehensive factors such as the delay of the link, the packet loss rate, the load and the bandwidth usage rate in the process of selecting the path, which not only ensures the quality of the transmission line but also improves the quality. Load balancing of each network acceleration node.
  • the influence of the network jitter on the delay data can be prevented to some extent; by using the path table, the user access request for the path detection can be ensured quickly. Assigned to a better path, avoiding the abnormality of the feedback rate of the secondary user access request due to the uncertainty of the path; by using the historical path list, the uniformity of the path used in the same network session can be guaranteed, and the cause is avoided.
  • the source station refuses to respond to the difference in road strength.
  • FIG. 3 illustrates a path selection system according to a preferred embodiment of the present invention.
  • the path selection system provided in this embodiment includes a detection module 10, an analysis processing module 20, a transceiver module 30, and an operation processing module 40.
  • the detection module 10 and the transceiver module 30 are respectively connected to the other three modules.
  • the path selection system is applied to each network acceleration node.
  • the detecting module 10 is configured to detect a link delay of the network acceleration node itself to reach another network acceleration node or to the source station IP, and obtain a packet loss rate of the network acceleration node and other network acceleration nodes and load and bandwidth usage of the network acceleration node. And transmitting the detection result to the analysis processing module 20 or the transceiver module 30. Specifically, if the network acceleration node is an ingress network acceleration node, when receiving the user access request, the detecting module 10 detects the delay, the packet loss rate, the load, and the bandwidth usage rate of the link that reaches the egress network acceleration node.
  • the network acceleration node is an egress network acceleration node, when the egress network acceleration node receives the probe request from the ingress network acceleration node, the detection module 10 reaches the source station IP
  • the link performs delay detection, and sends the detection result to the transceiver module 30, and the transceiver module 30 returns the detection result to the requesting node; if the network acceleration node is the intermediate network on the link of the ingress network acceleration node to the egress network acceleration node
  • the acceleration node detects when the ingress network acceleration node detects the link to the egress network acceleration node, and the detection module 10 detects the delay, packet loss rate, load, and bandwidth usage of the link to the next-hop intermediate network acceleration node. And transmitting the detection result to the transceiver module 30.
  • the analysis processing module 20 receives the detection result from the detection module 10, determines whether the egress network acceleration node is available based on the detection result, and sends the determination result to the transceiver module 20. Specifically, the analyzing the processing module 20 determining whether the egress network acceleration node is available includes the following steps:
  • the delay, the packet loss rate, the load, and the bandwidth usage rate of the link have corresponding preset thresholds, and the preset thresholds may be preset according to the actual state and actual requirements of each network acceleration node.
  • the threshold is saved in each network acceleration node, and can be saved in an unmodifiable form or in a modifiable form.
  • link delay, packet loss rate, the load, and the bandwidth usage rate can each have a corresponding priority.
  • the parameter of a certain priority is lower than its preset threshold, it can be determined that the egress network acceleration node is available. For example, the load and bandwidth usage rate is the first priority, the delay is the second priority, and the packet loss rate is the third priority.
  • the egress network acceleration node When the load and/or bandwidth usage is lower than the corresponding preset threshold, it may be determined that the egress network acceleration node is available. Alternatively, when the parameters of the second priority level and above are lower than their corresponding preset thresholds, that is, the load, the bandwidth usage rate, and the delay are both lower than their corresponding preset thresholds, the egress network acceleration node may be determined to be available.
  • At least one intermediate network acceleration node is further connected between the ingress network acceleration node and the egress network acceleration node.
  • the ingress network acceleration node includes, when determining whether the egress network acceleration node is available, To determine whether the intermediate network acceleration nodes are available, there are mainly two implementation manners. The first method: the intermediate network acceleration node sends the detection result of the detection module 10 to the next hop node to the ingress network acceleration node through the transceiver module 30.
  • the ingress network acceleration node determines whether the next hop node of the intermediate network acceleration node is available according to the foregoing determining method, that is, in this embodiment, the availability judgment of the intermediate network acceleration node and the egress network acceleration node are determined by the ingress network.
  • the acceleration processing node is configured to receive the detection result of the detection module 10 and determine whether the next hop network acceleration node of the intermediate network acceleration node is available, that is, in this embodiment. , the availability of each network acceleration node It has its previous hop off the network node to complete the acceleration, this way, the amount of dispersible task processing, better load balancing.
  • the operation processing module 40 obtains, according to the analysis processing result, the delay of the link to the available egress network acceleration node from the detection module 10, and obtains the link delay of the available egress network acceleration node from the transceiver module 30 to the source station IP, and calculates The sum of the two is selected as the path selection result, and the path selection result is sent to the transceiver module 30. It should be noted that, in this embodiment, the operation processing module 40 obtains the analysis result through the transceiver module 30, and in other embodiments of the present invention, the operation processing module 40 may directly obtain the analysis processing result from the analysis processing module 20, or The analysis processing results are obtained in other ways, and the present invention is not limited thereto.
  • the transceiver module 30 includes a receiving unit 31 and a sending unit 32.
  • the receiving unit 31 receives data from inside and outside of the network acceleration node, including user access requests from external sources and probe requests and detection results sent by other network acceleration nodes.
  • the detection result of the internal detection module 10, the analysis processing result of the analysis processing module 20, and the path selection result of the operation processing module 40, the sending unit 32 transmits relevant data to the outside of the network acceleration node, including sending a probe request to the available ingress network acceleration node, Send probe requests and path selection results to other network acceleration nodes.
  • the path selection system further includes a storage module 50 and a query module 60.
  • Storage module 50 is connected to the transceiver module 30, and obtains a path selection result from the transceiver module 30, and saves the path selection results in the path selection list L1, including user information, source station IP, and path.
  • the path selection list is in one cycle. Valid, will be cleared after the end of the cycle, and will be stored in the path selection result of the new cycle.
  • the query module 60 first queries the path selection list L1 in the storage module. If the same user accesses the corresponding record of the same source station, the corresponding path is selected for transmission and return.
  • the storage module 50 further includes a path selection list L2, which belongs to the same operator and the same
  • the source station IP of the geographic area is classified into the same IP packet, and a path list is used for each IP packet to save each path used by the source station IP in the IP packet and its usage count.
  • the query module 60 cannot find the corresponding record in the path selection list L1, and selects the path corresponding to the IP packet where the source station IP is located. The path with the highest number of paths used in the list L2 is used as the path of the current access request.
  • the source station requires that the requested user IP cannot change during the same session. This requires that the path of the same user IP to the source IP address cannot be changed for a period of time, otherwise it may be requested by the source station. re-register.
  • the storage module 50 further stores a historical path list L3, and the historical path list L3 stores a historical path of a user IP to a source station IP, and each historical path is set to be effective. time. If the request of the same user IP is sent to the same source IP address, the query module 60 first queries the historical path list L3.
  • the path selection system provided in the embodiment of the present invention runs on the network acceleration node, that is, the network acceleration node in the embodiment of the present invention may be an ingress network acceleration node during different user access request sessions. It is an egress network acceleration node or an intermediate network acceleration node. When the location of the network acceleration node is different, the functions of each module of the system are also adjusted accordingly. In this way, the network architecture formed by each network acceleration node is simple and flexible, and can be applied to different application scenarios, thereby better meeting actual needs.
  • the present invention also provides a network acceleration node that uses the path selection system described above to select an optimal path for itself to reach the source station IP.
  • FIG. 4 is a schematic structural diagram of a network acceleration system according to an embodiment of the present invention.
  • the network acceleration system includes a plurality of the above-mentioned network acceleration nodes 22, so that the user 21 selects an optimal path through the network acceleration node to access the corresponding source station 23IP.
  • the user sends an access request to the corresponding ingress network acceleration node, and the path selection system of the ingress network acceleration node selects the path, and records an optimal path from the network acceleration node to the source station IP, so that the user can
  • the optimal path enables access to the source station IP, thereby speeding up network access and providing a user experience.
  • the path selection method and system, the network acceleration node, and the network acceleration system of the present invention calculate the optimal path to the source station by the network acceleration node itself, without the intervention of the central server, speeding up the efficiency of path selection;
  • the network acceleration node can initiate path detection, reduce the number of path detection, optimize the quality of the path selection, and can select subsequent paths based on the previous path selection result, thereby improving the path selection efficiency. Therefore, the present invention effectively overcomes various shortcomings in the prior art and has high industrial utilization value.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Environmental & Geological Engineering (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

本发明提供一种路径选取方法及系统、网络加速节点、网络加速系统,包括入口网络加速节点探测到达各个出口网络加速节点的链路时延,并判断各个出口网络加速节点是否可用;入口网络加速节点从可用的出口网络加速节点获取该出口网络加速节点到达源站IP的链路时延;入口网络加速节点根据自身到达可用的出口网络加速节点的链路时延,以及可用的出口网络加速节点到达源站IP的链路时延,选取入口网络加速节点到达源站IP的最优路径。本发明的路径选取方法及系统、网络加速节点、网络加速系统由网络加速节点自身来计算到源站的最优路径,从而降低路径探测数量,优化路径选取的质量,加快了路径选取的效率。

Description

一种路径选取方法及系统、网络加速节点及网络加速系统 技术领域
本发明涉及网络加速技术领域,特别是涉及一种路径选取方法及系统、网络加速节点、网络加速系统。
背景技术
随着计算机网络技术的不断发展,互联网技术已成为人类生活的重要组成部分。用户对网络的传输速率及用户请求的反馈速度要求越来越大。在最初的技术架构中,用户的请求数据直接发送到源站服务器,源站服务器根据用户请求进行处理后,再直接返回给用户。由于用户和源站服务器之间可能存在地理位置距离远、归属不同的网络运营商、线路受损、线路过载等问题,用户与源站之间的通信会出现传输慢、反馈慢、费用高、传输中断等现象。如此一来,不仅影响了用户的上网体验,也增加了内容提供商的成本,甚至带来其他不可预估的损失。
随着技术的改良和进步,现有技术中,通过在用户和源站之间架设应用加速网络来缓解上述问题。应用加速网络是一种用于提升用户与源站之间数据传输速率的网络数据加速系统,可以针对不同区域的用户所访问的不同源站,分别选择质量较优的传输路径来为数据传输进行加速。应用加速网络主要采用中央服务器、网络加速节点的架构模式。这种模式下,网络加速节点负责探测自身到源站的链路状态以及到其他相连网络加速节点的链路状态,并且将探测结果上报至中央服务器,中央服务器根据来自各个网络加速节点的探测结果选择出用户到源站的最优路径,并将选择结果返回给各网络加速节点,网络加速节点根据选择结果进行用户请求的传输与反馈。
然而,现有网络加速模式存在以下不足:(1)当某个网络加速节点收到一个用户请求时,应用加速网络中的每个网络加速节点都新增探测自身到源站的链路状态的任务和探测到相连其他网络加速节点的链路状态的任务,如此一来,探测任务数量骤然增加,探测流量需要占用较大的带宽,甚至会影响到业务流量。(2)巨大的探测任务数量上报至中央服务器,导致中央服务器的计算压力特别大,反馈速度受到影响。如此一来,现有的应用加速网络虽然在一定程度上改善了用户和源站之间的数据传输速度和质量,但其自身存在一定弊端,无法满足日益增长的用户需求,也无法提供更优的用户体验。
发明内容
鉴于以上所述现有技术的不足,本发明提供了一种路径选取方法及系统、网络加速节点、网络加速系统。本发明所提供的技术方案中,网络加速节点自身来选择到源站的最优路径,从而降低路径探测数量,优化路径选取的质量,加快了路径选取的效率。
为实现上述目的及其他相关目的,本发明提供一种路径选取方法,包括以下步骤:入口网络加速节点探测到达各个出口网络加速节点的链路时延,并判断所述各个出口网络加速节点是否可用,其中所述入口网络加速节点用于接收用户访问请求;所述入口网络加速节点从可用出口网络加速节点获取该出口网络加速节点到达源站IP的链路时延;所述入口网络加速节点根据自身到达所述可用出口网络加速节点的链路时延,以及所述可用出口网络加速节点到达所述源站IP的链路时延,选取所述用户访问请求到达所述源站IP的最优路径。
于本发明一实施例中,所述入口网络加速节点判断所述出口网络加速节点是否可用包括以下步骤:所述入口网络加速节点探测并获取到达所述出口网络加速节点的链路的时延、丢包率、负载及带宽使用率;当所述到达所述出口网络加速节点的链路的时延、丢包率、负载及带宽使用率中的一项或者多项低于相应预设阈值时,则所述出口网络加速节点可用。
于本发明一实施例中,所述入口网络加速节点计算自身到达所述可用出口网络加速节点、所述可用的出口网络加速节点到达所述源站IP的总链路时延,选取总链路时延最小的一条路径作为最优路径。
于本发明一实施例中,所述入口网络加速节点到达所述出口网络加速节点的链路时延、以及所述出口网络加速节点到达所述源站IP的链路时延,采用平均值法计算得到。
于本发明一实施例中,包含一路径保存步骤,保存所述最优路径。
于本发明一实施例中,在所述探测步骤之前更包含一查询步骤,根据所述用户访问请求信息查询所保存的所述最优路径,若存在相应记录,则直接选取该路径作为本次访问请求的路径。
于本发明一实施例中,所述入口网络加速节点将属于同一运营商、同一地理区域的源站IP归为同一IP分组,并针对每一个所述IP分组采用一个路径列表保存所述IP分组使用过的每一条路径及其使用次数;当所述入口网络加速节点接收到用户针对某个所述源站IP的访问请求时,若尚未针对该所述源站IP进行路径选取,则选取所述源站IP所在所述IP分组对应的所述路径列表中路径使用次数最高的作为本次访问请求的路径。
于本发明一实施例中,所述入口网络加速节点上生成历史路径列表,所述历史路径列表用于保存用户IP到达源站IP的历史路径,且每个所述历史路径设有有效时间;在有效时间内,针对同一用户IP对同一源站IP的访问请求,采用所述历史路径列表中所保存的所述用 户IP到所述源站IP的历史路径,并重置有效时间。
同时,本发明还提供一种路径选取系统,包括探测模块、分析处理模块、收发模块、运算处理模块;所述探测模块分别与所述分析处理模块、所述收发模块和所述运算处理模块相连,所述收发模块分别与所述探测模块、所述分析处理模块和所述运算处理模块连接;所述探测模块探测网络加速节点自身到达其他网络加速节点或源站IP的链路时延和丢包率,并获取所述其他网络加速节点的负载和带宽使用率,并将探测结果发送至所述分析处理模块或所述收发模块;所述分析处理模块接收来自所述探测模块的探测结果,基于所述探测结果判断出口网络加速节点是否可用,并将判断结果发送至所述收发模块;所述出口网络加速节点为与源站相连的最后一跳网络加速节点;所述运算处理模块从所述收发模块获取和或所述收发模块所述网络加速节点自身到达可用出口网络加速节点链路的时延、可用出口网络加速节点到达源站IP的链路时延,并计算两者之和,选取时延最小的作为路径选取结果,并将所述路径选取结果发送至所述收发模块;所述收发模块接收用户访问请求、来自所述探测模块探测结果、所述分析处理模块的判断结果以及所述运算处理模块的路径选取结果,发送探测请求、所述探测结果及所述路径选取结果。
于本发明一实施例中,所述分析处理模块判断所述出口网络加速节点是否可用包括以下步骤:从所述探测模块获取到达所述出口网络加速节点的链路时延、丢包率及负载和带宽使用率;当所述链路时延、丢包率、负载及带宽使用率其中一项或多项低于相应预设阈值时,判定所述出口网络加速节点可用。
于本发明一实施例中,所述探测模块探测入口网络加速节点到达所述出口网络加速节点的链路时延、到达所述源站IP的链路时延时,通过平均值法计算得出。
于本发明一实施例中,还包含存储模块,所述存储模块与所述收发模块相连,用于从所述收发模块接收所述路径选取结果,并将所述路径选取结果存储在路径选取列表;当接收到用户访问请求时,先在所述路径选取列表中进行查询,若存在同一用户IP访问同一源站IP的记录,则选取对应的路径进行传输和回源。
于本发明一实施例中,还包括存储模块,存有路径列表,所述路径列表中对属于同一运营商、同一地理区域的源站IP归为同一IP分组,记录每一个所述IP分组中的所述源站IP使用过的每一条路径及其使用次数;当接收到用户针对某个源站IP的访问请求时,若在所述路径选取列表中无法查找到相应记录,则选取所述路径列表中所述源站IP所在IP分组对应的路径使用次数最高的路径作为本次访问请求的路径。
于本发明一实施例中,还包括存储模块,所述存储模块存有历史路径列表,所述历史路 径列表用于存储用户IP到源站IP的历史路径,及每个所述历史路径的有效时间;在所述历史路径有效时间内,若收到同一个用户IP发往同一个源站IP的请求,则采用所述历史路径列表中所保存的该用户IP到该源站IP的历史路径,并重置有效时间。
另外,本发明还提供一种网络加速节点,采用上述任一实施例所述的路径选取系统选取自身到达源站IP的最优路径。
相应的,本发明还提供一种网络加速系统,包括若干上述的网络加速节点,以使用户通过所述网络加速节点选取最优路径访问相应的源站IP。
如上所述,本发明的路径选取方法及系统、网络加速节点、网络加速系统,具有以下有益效果:
(1)由网络加速节点自身来计算到源站的最优路径,无需中央服务器介入,加快了路径选取的效率;
(2)仅向可用网络加速节点发起路径探测,降低了路径探测数量,优化了路径选取的质量;
(3)能够基于在前路径选取结果进行后续路径选取,提高了路径选取效率。
(4)通过平均值法计算链路时延,降低了因网络抖动而带来的数据异常对探测结果的影响。
(5)通过历史路径列表的使用,保证同一会话中的用户访问请求可以按照原路进行回源。
附图说明
图1绘示为本发明一较佳实施例所提供的路径选取方法的流程图;
图2绘示为本发明一实施例所提供的路径选取方法过程示意图;
图3绘示为本发明一较佳实施例所提供的路径选取系统的结构示意图;
图4绘示为本发明一实施所提供的网络加速系统结构示意图。
元件标号说明
10       探测模块
20       分析处理模块
30       收发模块
31       接收单元
32       发送单元
40       运算处理模块
50       存储模块
60       查询模块
21       用户
22       网络加速系统
23       源站
具体实施方式
以下通过特定的具体实例说明本发明的实施方式,本领域技术人员可由本说明书所揭露的内容轻易地了解本发明的其他优点与功效。本发明还可以通过另外不同的具体实施方式加以实施或应用,本说明书中的各项细节也可以基于不同观点与应用,在没有背离本发明的精神下进行各种修饰或改变。需说明的是,在不冲突的情况下,以下实施例及实施例中的特征可以相互组合。
需要说明的是,以下实施例中所提供的图示仅以示意方式说明本发明的基本构想,遂图式中仅显示与本发明中有关的组件而非按照实际实施时的组件数目、形状及尺寸绘制,其实际实施时各组件的型态、数量及比例可为一种随意的改变,且其组件布局型态也可能更为复杂。
首先,对本发明说明书中的技术词汇进行解释说明。以下说明是对本发明技术的解释,但本发明并不以此为限。
源站,互联网内容提供商为用户提供资源访问的网站,部署在不同地理位置、不同互联网运营商的服务器上。
源站IP,源站所对应的IP地址。同一个源站由于所在服务器的不同,所分配到的IP地址也不同。用户向源站发起请求时,无论是通过IP请求的方式,还是DNS请求方式,最终都会指向于一个具体的源站IP。
用户,各种用户应用程序,通过向源站发出访问请求来获取所需的资源服务。
网络加速节点,部署在不同地理位置、不同运营商网络中的代理服务设备,相互之间有连接,构成一个网络,该网络连接用户和源站。
入口网络加速节点,接收用户访问请求的网络加速节点。
出口网络加速节点,与源站相连的最后一跳网络加速节点。
中间网络加速节点,分别连接入口网络加速节点和出口网络加速节点的网络加速节点。
时延,指一个报文或分组从一个网络的一端传送到另一个端所需要的时间。
丢包率,指所丢失数据包数量占所发送数据组的比率。
负载,指任务队列数量与机器CPU核心数量的百分比。
带宽使用率,已使用带宽与链路总带宽的比率。
请参照图1,图1绘示本发明一较佳实施例所提供的的路径选取方法的流程图。如图1所示,当本发明中的入口网络加速节点接收到用户访问请求时,通过本实施例所提供的路径选取方法进行最优路径的选取,具体包括以下步骤:
步骤S1、入口网络加速节点探测到达各个出口网络加速节点的链路时延(RTT1),并判断各个出口网络加速节点是否可用。
具体地,入口网络加速节点判断出口网络加速节点是否可用包括以下步骤:
11)入口网络加速节点探测并获取到达出口网络加速节点的链路的时延、丢包率、负载及带宽使用率;
12)当出口网络加速节点的链路的时延、丢包率、负载和带宽使用率均低于相应的预设阈值时,判断该出口网络加速节点可用。
本发明中,链路的时延、丢包率、负载和带宽使用率都有相应的预设阈值。所述预设阈值可根据各网络加速节点的实际状态和实际需求来进行预设,预设阈值保存在各网络加速节点的程序中,可以以不可修改的形式保存,也可以以可修改的形式进行保存。
值得注意的是,本实施例中,在判断出口网络加速节点是否可用时,综合考虑了链路时延、丢包率、负载和带宽使用率是否均低于相应的预设阈值,而在本发明的其他实施例中,可根据实际需求仅考虑其中一项或多项是否低于相应阈值。另外,链路时延、丢包率、负载和带宽使用率可分别具有相应的优先级。当一定优先级的参数低于其预设阈值时,则可判断该出口网络加速节点可用。例如,负载和带宽使用率为第一优先级,时延为第二优先级,丢包率为第三优先级。当负载和/或带宽使用率低于对应预设阈值时,则可判断该出口网络加速节点可用。或者,当第二优先级及以上的参数低于其对应预设阈值时,即负载、带宽使用率和时延均低于其相应预设阈值时,则可判断该出口网络加速节点可用。
在本发明的一些实施例中,入口网络加速节点也同时是出口网络加速节点,也就是说,该网络加速节点不仅接受用户访问请求,而且是与源站相连的最后一跳网络加速节点。如此一来,在步骤S1中,入口网络加速节点探测到达各个出口网络加速节点的链路时延、丢包率、负载和带宽使用率时,包含探测到达入口网络加速节点自身的链路时延、丢包率、负载和带宽使用率。
在本发明的实施例中,入口网络加速节点与出口网络加速节点之间可以是直接相连的关系,也可以是通过一个或多个其他网络加速节点相连的关系。也就是说,入口网络加速节点到达各个出口网络加速节点的链路中可以是连接了一个或多个中间网络加速节点,也可以是没有连接其他网络加速节点的直接连接。在这种架构下,本发明的实施例中,对入口网络加速节点探测到出口网络节点的链路时延有两具体实施方式,方式一:中间网络加速节点对其下一跳节点的链路时延、丢包率、负载及带宽使用率发送给入口网络加速节点,入口网络加速节点根据上述判断方法对该中间网络加速节点的下一跳节点是否可用,也就是说,在该实施方式中,中间网络加速节点和出口网络加速节点的可用性判断都由入口网络加速节点来完成;方式二:入口网络加速节点或中间网络加速节点先对到达下一跳的网络加速节点的链路时延、丢包率、负载及带宽使用率进行探测,并通过上述方法对该下一跳网络加速节点进行可用判断,若可用则向该可用网络加速节点发出探测请求,该可用网络加速节点继续对其下一跳网络加速节点进行探测和判断,直到出口网络加速节点,因此,入口网络加速节点到达出口网络加速节点的时延为中间各段链路时延之和,且各相关节点的可用性判断任务也分散在各中间节点中,从而降低入口网络加速节点的任务量,实现负载均衡。
步骤S2、入口网络加速节点从可用的出口网络加速节点获取该出口网络加速节点到达源站IP的链路时延(RTT2)。
具体地,步骤S2包括以下步骤:
21)入口网络加速节点仅向可用的出口网络加速节点发起针对源站IP的探测请求;
22)入口网络加速节点接收出口网络加速节点发送来的该出口网络加速节点到达源站IP的链路时延(RTT2)。
其中,出口网络加速节点接收到针对源站IP的探测请求时,探测从自身到达源站IP的链路时延(RTT2),并将探测得到的链路时延(RTT2)返回给入口网络加速节点。也就是说,出口网络加速节点仅在接收到探测请求后,才探测自身到源站IP的链路时延。
步骤S3、入口网络加速节点根据自身到达可用的出口网络加速节点的链路时延,以及可用的出口网络加速节点到达源站IP的链路时延,选取入口网络加速节点到达源站IP的最优路径。
具体地,入口网络加速节点计算自身到达可用的出口网络加速节点、可用的出口网络加速节点到达源站IP的总链路时延,选取总链路时延最小的一条路径作为最优路径。
优选地,获取入口网络加速节点到达出口网络加速节点的链路时延、以及出口网络加速节点到达源站IP的链路时延时,采用平均值法计算,对最近几次或者一个时间周期内的链路 时延进行计算平均值,以作为该链路的时延数据,通过这种方式,能够减少因网络链路抖动所产生异常数据对路径选择的影响。
下面通过具体实施例来阐述本发明的路径选取方法。请参照图2,图2绘示本发明一实施例所提供的路径选取方法过程示意图。如图2所示,网络加速节点A、B、C和D中,A为入口网络加速节点,直接接收用户U的访问请求;B、C和D为距离源站X仅有一跳的出口网络加速节点。A针对到用户U的访问请求进行路径选取时,首先分别探测到达B、C和D的链路时延RTTB1、RTTC1和RTTD1,并根据获取到的到达B、C、D的链路时延、丢包率、负载、带宽使用率及相应的预设阈值,判断B、C和D是否为可用的出口网络加速节点。若B的负载高于预设阈值,则判断B为不可用。因此,A仅向C和D发出探测请求,C和D接收到A发出的探测请求后,探测自身到达源站X的链路时延RTTC2和RTTD2,并将探测结果返回至A。A分别计算RTTC1与RTTC2之和、RTTD1与RTTD2之和,最后选择较小的RTTC1与RTTC2之和所对应的路径AC作为路径探测的结果,即选择入口网络加速节点A-出口网络加速节点C-源站X作为用户U访问源站X的最优路径。
优选地,所获取的RTTB1、RTTC1、RTTC2、RTTD1和RTTD2采用平均值法计算得出,即对最近几次或者一个时间周期内的相应链路时延进行计算平均值,以作为该链路的时延数据。
本发明的实施例中,更包含保存根据上述路径选取方法所获得的最优路径的步骤,在同一个探测周期内,同一用户向同一源站IP发出访问请求时,先查询保存的最优路径中是否有相应的记录,若有,则直接选取该相应路径为本次访问路径,若没有相应记录,则按照上述方法进行最优路径的探测。该些保存的最优路径记录保存时间为一个探测周期,当前探测周期结束后将被清除,并继续存入新的探测周期中探测得到的最优路径,所述清除的方法可以是直接将记录删除,也可以是将其隐藏起来,本发明并不做限制。
由此可见,该最优路径是在当前探测周期内,同一用户向同一源站IP发出访问请求时的最优路径,也就是说,在同一个探测周期内,当同一用户再次访问该源站IP时,访问请求将根据先前所探测到的最优路径,从入口网络加速节点传输到源站IP,源站IP按照原路进行回源,从而保证用户请求能得到最快的传输和反馈。当新的周期开始时,则需要重新进行探测。
在本发明的实施例中,入口网络加速节点将源站IP根据网络运营商和地理位置进行分组,将属于同一运营商、同一地理区域的源站IP归为同一IP分组,并针对每一个IP分组采用一个路径列表保存该IP分组中的源站IP使用过的每一条路径及其使用次数。优选地,在路径列表中,按照该IP分组使用过的每一条路径的使用次数从大到小的顺序对路径进行排序。每 当网络加速节点完成针对一个源站IP的路径选择时,就在该源站IP所属的IP分组的路径列表中,为所选取的最优路径的使用次数加1。
当入口网络加速节点接收到用户针对某个源站IP的访问请求时,若尚未针对该源站IP进行路径选取(即在所保存的最优路径记录中无法查询到相应记录),则选取该源站IP所在IP分组对应的路径列表中路径使用次数最高的路径作为本次访问请求的路径,并同时根据上述路径选取方法针对本次访问请求进行路径选择,保存探测结果,以供后续相同用户的相同访问请求使用。
若既未对该源站IP进行路径选取,也没有对应的IP分组记录,则本次访问路径根据预先制定的默认路径进行传输和回源,并同时根据上述路径选取方法,针对本次访问请求进行路径选择,保存探测结果。
如此一来,在一个探测周期中首次出现的用户访问请求时,入口网络加速节点可以根据路径列表或默认路径及时地选择一个较佳的路径进行传输和回源,保证了用户访问请求传输和反馈的稳定性,且在一定程度上保证了传输路径的质量。
本发明的实施例中,入口网络加速节点上生成有历史路径列表,该历史路径列表用于保存用户IP到达源站IP的历史路径,并为每个历史路径配置有效时间。在有效时间内,针对同一用户IP对同一源站IP的访问请求,采用历史路径列表中所保存的该用户IP到该源站IP的历史路径,并重置有效时间。
具体地,在一些情况下,如基于TCP协议类型的请求,源站要求在同一个会话期间,请求的用户IP不可以发生变化,这就要求同一用户IP到达源站IP的路径在这个会话期间内不可以变化,否则可能会被源站要求重新登录,从而无法建立有效连接。针对这种情况,在网络加速节点上生成一个历史路径列表,用来保存一个用户IP到一个源站IP的历史路径,且每个历史路径都设置有有效时间。在该历史路径有效时间内,若收到同一个用户IP发往同一个源站IP的请求,则采用历史路径列表中所保存的该用户IP到该源站IP的历史路径,并重置有效时间。
本发明所提供的路径选取方法通过入口网络加速节点实现路径选择,且出口网络加速节点只在接收到探测请求时才需对源站IP进行探测。如此一来不仅简化了原有加速网络的架构,而且降低了网络加速节点的探测任务数量及最终的计算选择量,从而释放了探测量过大所占用的宽带。不仅如此,本发明所提供的路径选取方法在选择路径的过程中还综合考虑了链路的时延、丢包率、负载和带宽使用率等综合因素,不仅保证了传输线路的质量,而且改善了各网络加速节点的负载均衡。
更进一步的,通过使用平均值法计算获得链路的时延,可在一定程度上防止因网络抖动对时延数据的影响;通过使用路径表,可保证尚未进行路径探测的用户访问请求可快速分配到较优的路径,避免当次用户访问请求的反馈速率因路径的不确定性而出现异常;通过使用历史路径列表,可保证在同一网络会话过程中所使用的路径的统一性,避免因路劲的不同而出现源站拒绝响应的情况。
此外,本发明还提供一种路径选取系统。请参照图3,图3绘示本发明一较佳实施例所提供的路径选取系统。本实施例所提供的路径选取系统,包括探测模块10、分析处理模块20、收发模块30、运算处理模块40,其中探测模块10和收发模块30分别连接其他三个模块。该路径选取系统应用在各网络加速节点上。
探测模块10,用于探测网络加速节点自身到达其他网络加速节点或到达源站IP的链路时延,以及获取自身到达其他网络加速节点的丢包率和其他网络加速节点的负载和带宽使用率,并将探测结果发送给分析处理模块20或收发模块30。具体地,若该网络加速节点为入口网络加速节点,在其接收到用户访问请求时,探测模块10对到达出口网络加速节点的链路进行时延、丢包率、负载和带宽使用率进行探测,并将探测结果发送给分析处理模块20;若该网络加速节点为出口网络加速节点,当该出口网络加速节点接收到来自入口网络加速节点的探测请求时,探测模块10对到达源站IP的链路进行时延探测,并将探测结果发送给收发模块30,收发模块30将探测结果返回给请求节点;若该网络加速节点为入口网络加速节点到达出口网络加速节点的链路上的中间网络加速节点,在入口网络加速节点对到达出口网络加速节点链路进行探测时,探测模块10对到达下一跳中间网络加速节点的链路进行时延、丢包率、负载和带宽使用率的探测,并将探测结果发送给收发模块30。
分析处理模块20,入口网络加速节点的分析模块20接收来自探测模块10的探测结果,基于所述探测结果判断出口网络加速节点是否可用,并将判断结果发送至收发模块20。具体地,分析处理模块20判断出口网络加速节点是否可用包括以下步骤:
11)接收来自探测模块10的自身到达出口网络加速节点的链路的丢包率、负载、带宽使用率;
12)当链路的时延、丢包率、负载和带宽使用率均低于相应的预设阈值时,判断该出口网络加速节点可用。
本发明中,链路的时延、丢包率、负载和带宽使用率都有相应的预设阈值,该些预设阈值可根据各网络加速节点的实际状态和实际需求来进行预设,预设阈值保存在各网络加速节点中,可以以不可修改的形式保存,也可以以可修改的形式进行保存。
值得注意的是,本实施例中,在判断出口网络加速节点是否可用时,综合考虑了链路时延、丢包率、负载和带宽使用率是否均低于相应的预设阈值,而在本发明的其他实施例中,可根据实际需求仅考虑其中一项或多项是否低于相应阈值。另外,链路时延、丢包率、负载和带宽使用率可分别具有相应的优先级。当一定优先级的参数低于其预设阈值时,则可判断该出口网络加速节点可用。例如,负载和带宽使用率为第一优先级,时延为第二优先级,丢包率为第三优先级。当负载和/或带宽使用率低于对应预设阈值时,则可判断该出口网络加速节点可用。或者,当第二优先级及以上的参数低于其对应预设阈值时,即负载、带宽使用率和时延均低于其相应预设阈值时,则可判断该出口网络加速节点可用。
本发明的一些实施例中,入口网络加速节点与出口网络加速节点之间还连接有至少一个中间网络加速节点,在这种情况下,入口网络加速节点在判断出口网络加速节点是否可用时,包含判断该些中间网络加速节点是否可用,主要有以下两种实施方式,方式一:中间网络加速节点将其探测模块10的对下一跳节点的探测结果通过收发模块30发送给入口网络加速节点,入口网络加速节点根据上述判断方法对该中间网络加速节点的下一跳节点是否可用进行判断,也就是说,在该实施方式中,中间网络加速节点和出口网络加速节点的可用性判断都由入口网络加速节点来完成;方式二:中间网络加速节点的分析处理模块20接收探测模块10的探测结果并判断该中间网络加速节点的下一跳网络加速节点是否可用,也就是说,在该实施方式中,各网络加速节点的可用性判断均有其上一跳网络加速节点来完成,如此一来,可分散任务处理量,更好实现负载均衡。
运算处理模块40,根据分析处理结果,从探测模块10获取自身到达可用出口网络加速节点链路的时延、从收发模块30获取可用出口网络加速节点到达源站IP的链路时延,并计算两者之和,选取时延最小的路径作为路径选取结果,并将路径选取结果发送给收发模块30。值得注意的是,本实施例中,运算处理模块40是通过收发模块30获取分析结果,而本发明的其他实施例中,运算处理模块40可直接从分析处理模块20获取分析处理结果,或通过其他方式获得分析处理结果,本发明并不以此为限。
收发模块30,包含接收单元31和发送单元32,其中接收单元31接收来自网络加速节点内部及外部的数据,包含来自外部的用户访问请求和其他网络加速节点发送来的探测请求及探测结果,来自内部探测模块10的探测结果、分析处理模块20的分析处理结果以及运算处理模块40的路径选取结果,发送单元32向网络加速节点外部发送相关数据,包含向可用的入口网络加速节点发送探测请求、向其他网络加速节点发送探测请求和路径选取结果。
进一步的,本实施例中,路径选取系统更包含一存储模块50和查询模块60。存储模块 50连接于收发模块30,从收发模块30中获取路径选取结果,并将该些路径选取结果保存在路径选取列表L1中,包含用户信息、源站IP及路径,该路径选取列表在一个周期内有效,周期结束后将被清空,并存入新周期所发生的路径选取结果。查询模块60在接收到用户访问请求时,先对存储模块中的路径选取列表L1查询,若存在同一用户访问同一源站的相应记录,则选取对应的路径进行传输和回源。
当路径选取列表L1中没有相应的记录时,需及时为当前用户访问请求分配一个有效路径,在本发明的实施例中,存储模块50中更包含路径选取列表L2,将属于同一运营商、同一地理区域的源站IP归为同一IP分组,并针对每一个IP分组采用一个路径列表保存该IP分组中的源站IP使用过的每一条路径及其使用次数。当入口网络加速节点接收到用户发来的针对某个源站IP的访问请求时,查询模块60在路径选取列表L1中无法查找到相应记录,则选取该源站IP所在IP分组对应的路径选取列表L2中路径使用次数最高的路径作为本次访问请求的路径。
在一些情况下,源站要求在同一个会话期间,请求的用户IP不可以发生变化,这就要求同一用户IP到达源站IP的路径在一段时间内不可以变化,否则可能会被源站要求重新登录。针对这种情况,本发明的实施例中,存储模块50中更保存有历史路径列表L3,历史路径列表L3保存一个用户IP到一个源站IP的历史路径,且每个历史路径都设置有有效时间。在该历史路径有效时间内,若收到同一个用户IP发往同一个源站IP的请求,查询模块60首先对历史路径列表L3进行查询,若查询到相应记录,且处于有效时间内,则采用历史路径列表中所保存的该用户IP到该源站IP的历史路径,并重置有效时间,从而保证了同一个会话期间同一个用户访问请求按照相同路径进行传输。
本发明实施例中所提供的路径选取系统均运行在网络加速节点上,也就是说,本发明实施例中的网络加速节点在不同的用户访问请求会话期间,既可是入口网络加速节点,也可以是出口网络加速节点,或中间网络加速节点,当网络加速节点的定位不同时,系统各模块的功能也会进行相应调整。如此一来,各网络加速节点所构成的网络架构简单、灵活,可适用于不同的应用场景,从而更好地满足实际需求。
本发明还提供一种网络加速节点,其采用上述的路径选取系统选取自身到达源站IP的最优路径。
参照图4,图4绘示本发明一实施所提供的网络加速系统结构示意图。如图4所示,网络加速系统,包括若干上述网络加速节点22,以使用户21通过网络加速节点选取最优路径访问相应的源站23IP。
具体地,用户发送访问请求至对应的入口网络加速节点,该入口网络加速节点的路径选取系统对路径进行选取,记录由该网络加速节点至源站IP的最优路径,使得用户能够基于该最优路径实现对源站IP的访问,从而加快了网络访问速度,提供了用户体验。
综上所述,本发明的路径选取方法及系统、网络加速节点、网络加速系统由网络加速节点自身来计算到源站的最优路径,无需中央服务器介入,加快了路径选取的效率;仅向可用网络加速节点发起路径探测,降低了路径探测数量,优化了路径选取的质量;能够基于在前路径选取结果进行后续路径选取,提高了路径选取效率。所以,本发明有效克服了现有技术中的种种缺点而具高度产业利用价值。
上述实施例仅例示性说明本发明的原理及其功效,而非用于限制本发明。任何熟悉此技术的人士皆可在不违背本发明的精神及范畴下,对上述实施例进行修饰或改变。因此,举凡所属技术领域中具有通常知识者在未脱离本发明所揭示的精神与技术思想下所完成的一切等效修饰或改变,仍应由本发明的权利要求所涵盖。

Claims (16)

  1. 一种路径选取方法,其特征在于,包括以下步骤:
    入口网络加速节点探测到达各个出口网络加速节点的链路时延,并判断所述各个出口网络加速节点是否可用,其中所述入口网络加速节点用于接收用户访问请求;
    所述入口网络加速节点从可用出口网络加速节点获取该出口网络加速节点到达源站IP的链路时延;
    所述入口网络加速节点根据自身到达所述可用出口网络加速节点的链路时延,以及所述可用出口网络加速节点到达所述源站IP的链路时延,选取所述用户访问请求到达所述源站IP的最优路径。
  2. 根据权利要求1所述的路径选取方法,其特征在于,所述入口网络加速节点判断所述出口网络加速节点是否可用包括以下步骤:
    所述入口网络加速节点探测并获取到达所述出口网络加速节点的链路的时延、丢包率、负载及带宽使用率;
    当所述到达所述出口网络加速节点的链路的时延、丢包率、负载及带宽使用率中的一项或者多项低于相应预设阈值时,则所述出口网络加速节点可用。
  3. 根据权利要求1所述的路径选取方法,其特征在于,所述入口网络加速节点计算自身到达所述可用出口网络加速节点、所述可用的出口网络加速节点到达所述源站IP的总链路时延,选取总链路时延最小的一条路径作为最优路径。
  4. 根据权利要求1所述的路径选取方法,其特征在于,所述入口网络加速节点到达所述出口网络加速节点的链路时延、以及所述出口网络加速节点到达所述源站IP的链路时延,采用平均值法计算得到。
  5. 根据权利要求1所述的路径选取方法,其特征在于,包含一路径保存步骤,保存所述最优路径。
  6. 根据权利要求5所述的路径选取方法,其特征在于,在所述探测步骤之前更包含一查询步骤,根据所述用户访问请求信息查询所保存的所述最优路径,若存在相应记录,则直接选取该路径作为本次访问请求的路径。
  7. 根据权利要求1所述的路径选取方法,其特征在于,所述入口网络加速节点将属于同一运营商、同一地理区域的源站IP归为同一IP分组,并针对每一个所述IP分组采用一个路径列表保存所述IP分组使用过的每一条路径及其使用次数;当所述入口网络加速节点接收到用户针对某个所述源站IP的访问请求时,若尚未针对该所述源站IP进行路径选取,则选取所述源站IP所在所述IP分组对应的所述路径列表中路径使用次数最高的作为本次访问请求的路径。
  8. 根据权利要求1所述的路径选取方法,其特征在于,所述入口网络加速节点上生成历史路径列表,所述历史路径列表用于保存用户IP到达源站IP的历史路径,且每个所述历史路径设有有效时间;在有效时间内,针对同一用户IP对同一源站IP的访问请求,采用所述历史路径列表中所保存的所述用户IP到所述源站IP的历史路径,并重置有效时间。
  9. 一种路径选取系统,其特征在于,包括探测模块、分析处理模块、收发模块、运算处理模块;
    所述探测模块探测网络加速节点自身到达其他网络加速节点或源站IP的链路时延和丢包率,并获取所述其他网络加速节点的负载和带宽使用率,并将探测结果发送至所述分析处理模块或所述收发模块;
    所述分析处理模块接收来自所述探测模块和或所述收发模块发送而来的探测结果,基于所述探测结果判断出口网络加速节点是否可用,并将分析处理结果发送至所述收发模块;所述出口网络加速节点为与源站相连的最后一跳网络加速节点;
    所述运算处理模块从所述探测模块获取所述网络加速节点自身到达可用出口网络加速节点链路的时延、从所述收发模块获取可用出口网络加速节点到达源站IP的链路时延,并计算两者之和,选取时延最小的作为路径选取结果,并将所述路径选取结果发送至所述收发模块;
    所述收发模块接收用户访问请求、来自所述探测模块探测结果、所述分析处理模块的判断结果以及所述运算处理模块的路径选取结果和其他所述网络加速节点发送的探测请求,发送探测请求、所述探测结果及所述路径选取结果。
  10. 根据权利要求9所述的路径选取系统,其特征在于,所述分析处理模块判断所述出口网络加速节点是否可用包括以下步骤:
    从所述探测模块获取到达所述出口网络加速节点的链路时延、丢包率及负载和带宽使用率;
    当所述链路时延、丢包率、负载及带宽使用率其中一项或多项低于相应预设阈值时,判定所述出口网络加速节点可用。
  11. 根据权利要求9所述的路径选取系统,其特征在于,所述探测模块探测入口网络加速节点到达所述出口网络加速节点的链路时延、到达所述源站IP的链路时延时,通过平均值法计算得出。
  12. 根据权利要求9所述的路径选取系统,其特征在于,还包含存储模块,所述存储模块与所述收发模块相连,用于从所述收发模块接收所述路径选取结果,并将所述路径选取结果存储在路径选取列表;当接收到用户访问请求时,先在所述路径选取列表中进行查询,若存在同一用户IP访问同一源站IP的记录,则选取对应的路径进行传输和回源。
  13. 根据权利要求12所述的路径选取系统,其特征在于,还包括存储模块,存有路径列表,所述路径列表中对属于同一运营商、同一地理区域的源站IP归为同一IP分组,记录每一个所述IP分组中的所述源站IP使用过的每一条路径及其使用次数;当接收到用户针对某个源站IP的访问请求时,若在所述路径选取列表中无法查找到相应记录,则选取所述路径列表中所述源站IP所在IP分组对应的路径使用次数最高的路径作为本次访问请求的路径。
  14. 根据权利要求9所述的路径选取系统,其特征在于,还包括存储模块,所述存储模块存有历史路径列表,所述历史路径列表用于存储用户IP到源站IP的历史路径,及每个所述历史路径的有效时间;在所述历史路径有效时间内,若收到同一个用户IP发往同一个源站IP的请求,则采用所述历史路径列表中所保存的该用户IP到该源站IP的历史路径,并重置有效时间。
  15. 一种网络加速节点,其特征在于,采用权利要求9至14中任一项所述的路径选取系统选取自身到达源站IP的最优路径。
  16. 一种网络加速系统,其特征在于,包括若干权利要求15所述的网络加速节点,以使 用户通过所述网络加速节点选取最优路径访问相应的源站IP。
PCT/CN2017/078398 2017-02-21 2017-03-28 一种路径选取方法及系统、网络加速节点及网络加速系统 WO2018152919A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
US16/326,779 US10951733B2 (en) 2017-02-21 2017-03-28 Route selection method and system, network acceleration node, and network acceleration system
EP17897430.9A EP3528440B1 (en) 2017-02-21 2017-03-28 Path selection method and system, network acceleration node, and network acceleration system

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710093502.7A CN106656800B (zh) 2017-02-21 2017-02-21 一种路径选取方法及系统、网络加速节点及网络加速系统
CN201710093502.7 2017-02-21

Publications (1)

Publication Number Publication Date
WO2018152919A1 true WO2018152919A1 (zh) 2018-08-30

Family

ID=58845763

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/078398 WO2018152919A1 (zh) 2017-02-21 2017-03-28 一种路径选取方法及系统、网络加速节点及网络加速系统

Country Status (4)

Country Link
US (1) US10951733B2 (zh)
EP (1) EP3528440B1 (zh)
CN (1) CN106656800B (zh)
WO (1) WO2018152919A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112910949A (zh) * 2021-01-12 2021-06-04 广州虎牙科技有限公司 跨区域的网络访问方法、装置、通讯设备、存储介质
EP3869750A4 (en) * 2018-10-25 2021-12-22 Wangsu Science & Technology Co., Ltd. DATA TRAFFIC TRANSMISSION APPARATUS AND METHOD
WO2023040366A1 (zh) * 2021-09-14 2023-03-23 天翼数字生活科技有限公司 一种基于icmp的vpn加速通道网络质量探测系统

Families Citing this family (34)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107171955A (zh) * 2017-05-24 2017-09-15 深圳市乃斯网络科技有限公司 网络中数据链路的选择方法及系统
WO2018214059A1 (zh) * 2017-05-24 2018-11-29 深圳市乃斯网络科技有限公司 网络中数据链路的选择方法及系统
CN107277118B (zh) * 2017-05-31 2020-06-30 北京京东尚科信息技术有限公司 生成节点的常用访问路径的方法和装置
CN109005078B (zh) * 2017-06-06 2021-03-19 腾讯科技(深圳)有限公司 路由策略生成方法、装置及计算机可读存储介质
CN107317879B (zh) * 2017-08-02 2019-07-05 网宿科技股份有限公司 一种用户请求的分发方法及系统
CN110392074B (zh) * 2018-04-19 2022-05-17 贵州白山云科技股份有限公司 一种基于动态加速的调度方法及装置
CN108881057B (zh) * 2018-04-20 2022-08-02 网宿科技股份有限公司 回源线路的选择方法和流量分发器
CN108810558B (zh) * 2018-06-12 2021-03-02 北京奇艺世纪科技有限公司 一种直播回源的方法及系统
US11077365B2 (en) 2018-06-27 2021-08-03 Niantic, Inc. Low latency datagram-responsive computer network protocol
CN108833276B (zh) * 2018-07-03 2021-02-05 网宿科技股份有限公司 优选路径的确定方法及装置、服务器
CN109285317A (zh) * 2018-09-17 2019-01-29 北海华源电子有限公司 基于智能手环和智能手机的安全监测系统
CN109008098A (zh) * 2018-09-17 2018-12-18 北海华源电子有限公司 安全监测系统
CN109118712A (zh) * 2018-09-17 2019-01-01 北海华源电子有限公司 用于对儿童进行监测的安全监测系统
CN109275993A (zh) * 2018-09-17 2019-01-29 北海华源电子有限公司 便携式安全监测系统
CN110971626B (zh) * 2018-09-28 2024-01-19 贵州白山云科技股份有限公司 一种企业分支机构访问请求处理方法、装置及系统
CN111181849B (zh) * 2018-11-09 2022-04-15 北京嘀嘀无限科技发展有限公司 回源路径确定方法、确定装置、计算机设备和存储介质
CN109194772A (zh) * 2018-11-13 2019-01-11 杭州领智云画科技有限公司 Cdn边缘节点的回源方法和系统
CA3130880C (en) 2019-02-25 2024-01-02 Niantic, Inc. Augmented reality mobile edge computing
CN109862115B (zh) * 2019-03-13 2022-09-16 网宿科技股份有限公司 web请求的自动化处理方法及BGP路由系统
CN109787868B (zh) * 2019-03-18 2022-03-01 网宿科技股份有限公司 一种选取路由路径的方法、系统及服务器
CN109981765B (zh) 2019-03-18 2023-03-24 北京百度网讯科技有限公司 用于确定内容分发网络的访问路径的方法和装置
CN110809040B (zh) * 2019-10-30 2022-04-29 四川速宝网络科技有限公司 加速服务提供系统及方法
AU2020409015A1 (en) * 2019-12-20 2022-08-25 Niantic, Inc. Data hierarchy protocol for data transmission pathway selection
CN111600929B (zh) * 2020-04-07 2023-05-19 厦门网宿有限公司 传输线路探测方法、路由策略生成方法及代理服务器
CN111683013B (zh) * 2020-06-08 2022-03-08 腾讯科技(深圳)有限公司 一种加速网络的路由方法和加速网络
CN113950125B (zh) * 2020-07-16 2024-05-17 腾讯科技(深圳)有限公司 一种应用程序中网络加速的方法、装置以及存储介质
CN112637132B (zh) * 2020-12-01 2022-03-11 北京邮电大学 一种网络异常检测方法、装置、电子设备和存储介质
CN112954765B (zh) * 2021-02-24 2022-12-09 四川速宝网络科技有限公司 一种基于网络状态监控的动态算子选路方法
CN114024876B (zh) * 2021-10-15 2023-06-16 中国联合网络通信集团有限公司 一种网络拨测方法、装置、设备及存储介质
CN114448872B (zh) * 2021-12-23 2023-06-23 天翼云科技有限公司 一种cdn中的路径选择方法及系统
CN114520784B (zh) * 2021-12-24 2023-10-10 天翼云科技有限公司 一种动态内容加速访问方法及装置
CN115242669B (zh) * 2022-06-30 2023-10-03 北京华顺信安科技有限公司 一种网络质量监测方法
CN115277539A (zh) * 2022-07-29 2022-11-01 天翼云科技有限公司 一种数据传输方法、选路集群以及边缘节点
CN116094986A (zh) * 2023-01-05 2023-05-09 中国联合网络通信集团有限公司 一种业务保障方法、装置、电子设备及存储介质

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2004073269A1 (ja) * 2003-02-13 2004-08-26 Fujitsu Limited 伝送システム,配信経路制御装置,負荷情報収集装置および配信経路制御方法
CN104618237A (zh) * 2015-01-21 2015-05-13 网宿科技股份有限公司 一种基于tcp/udp的广域网加速系统及方法
CN104954219A (zh) * 2014-03-27 2015-09-30 北京奇虎科技有限公司 Cdn网络访问路径优化方法、装置及系统
CN105515915A (zh) * 2015-12-25 2016-04-20 厦门网宿软件科技有限公司 节点探测方法及装置、路径选取方法及装置、及网络系统
CN105657006A (zh) * 2015-12-29 2016-06-08 厦门网宿软件科技有限公司 一种基于上网加速网络的首次访问加速方法和系统
CN105704034A (zh) * 2016-04-05 2016-06-22 网宿科技股份有限公司 接入网络的方法和中央服务器

Family Cites Families (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7027426B2 (en) 2002-08-05 2006-04-11 Harris Corporation Multi-channel mobile ad hoc network
US8065559B2 (en) * 2008-05-29 2011-11-22 Citrix Systems, Inc. Systems and methods for load balancing via a plurality of virtual servers upon failover using metrics from a backup virtual server
US8040808B1 (en) * 2008-10-20 2011-10-18 Juniper Networks, Inc. Service aware path selection with a network acceleration device
US9071541B2 (en) * 2012-04-25 2015-06-30 Juniper Networks, Inc. Path weighted equal-cost multipath
CN103051709A (zh) * 2012-12-20 2013-04-17 新浪网技术(中国)有限公司 数据传输路径确定方法、网络节点及内容分发网络系统
US10148544B2 (en) * 2013-04-06 2018-12-04 Citrix Systems, Inc. Systems and methods for cluster statistics aggregation
CN104168300B (zh) * 2013-05-17 2017-06-27 中国电信股份有限公司 内容加速方法与系统
US9749148B2 (en) * 2013-07-31 2017-08-29 Citrix Systems, Inc. Systems and methods for load balancing non-IP devices
CN103532817B (zh) * 2013-10-12 2017-01-18 无锡云捷科技有限公司 一种cdn动态加速的系统及方法
CN105897581A (zh) * 2015-11-20 2016-08-24 乐视云计算有限公司 Cdn网络中的媒体文件回源路径的确定方法及装置
US10305807B2 (en) * 2016-05-03 2019-05-28 Citrix Systems, Inc. Systems and methods to choose an optimal path from multiple high latency links
US10455053B2 (en) * 2016-05-23 2019-10-22 Citrix Systems, Inc. Systems and methods for provisioning network automation by logically separating L2-L3 entities from L4-L7 entities using a software defined network (SDN) controller
CN106357792B (zh) * 2016-10-10 2019-09-06 网宿科技股份有限公司 节点选路方法及系统

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2004073269A1 (ja) * 2003-02-13 2004-08-26 Fujitsu Limited 伝送システム,配信経路制御装置,負荷情報収集装置および配信経路制御方法
CN104954219A (zh) * 2014-03-27 2015-09-30 北京奇虎科技有限公司 Cdn网络访问路径优化方法、装置及系统
CN104618237A (zh) * 2015-01-21 2015-05-13 网宿科技股份有限公司 一种基于tcp/udp的广域网加速系统及方法
CN105515915A (zh) * 2015-12-25 2016-04-20 厦门网宿软件科技有限公司 节点探测方法及装置、路径选取方法及装置、及网络系统
CN105657006A (zh) * 2015-12-29 2016-06-08 厦门网宿软件科技有限公司 一种基于上网加速网络的首次访问加速方法和系统
CN105704034A (zh) * 2016-04-05 2016-06-22 网宿科技股份有限公司 接入网络的方法和中央服务器

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3528440A4 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3869750A4 (en) * 2018-10-25 2021-12-22 Wangsu Science & Technology Co., Ltd. DATA TRAFFIC TRANSMISSION APPARATUS AND METHOD
CN112910949A (zh) * 2021-01-12 2021-06-04 广州虎牙科技有限公司 跨区域的网络访问方法、装置、通讯设备、存储介质
CN112910949B (zh) * 2021-01-12 2022-05-17 广州虎牙科技有限公司 跨区域的网络访问方法、装置、通讯设备、存储介质
WO2023040366A1 (zh) * 2021-09-14 2023-03-23 天翼数字生活科技有限公司 一种基于icmp的vpn加速通道网络质量探测系统

Also Published As

Publication number Publication date
EP3528440A4 (en) 2019-10-30
CN106656800B (zh) 2020-04-14
US20190182351A1 (en) 2019-06-13
EP3528440B1 (en) 2021-09-08
EP3528440A1 (en) 2019-08-21
CN106656800A (zh) 2017-05-10
US10951733B2 (en) 2021-03-16

Similar Documents

Publication Publication Date Title
WO2018152919A1 (zh) 一种路径选取方法及系统、网络加速节点及网络加速系统
CN111181849B (zh) 回源路径确定方法、确定装置、计算机设备和存储介质
JP5048134B2 (ja) パケットルーティングの方法、システム、装置、及びバックアップリソースを選択する方法、システム
JP7313480B2 (ja) スライスベースネットワークにおける輻輳回避
CN109040243B (zh) 一种报文处理方法及装置
CN109151009B (zh) 一种基于mec的cdn节点分配方法和系统
US8948191B2 (en) Intelligent traffic quota management
CN112087382B (zh) 一种服务路由方法及装置
CN104754640A (zh) 一种网络资源调度方法及网络资源管理服务器
US20240259475A1 (en) Service scheduling method and apparatus, device, and computer-readable storage medium
CN110601989A (zh) 一种网络流量均衡方法及装置
US20220070736A1 (en) Traffic steering device
US11290379B2 (en) Egress traffic steering controller
US12015665B2 (en) Network optimization system using server latency measurements
WO2024032011A1 (zh) Cdn调度方法、cdn调度系统、存储介质
WO2011110104A1 (zh) 终端接入方法、设备及系统
CN110380981B (zh) 一种流量分发方法及设备
KR101720710B1 (ko) 무선 메쉬 네트워크에서 sdn 기반 정보 캐시 및 캐시 대상 라우터 선택 방법
Seyyed Hashemi et al. Analytical characterization of cache replacement policy impact on content delivery time in information‐centric networks
CN107404438A (zh) 网络路由方法和网络路由系统
Bays et al. Flow based load balancing: Optimizing web servers resource utilization
CN106254576B (zh) 一种报文转发方法及装置
CN115665262A (zh) 一种请求处理方法、装置、电子设备及存储介质
CN109451074B (zh) 一种基于portal协议的服务器负载均衡处理办法
CN109691181B (zh) 用于传输网络中的分组流优化的方法、流控制实体和承载控制实体

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17897430

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2017897430

Country of ref document: EP

Effective date: 20190517

NENP Non-entry into the national phase

Ref country code: DE