Method and apparatus for determining network performance state, electronic device, and storage medium

Abstract

The application provides a network performance state determination method and device, electronic equipment and storage medium, and relates to the technical field of computing power networks. The method is applied to a first computing power gateway and comprises the following steps: measuring first network performance states between the first computing power gateway and at least one second computing power gateway; receiving second network performance states between each second computing power gateway and each computing power service instance sent by each second computing power gateway; determining comprehensive network performance states between the first computing power gateway and each computing power service instance based on the first network performance states and the second network performance states; and using the comprehensive network performance states to determine optimal computing power routing of a user access request. The application reduces the number of network detection sessions of the user-side computing power gateway through a segmented network state detection mechanism, reduces the network burden in the computing power network domain, and further improves the efficiency of detecting network performance states.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computing power network, and particularly relates to a network performance state determination method and device, electronic equipment and storage medium. BACKGROUND

[0002] The computing power network is a key bridge connecting users and computing power resources. When the computing power network schedules access to computing power services for users, the network performance state contained in the computing power routing needs to be referred to in combination with the requirements of the users on the network access performance, so as to select appropriate computing power service instances that meet the requirements of the users for the users.

[0003] However, the existing network performance state is directly initiated by the user-side computing power gateway to a large number of computing power service instances for network detection. The large number of computing power service instances increases the number of network detection sessions of the user-side computing power gateway, and a large number of network detection packets increase the network burden, further reducing the efficiency of detecting the network performance state. SUMMARY

[0004] The present application provides a network performance state determination method and device, electronic equipment and storage medium, to solve the defect of low efficiency of detecting the network performance state in the prior art, and to reduce the number of network detection sessions of the user-side computing power gateway by using a segmented network state detection mechanism, reduce the network burden in the computing power network domain, and further improve the efficiency of detecting the network performance state.

[0005] The present application provides a network performance state determination method applied to a first computing power gateway, comprising:

[0006] The first network performance state between the first computing power gateway and at least one second computing power gateway is measured respectively; the first computing power gateway is a computing power gateway connected to a user terminal, and the second computing power gateway is a computing power gateway connected to a computing power resource pool providing computing power resources;

[0007] The second network performance state between each second computing power gateway and each computing power service instance is received, which is sent by each second computing power gateway; the computing power service instance is deployed in each computing power resource in the computing power resource pool connected to the second computing power gateway;

[0008] Based on the first network performance state and the second network performance state, the comprehensive network performance state between the first computing power gateway and each computing power service instance is determined; wherein the comprehensive network performance state is used to determine the optimal computing power routing of the user access request.

[0009] According to the network performance state determination method provided by the application, the first network performance state between the first computing power gateway and at least one second computing power gateway is measured, including: sending at least one detection request packet to each second computing power gateway; receiving a response packet corresponding to each detection request packet sent by each second computing power gateway; determining the first network performance state based on the detection request packet and the response packet corresponding to the detection request packet.

[0010] According to the network performance state determination method provided by the application, the first network performance state between the first computing power gateway and at least one second computing power gateway is measured, including: sending at least one detection request packet to each second computing power gateway; receiving a response packet corresponding to each detection request packet sent by each second computing power gateway; determining the first network performance state based on the detection request packet and the response packet corresponding to the detection request packet.

[0011] According to the network performance state determination method provided by the application, the first network performance state includes: packet loss rate and average delay; determining the packet loss rate based on the detection request packet and the response packet corresponding to the detection request packet, including: determining the first number of detection request packets and the second number of response packets; determining the packet loss rate based on the first number and the second number; determining the average delay based on the detection request packet and the response packet corresponding to the detection request packet, including: determining the time difference between sending each detection request packet and receiving the response packet corresponding to the detection request packet; determining the average delay based on each time difference.

[0012] According to the network performance state determination method provided by the application, the first network performance state also includes: delay jitter; determining the delay jitter, including: determining the minimum time difference and the maximum time difference from each time difference respectively; determining the delay jitter based on the minimum time difference, the maximum time difference and the average delay.

[0013] According to a method for determining a network performance status provided by the present invention, the method also includes: updating the first network performance status based on a first change value to obtain a third network performance status; the first change value is used to represent the change value between the first network performance status in the current time period and the first network performance status in the previous time period; receiving a fourth network performance status; the fourth network performance status is obtained by the second computing power gateway updating the second network performance status based on the second change value; the second change value is used to represent the change value between the second network performance status in the current time period and the second network performance status in the previous time period; based on the third network performance status and the fourth network performance status, recalculating the comprehensive network performance status from the first computing power gateway to each of the computing power service instances to obtain an updated comprehensive network performance status.

[0014] According to a method for determining a network performance status provided by the present invention, the second network performance status between each second computing power gateway and each computing power service instance sent by each second computing power gateway is received to obtain the second network performance status, including: establishing a neighbor relationship with each second computing power gateway respectively; using each of the neighbor relationships, receiving a second network performance status message sent by the second computing power gateway corresponding to each of the neighbor relationships; parsing the second network performance status message to obtain the second network performance status.

[0015] The present invention also provides a device for determining network performance status, comprising:

[0016] A measurement module, configured to respectively measure a first network performance status between a first computing power gateway and at least one second computing power gateway; the first computing power gateway is a computing power gateway connected to a user terminal, and the second computing power gateway is a computing power gateway connected to a computing power resource pool that provides computing power resources;

[0017] A receiving module, configured to receive a second network performance status between each second computing power gateway and each computing power service instance sent by each second computing power gateway; the computing power service instance is deployed in each computing power resource in the computing power resource pool connected to the second computing power gateway;

[0018] A determination module is used to determine the comprehensive network performance status between the first computing power gateway and each of the computing power service instances based on the first network performance status and the second network performance status; wherein the comprehensive network performance status is used to determine the optimal computing power route for user access requests.

[0019] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for determining the network performance status as described above is implemented.

[0020] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for determining the network performance status.

[0021] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above methods for determining the network performance status.

[0022] The method, device, electronic device and storage medium for determining the network performance status provided by the present invention respectively measure the first network performance status between the first computing power gateway and each second computing power gateway through the first computing power gateway, wherein the first computing power gateway is the computing power gateway connected to the user terminal, and the second computing power gateway is the computing power gateway connected to the computing power device providing computing power resources; receive the second network performance status between each second computing power gateway and each computing power service instance, wherein the comprehensive network performance status is used to determine the optimal computing power route for the user access request; according to the first network performance status and the second network performance status, obtain the end-to-end comprehensive network performance status to provide the optimal computing power route for the user access request. In this way, by detecting the network performance status in a segmented manner, compared with the network detection directly initiated from the user-side computing power gateway to a large number of computing power service instances, the number of network detection sessions of the first computing power gateway is reduced, the bandwidth occupied by the network detection messages in the computing power network domain is reduced, the network burden is reduced, and the efficiency and pertinence of the network performance status detection are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0024] Figure 1 This is one of the flow charts of the method for determining the network performance status provided by the present invention.

[0025] Figure 2 This is one of the structural diagrams of the computing power network system provided by the present invention.

[0026] Figure 3 This is the second flow chart of the method for determining the network performance status provided by the present invention.

[0027] Figure 4 This is the second structural diagram of the computing power network system provided by the present invention.

[0028] Figure 5 This is a schematic diagram of the BGP / BGP4+ update message format provided by the present invention.

[0029] Figure 6 It is a structural diagram of the device for determining network performance status provided by the present invention.

[0030] Figure 7 It is a schematic diagram of the physical structure of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0031] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0032] With the widespread adoption of technologies such as cloud computing, big data, and artificial intelligence, computing power, as the cornerstone supporting these technologies, has become increasingly important. The computing network serves as a critical bridge connecting users with computing resources. Network performance directly impacts the performance of various application services and the user experience with computing services. When scheduling computing services for users, the computing network must consider the user's network access performance requirements and refer to the real-time network performance status contained in the computing routes to select a suitable computing service instance that meets their needs. Therefore, using the destination address of the computing service instance contained in the computing route as the endpoint, network status information is generated from the user-side computing gateway device to the destination address of the computing service instance. This provides a basis for computing route selection and has become a key research topic in computing networks.

[0033] However, the existing network performance status is directly initiated by the user-side computing power gateway to a large number of computing power service instances for network detection. The large number of computing power service instances leads to an increase in the number of network detection sessions of the user-side computing power gateway. The large number of network detection messages increases the network burden, further reducing the efficiency of detecting network performance status.

[0034] Based on the above problems, the present invention provides a method for determining the network performance status. By detecting the network performance status in a segmented manner, compared with the network detection directly initiated from the user-side computing power gateway to a large number of computing power service instances, the method reduces the number of network detection sessions of the first computing power gateway, reduces the bandwidth occupied by network detection messages within the computing power network domain, reduces the network burden, and improves the efficiency and pertinence of network performance status detection.

[0035] The following combination Figures 1-4The present invention describes a method for determining network performance status, which can be applied to any computing network. The method can be implemented by an electronic device or a method for determining network performance status implemented within the electronic device. The network performance status determination device can be implemented using software, hardware, or a combination of both.

[0036] Figure 1 This is one of the flow charts of the method for determining the network performance status provided by the present invention, which is applied to the first computing power gateway, such as Figure 1 As shown, the method includes the following:

[0037] Step 101: Measure the first network performance status between a first computing power gateway and at least one second computing power gateway respectively.

[0038] Among them, the first computing power gateway is a computing power gateway connected to the user terminal, and the second computing power gateway is a computing power gateway connected to the computing power resource pool that provides computing power resources.

[0039] Here, the first computing power gateway can be a user-side computing power gateway, and the second computing power gateway can be a computing power-side computing power gateway. The first computing power gateway and the second computing power gateway can be connected through a computing power network. The first computing power gateway is responsible for connecting the user's computing power needs to the computing power network and implementing computing network collaborative scheduling. The second computing power gateway is responsible for accessing and managing computing power resources.

[0040] It should be noted that the first network performance status is used to characterize the network quality between the first computing power gateway and the second computing power gateway. The first network performance status may include at least one of the following information: packet loss rate, average delay, and delay jitter. Among them, the smaller the packet loss rate, the lower the average delay, and the smaller the delay jitter, the better the network quality; conversely, the worse the network quality.

[0041] The first computing power gateway may measure the performance status of the first network in any appropriate manner, such as sending a detection request message, a traceroute test method, a carrier monitoring and response mechanism, and the like.

[0042] Step 102: Receive the second network performance status between each second computing power gateway and each computing power service instance sent by each second computing power gateway.

[0043] The computing service instance is deployed in each computing resource in a computing resource pool connected to the second computing gateway. Examples include banking services and image AI generation. Computing resources can include resources such as a central processing unit (CPU) or a graphics processing unit (GPU). Computing resources reside in a computing resource pool, which can include cloud data centers, edge computing nodes, or computer rooms.

[0044] Here, the second network performance status is used to represent the network quality between the second computing power gateway and the computing power service instance. The second network performance status is measured by the second computing power gateway and may include at least one of packet loss rate, average latency, and latency jitter. The method for measuring the first network performance status by the second computing power gateway can refer to the method for measuring the first network performance status and is not further described here.

[0045] It should be noted that the second computing power gateway can send multiple entry information at one time.

[0046] Step 103: Based on the first network performance status and the second network performance status, determine the comprehensive network performance status between the first computing power gateway and each of the computing power service instances.

[0047] The comprehensive network performance status is used to determine the optimal computing power routing for user access requests.

[0048] When a user sends an access request, the First Computing Gateway compares the user's performance requirements (such as low latency and low packet loss rate) with the overall network performance to make a decision and select the optimal computing service instance to respond to the user's service request. If the packet loss rate, average latency, and latency jitter recorded in the computing route from the First Computing Gateway to the destination address of the computing service instance fall below certain indicators, the corresponding computing route will be set to the network unavailable state. The computing service instance pointed to by the computing route will no longer provide services to the user and will not participate in computing route selection.

[0049] In an embodiment of the present invention, the first network performance status between the first computing power gateway and each second computing power gateway is measured respectively through the first computing power gateway, wherein the first computing power gateway is a computing power gateway connected to the user terminal, and the second computing power gateway is a computing power gateway connected to the computing power resource pool that provides computing power resources; the second network performance status between each second computing power gateway and each computing power service instance is received, wherein the comprehensive network performance status is used to determine the optimal computing power route for the user access request; based on the first network performance status and the second network performance status, the end-to-end comprehensive network performance status is obtained to provide the optimal computing power route for the user access request. In this way, by segmented detection of the network performance status, compared with the network detection directly initiated from the user-side computing power gateway to a large number of computing power service instances, the number of network detection sessions of the first computing power gateway is reduced, the bandwidth occupied by the network detection messages in the computing power network domain is reduced, the network burden is reduced, and the efficiency and pertinence of the network performance status detection are improved.

[0050] Exemplarily, the respectively measuring of the first network performance status between the first computing power gateway and at least one second computing power gateway includes: sending at least one detection request message to each of the second computing power gateways; receiving a response packet corresponding to each of the detection request messages sent by each of the second computing power gateways; and determining the first network performance status based on the detection request message and the response packet corresponding to the detection request message.

[0051] Here, the detection request message can be any message sent by a suitable network detection tool, such as an Internet Control Message Protocol (ICMP) control message sent by a Packet Internet Groper (ping) detection tool, a message sent by a Network Statistics (Netstat) detection tool, etc.

[0052] It should be noted that every time the first computing power gateway sends a detection request message to the second computing power gateway, under normal circumstances, the second computing power gateway will return a corresponding response packet to the first computing power gateway. Here, the first network performance status table can be determined once based on one detection request message and the corresponding response packet, or the first network performance status table can be determined once based on multiple detection request messages and the corresponding response packets.

[0053] Exemplarily, the first computing power gateway acts as the initiator and the second computing power gateway acts as the receiver. The first computing power gateway sends a detection request message to the destination address of the second computing power gateway at a period of N, and receives a response packet from the second computing power gateway. The first computing power gateway sends M detection request messages to the destination address of the second computing power gateway. When the first computing power gateway receives the response packet corresponding to the last detection request message or when the timer for the response packet corresponding to the last detection request message times out, it calculates the network performance status between itself and the second computing power gateway.

[0054] In one example, all first network performance states may be stored in a local first preset table to obtain a first network performance state table for easy subsequent search.

[0055] In an embodiment of the present invention, a detection request message and a corresponding response packet are used to accurately determine any first network performance status between the first computing power gateway and the second computing power gateway, which is conducive to the subsequent supplementation of attribute information in the computing power routing.

[0056] Exemplarily, the comprehensive network performance status between the first computing power gateway and each of the computing power service instances is determined based on the first network performance status and the second network performance status, including: determining the target second computing power gateway corresponding to the first computing power gateway and the target computing power service instance corresponding to the target second computing power gateway; searching for the first path network performance status between the first computing power gateway and the target second computing power gateway from all the first network performance statuses; searching for the second path network performance status between the target second computing power gateway and the target computing power service instance from all the second network performance statuses; integrating the first path network performance status and the second path network performance status to obtain the comprehensive network performance status.

[0057] Here, the target second computing gateway and target computing service instance can be determined using the computing routing table. The computing routing table contains forwarding path information for all destination addresses of the computing service instance. The computing routing table records the next hop information as the IP address of the second computing gateway for the destination address of the computing service instance. The computing routing table is generated after the computing service instance is deployed. In other words, the computing routing table only records the path from the first computing gateway to the computing service instance and does not include the corresponding network performance status.

[0058] Specifically, in the computing power routing table, determine the IP address of the target second computing power gateway and the destination address of the target computing power service instance. Based on the two pieces of information, the IP address of the second computing power gateway and the destination address of the computing power service instance, find the network performance status from the first computing power gateway to the second computing power gateway in all first network performance states, and find the network performance status from the second computing power gateway to the destination address of the computing power service instance in all second network performance states. The two pieces of information are integrated and stored in the network performance status attribute of the corresponding computing power route in the computing power routing table.

[0059] Exemplarily, the method of integrating the first network performance status and the second network performance status to obtain a comprehensive network performance status may specifically include: if the network performance status is packet loss rate, then including the first packet loss rate and the second packet loss rate, and selecting the largest packet loss rate as the comprehensive packet loss rate; if the network performance status is average delay, then including the first average delay and the second average delay, and using the sum of the first average delay and the second average delay as the comprehensive average delay; if the network performance status is delay jitter, then including the first delay jitter and the second delay jitter, and using the sum of the first delay jitter and the second delay jitter as the comprehensive network performance status. If the network performance status includes three pieces of information: packet loss rate, average delay, and delay jitter, it is necessary to comprehensively calculate the three values ​​to obtain the comprehensive network performance status.

[0060] In an embodiment of the present invention, the segmented network status detection mechanism breaks down complex network detection tasks into multiple independent network layers or regional detection stages, and then completes data aggregation through protocol notifications. By integrating multiple segments of data, the end-to-end network status is obtained. Compared with network detection directly initiated from the user-side computing power gateway to a large number of computing power service instances, the number of network detection sessions of the user-side computing power gateway is reduced, the bandwidth occupied by network detection messages in the computing power network domain is reduced, the network burden is effectively reduced, and the pertinence and efficiency of network performance status detection are further improved. At the same time, the network performance status is mapped one-to-one with the path in the computing power routing table. The first computing power gateway can accurately perceive the network status of each specific computing power service instance, which enables the computing network scheduling to more precisely match user needs and computing power resources, thereby improving the service quality and user experience of the entire computing network system.

[0061] In one embodiment, the first network performance status may include: a packet loss rate and an average delay; determining the packet loss rate based on the detection request message and the response packet corresponding to the detection request message includes: determining a first number of the detection request messages and a second number of the response packets; determining the packet loss rate based on the first number and the second number;

[0062] Determining the average delay based on the detection request message and the response packet corresponding to the detection request message includes: determining the time difference between sending each detection request message and receiving the response packet corresponding to the detection request message; and determining the average delay based on each time difference.

[0063] It should be noted that the packet loss rate and average latency require multiple sets of detection request packets and response packets to determine. Here, detection request packets can be continuously sent within a preset time period to determine a first number of detection request packets and a second number of response packets; or a detection request packet can be sent at a preset period N, and M detection request packets can be continuously sent to determine the first number of detection request packets and the second number of response packets.

[0064] The packet loss rate may be a ratio of the second number to the first number, or a ratio of a weighted value of the second number to a weighted value of the first number.

[0065] Here, the time difference, also known as latency, is the difference between the time the first computing power gateway sends the detection request packet and the time it receives the corresponding response packet. If the first computing power gateway does not receive the corresponding response packet, the time difference is set to infinity.

[0066] After determining the time difference between each detection request message and the corresponding response packet, the average of multiple time differences is calculated to obtain the average delay.

[0067] In another embodiment, the first network performance state also includes: delay jitter; determining the delay jitter includes: determining the minimum time difference and the maximum time difference from each of the time differences respectively; determining the delay jitter based on the minimum time difference, the maximum time difference and the average delay.

[0068] It should be noted that delay jitter includes positive offset jitter and negative offset jitter. Positive offset jitter is the difference between the maximum delay and the average delay, or the difference between the weighted maximum delay and the weighted average delay. Negative offset jitter is the difference between the average delay and the minimum delay, or the difference between the weighted average delay and the weighted minimum delay.

[0069] Specifically, after determining the time difference between each detection request message and the corresponding response packet, the minimum delay and maximum delay are found, the positive offset jitter is calculated based on the maximum delay and average delay, and the negative offset jitter is calculated based on the minimum delay and average delay.

[0070] It should be noted that the network performance status may include one or more items of information including packet loss rate, average delay, and delay jitter. If the network performance status includes three items of information including packet loss rate, average delay, and delay jitter, it is necessary to calculate the packet loss rate, average delay, and delay jitter separately to obtain the network performance status.

[0071] In an embodiment of the present invention, by calculating multiple item information such as packet loss rate, average delay and delay jitter, the first network performance status is comprehensively determined based on the multiple item information, the accuracy of the routing quality evaluation between the first computing power gateway and the second computing power gateway is improved, and the user experience of the computing power service is improved.

[0072] Furthermore, the method further comprises:

[0073] Based on the first change value, the first network performance state is updated to obtain a third network performance state; the first change value is used to represent the change value between the first network performance state in the current time period and the first network performance state in the previous time period;

[0074] Receive a fourth network performance state; the fourth network performance state is obtained by the second computing power gateway updating the second network performance state based on the second change value; the second change value is used to represent the change value between the second network performance state in the current time period and the second network performance state in the previous time period;

[0075] Based on the third network performance status and the fourth network performance status, the comprehensive network performance status from the first computing power gateway to each of the computing power service instances is recalculated to obtain an updated comprehensive network performance status.

[0076] It should be noted that in order to ensure the real-time nature of the network performance status, when there is a significant change between the first network performance status and / or the second network performance status of two adjacent measurements, it is necessary to update the previous first network performance status and / or the second network performance status, and then update the comprehensive network performance status.

[0077] Exemplarily, when the first change value satisfies a first preset change condition, the first network performance state is updated, wherein the first preset change condition may include the current packet loss rate increasing or decreasing by a first preset value (such as 10%, 11%, etc.) compared with the previous packet loss rate, the current average delay increasing or decreasing by a second preset value (such as 10%, 11%, etc.) compared with the previous average delay, and the current jitter delay increasing or decreasing by a third preset value (such as 10%, 11%, etc.) compared with the previous jitter delay.

[0078] Here, the update method of the second network performance status may refer to the update method of the first network performance status.

[0079] It should be noted that the second computing power gateway sends the second network performance status change message to the first computing power gateway at most once per cycle. By setting an appropriate cycle length, both sending timeliness and saving network overhead can be taken into account.

[0080] Specifically, after obtaining the updated first network performance state (i.e., the third network performance state) and the updated second network performance state (i.e., the fourth network performance), the updated path can be determined according to the computing power routing table, and the updated comprehensive network performance state can be stored in the network performance state attribute of the corresponding path.

[0081] In one embodiment, the second computing power gateway may send the fourth network performance status to the first computing power gateway by configuring the network segment route to the computing power resource pool as a static route and configuring a network performance status carrying identifier. When the network performance status detected and generated in the second computing power gateway changes, the route configured with the network performance status carrying identifier is used to send a BGP update message, carrying the fourth network performance status, to implement network performance status information notification.

[0082] In an embodiment of the present invention, by updating the first network status and the second network status in real time, the blind spot of computing power routing to the network path performance status is eliminated, ensuring that computing power scheduling can make the optimal computing power routing based on the latest comprehensive network performance status.

[0083] In one example, the receiving of the second network performance status between each second computing power gateway and each computing power service instance sent by each second computing power gateway to obtain the second network performance status includes: establishing a neighbor relationship with each second computing power gateway respectively; using each neighbor relationship to receive the second network performance status message sent by the second computing power gateway corresponding to each neighbor relationship; parsing the second network performance status message to obtain the second network performance status.

[0084] Here, the neighbor relationship may be a Border Gateway Protocol (BGP) session established between the first computing power gateway and the second computing power gateway.

[0085] When the second computing power gateway measures the second network performance status, it communicates the second network performance status via a BGP message carrying the second network performance status through a designated route. If this is the first time the second computing power gateway has measured the second network performance status, it must communicate the second network performance status to the first computing power gateway. If this is not the first time the second network performance status has been measured, it must communicate the second network performance status to the first computing power gateway only if the change between the current second network performance status and the previous network performance status meets a preset change condition.

[0086] Here, the first computing power gateway can store the received second network performance status in a second preset table to obtain a second network performance status table for subsequent convenient search.

[0087] In the embodiment of the present application, the second network performance state is notified to the first computing power gateway in the form of a message through the second computing power gateway, which greatly reduces the overall number of detection messages in the computing power network, effectively reduces the occupation of network state detection on service bandwidth, and reduces the network burden.

[0088] Figure 2 is one of the structural schematic diagrams of the computing power network system provided by the present application, as shown in Figure 2 The user 201 in the computing power network system 200 sends a user access request to the first computing power gateway 202, the first computing power gateway 202 and the second computing power gateway 203 access the computing power network 204, each second computing power gateway 203 connects one or more computing power resource pools 205, and a plurality of computing power service instances are deployed in the computing power resource pool.

[0089] Figure 3 is the second flowchart of the network performance state determination method provided by the present application, as shown in Figure 3 , comprising:

[0090] Step 301, the first computing power gateway performs ping detection on the second computing power gateway to obtain the first network performance state, and the second computing power gateway performs ping detection on the computing power service instances in the computing power resource pool to obtain the second network performance state;

[0091] Step 302, the second computing power gateway performs BGP notification to the first computing power gateway, and the network performance state information is extended and carried;

[0092] Here, the network performance state information can be the second network performance state.

[0093] Step 303, the first computing power gateway integrates the network performance state information.

[0094] The following is an application scenario of the network performance state determination method provided by the embodiment of the present application.

[0095] Figure 4 is the second structural schematic diagram of the computing power network system provided by the present application, as shown in Figure 4 User A and user B access the computing power network through the first computing power gateway 1 and the second computing power gateway 2, two computing power resource pools access the computing power network through the second computing power gateway 1 and the second computing power gateway 2, the three-layer network communication of the two first computing power gateways and the two second computing power gateways is connected, and the intermediate layer network can be a normal IP network or a special network connection. Among them, two computing power service instances are deployed in the computing power resource pool 1: A type computing power service instance a-1; B type computing power service instance b-1; one computing power service instance is deployed in the computing power resource pool 2: A type computing power service instance a-2.

[0096] In this application scenario, the first computing power gateway 1 and the second computing power gateway 2 are connected through the three-layer network communication, and the network performance state information of the two computing power resource pools is carried in the BGP message. Figure 4 The method for determining the network performance status provided by the present invention is illustrated by taking the computing power network system in the example.

[0097] First, measuring a first network performance state and a second network performance state;

[0098] 1. The first network status measurement method includes: a) First computing power gateway 1 initiates a ping network test to second computing power gateway 1 and second computing power gateway 2. It sends a ping request every 3 seconds. After sending 10 ping request messages, upon receiving the ping relay response packet corresponding to the last ping request or upon the timer waiting for the ping relay response packet to the last ping request expires, it calculates the network performance status of itself and the second computing power gateway 1 and second computing power gateway 2, and updates the calculated results to the local user-side network performance status table. This detection process is repeated. At a certain moment, the network performance status table of first computing power gateway 1 records the data shown in Table 1 below. b) First computing power gateway 2 initiates a ping network test to second computing power gateway 1 and second computing power gateway 2. It sends a ping request every 3 seconds. After sending 10 ping request messages, upon receiving the ping relay response packet corresponding to the last ping request or upon the timer waiting for the ping relay response packet to the last ping request expires, it calculates the network performance status of itself and the destination end, and updates the calculated results to the local user-side network performance status table. This detection process is repeated. The data recorded in the user-side network performance status table at a certain moment is as follows:

[0099] Table 1

[0100]

[0101] Table 2

[0102]

[0103] 2. Each second computing gateway measures the network performance status of all computing service instances in all connected computing resource pools: a) Second computing gateway 1 initiates a ping network test to two computing service instances a-1 and b-1 in computing resource pool 1. It sends a ping request every 3 seconds. After sending 10 ping requests, upon receiving the ping relay response packet corresponding to the last ping request or upon the expiration of the ping relay response timer, it calculates the network performance status of itself and the destination, and updates the calculation result to the local computing network performance status table. This test process continues. The data recorded in the computing network performance status table at a certain moment is shown in Table 3 below. b) Second computing gateway 2 initiates a ping network test to a computing service instance a-2 in computing resource pool 2. A ping request is sent every 3 seconds. After sending 10 ping request messages, when the ping relay response packet corresponding to the last ping request is received or the timer for the ping relay response packet corresponding to the last ping request expires, the network performance status of the network itself and the destination is calculated and the calculation results are updated to the local computing power side network performance status table. The above detection process is continued. The data recorded in the computing power side network performance status table at a certain moment is as follows Table 4:

[0104] Table 3

[0105]

[0106] Table 4

[0107]

[0108] Next, the second computing power gateway notifies the first computing power gateway of the second network performance status:

[0109] 1. BGP Neighborhood Establishment: First Computing Gateway 1 establishes neighbor relationships with Second Computing Gateways 1 and 2. First Computing Gateway 2 establishes neighbor relationships with Second Computing Gateways 1 and 2. That is, each First Computing Gateway must establish neighbor relationships with all Second Computing Gateways to obtain network performance status measurements from all Second Computing Gateways to all Computing Service Instances in the Computing Resource Pool connected to it.

[0110] 2. Second Computing Power Gateway Network Performance Status Notification: When the entry information in the computing power-side network performance status table stored on the second computing power gateway changes by more than a certain indicator (such as a 10% increase or decrease in network packet loss rate, a 10% increase or decrease in average latency, or a 10% increase or decrease in latency jitter), the second computing power gateway notifies the local computing power-side network performance status table of the information change via a BGP update message, using a designated route carrying the network status attribute. The second computing power gateway configures the network segment route to the computing power resource pool as a static route and configures it with a network performance status indicator. When the network performance status detected and generated by the second computing power gateway changes, it sends a BGP update message carrying the network performance status via this route configured with the network performance status indicator, thereby implementing network performance status information notification. (This description is an example method; other notification methods can be customized.)

[0111] 3. BGP update message extension: Figure 5 This is a schematic diagram of the BGP / BGP4+ update message format provided by the present invention, such as Figure 5 As shown in the figure, the BGP / BGP4+ update message format includes Withdrawn Routes Length, Withdrawn Routes variable, Total Path Attribute Length, Path Attribute variable, and Network Layer Reachability Information variable. Path attributes in BGP / BGP4+ update messages include a newly defined computing network performance attribute, which supports the notification of network performance status information of reaching a specific computing service instance measured by the second computing gateway in the computing network.

[0112] The Path Attribute contains a list of route attributes to be updated, sorted in ascending order by type number. All attributes of the updated route are populated. A specific type number is defined to store the computing power network performance attribute. Each attribute unit consists of three parts: attribute type, attribute length, and attribute value. It is encoded in TLV format. The type field occupies 2 bytes and is divided into two fields: the attribute tag (Attr.Flags) and the attribute type value (Attr.Type Code). The BGP Path Attribute extension supports the computing power network performance attribute type definition as shown in Table 5. In Table 5, bits O, T, and E of Attr.Flags in the computing power network performance attribute should be set to 1. O=1 and T=1 indicate that the attribute is optional and must be transmitted. This means that devices that do not recognize the attribute will still receive it and forward it to other BGP peers. E=1 indicates that the attribute length is extended to 2 bytes. The type value of the computing power network performance attribute is defined as 65, that is, Attr.Type Code = 65.

[0113] Table 5

[0114]

[0115] The BGP Path Attribute extension supports computing network performance attributes, as defined in Table 6. In Table 6, the Router ID of the second computing gateway itself refers to the BGP Router ID configured on the second computing gateway, indicating the source of the computing routing attributes. The computing service instance IP address indicates that the network performance status measurement results from the second computing gateway to the specific computing service instance IP address are being advertised. The total length of the network performance attribute information is the total number of bytes in the network performance attribute information field carried in the message. The network performance attribute information consists of a set of TLVs that carry the network performance attribute information to be advertised. The specific format definition is: type occupies 1 byte, length occupies 1 byte, and value length depends on length. Type definitions include: 0x01: Packet loss rate; 0x02: Average delay; 0x03: Positive offset jitter; 0x04: Negative offset jitter.

[0116] Table 6

[0117]

[0118] 4、Based on the above message extension format, each second computing power gateway encapsulates the network performance state measurement information of all computing power service instances in all computing power resource pools connected to itself into a message and announces it to the first computing power gateway neighbors. That is, the second computing power gateway 1 announces the network performance state information of the two computing power service instances a-1 and b-1 in the computing power resource pool 1 to the first computing power gateways 1 and 2 respectively, and table 3 corresponds to the network performance state complete announcement transmission; the second computing power gateway 2 announces the network performance state information of the computing power service instance a-2 in the computing power resource pool 2 to the first computing power gateways 1 and 2 respectively, and table 4 corresponds to the network performance state complete announcement transmission.

[0119] Next, the first computing power gateway integrates the network performance state:

[0120] 1、The first computing power gateway receives the BGP update message announced by the second computing power gateway, obtains the contents of table 3 and table 4, and records the generated computing power side network performance state table locally. And combined with the local user side network performance state table information, the end-to-end network state information is obtained by integration, and the computing power routing information is updated.

[0121] 2、The first computing power gateway maintains a computing power routing table, and the computing power routing table generates forwarding path information to all computing power service instance destination addresses. The next hop information in the computing power routing table is the second computing power gateway IP address when the destination address is the computing power service instance destination address. According to the second computing power gateway IP address and the computing power service instance destination address, the network performance state information from the first computing power gateway to the second computing power gateway is found in the user side network performance state table, and the network performance state information from the second computing power gateway and the computing power service instance destination address is found in the computing power side network performance state table. Process two pieces of information and store them in the corresponding computing power routing network performance state information attribute to obtain the network performance state in the computing power routing. The network performance state in the first computing power gateway 1 computing power routing in table 7 and the network performance state in the first computing power gateway 2 computing power routing in table 8:

[0122] Table 7

[0123]

[0124] 3. The calculated packet loss rate, average latency, and latency jitter from the first computing gateway to the destination address of a computing service instance are updated to the corresponding computing route as network performance attributes of the computing route. When a user issues a service request, the first computing gateway compares the user's performance requirements (such as low latency and low packet loss rate) with the current network performance of each computing service instance in the computing route. It then makes a decision and selects the most appropriate computing service instance to respond to the user's service request. For example, if user A connects to the first computing gateway 1 and requests access to computing service instance A via the shortest latency path, the first computing gateway 1 will prioritize dispatching the traffic to computing service instance a-2, forwarding it through the computing route with the next hop being the second computing gateway 2.

[0125] 4. When the user-side network performance status table and the computing power-side network performance status table maintained by the first computing power gateway are updated, the relevant computing power route is retrieved, and the packet loss rate, average delay, and delay jitter information from the first computing power gateway to the destination address of a computing power service instance are recalculated and updated to the corresponding computing power route.

[0126] Table 8

[0127]

[0128] The method for determining the network performance status provided by the present invention ensures that the computing network can make the best decision based on the latest network status information by dynamically sensing the network status of the computing service instance, including key indicators such as latency, latency jitter, packet loss rate and service reachability. The first computing power gateway can accurately sense the network status of each specific computing power service instance, which enables the computing network scheduling to more finely match user needs with computing power resources, thereby improving the service quality and user experience of the entire computing network system. The innovative segmented network status detection mechanism breaks down complex network detection tasks into multiple independent network layers or regional detection stages, and then completes data aggregation through protocol notifications. By integrating multiple segments of data, the end-to-end network status is obtained. Compared with network detection directly initiated from the first computing power gateway to a large number of computing power service instances, the number of network detection sessions of the first computing power gateway is reduced, and the bandwidth occupied by network detection messages within the computing power network domain is reduced. This method not only improves the pertinence and efficiency of detection, but also significantly reduces the overall number of detection messages in the network, effectively reducing the network burden.

[0129] The following describes the device for determining the network performance status provided by the present invention. The device for determining the network performance status described below and the method for determining the network performance status described above can refer to each other.

[0130] Figure 6 FIG. 1 is a schematic diagram of the structure of the device for determining the network performance status provided by the present invention, such as Figure 6 As shown, the network performance status determination device 600 includes:

[0131] A measurement module 601 is configured to measure a first network performance status between a first computing power gateway and at least one second computing power gateway, respectively; the first computing power gateway is a computing power gateway connected to a user terminal, and the second computing power gateway is a computing power gateway connected to a computing power resource pool that provides computing power resources;

[0132] Receiving module 602, configured to receive a second network performance status between each second computing power gateway and each computing power service instance sent by each second computing power gateway; the computing power service instance is deployed in each computing power resource in the computing power resource pool connected to the second computing power gateway;

[0133] Determination module 603 is used to determine the comprehensive network performance status between the first computing power gateway and each of the computing power service instances based on the first network performance status and the second network performance status; wherein the comprehensive network performance status is used to determine the optimal computing power route for user access requests.

[0134] In some embodiments, the measurement module 601 is specifically used to: send at least one detection request message to each of the second computing power gateways; receive a response packet corresponding to each of the detection request messages sent by each of the second computing power gateways; and determine the first network performance status based on the detection request message and the response packet corresponding to the detection request message.

[0135] In some embodiments, the determination module 603 is specifically used to: determine the target second computing power gateway corresponding to the first computing power gateway and the target computing power service instance corresponding to the target second computing power gateway; search for the first path network performance status between the first computing power gateway and the target second computing power gateway from all the first network performance states; search for the second path network performance status between the target second computing power gateway and the target computing power service instance from all the second network performance states; integrate the first path network performance status and the second path network performance status to obtain a comprehensive network performance status.

[0136] In some embodiments, the first network performance status includes: packet loss rate and average delay; the measurement module 601 is further specifically used to: determine a first number of the detection request messages and a second number of the response packets; determine the packet loss rate based on the first number and the second number; determine the average delay based on the detection request message and the response packet corresponding to the detection request message, including: determining the time difference between sending each of the detection request messages and receiving the response packet corresponding to the detection request message; and determine the average delay based on each of the time differences.

[0137] In some embodiments, the measurement module 601 is further specifically used to: the first network performance state also includes: delay jitter; determining the delay jitter includes: determining the minimum time difference and the maximum time difference from each of the time differences respectively; determining the delay jitter based on the minimum time difference, the maximum time difference and the average delay.

[0138] In some embodiments, the network performance status determination device 600 also includes an update module, which is specifically used to: update the first network performance status based on a first change value to obtain a third network performance status; the first change value is used to represent the change value between the first network performance status in the current time period and the first network performance status in the previous time period; receive a fourth network performance status; the fourth network performance status is obtained by the second computing power gateway updating the second network performance status based on the second change value; the second change value is used to represent the change value between the second network performance status in the current time period and the second network performance status in the previous time period; based on the third network performance status and the fourth network performance status, recalculate the comprehensive network performance status from the first computing power gateway to each of the computing power service instances to obtain an updated comprehensive network performance status.

[0139] In some embodiments, the receiving module 602 is specifically used to: establish a neighbor relationship with each of the second computing power gateways respectively; use each of the neighbor relationships to receive the second network performance status message sent by the second computing power gateway corresponding to each of the neighbor relationships; parse the second network performance status message to obtain the second network performance status.

[0140] Figure 7 This is a schematic diagram of the physical structure of the electronic device provided by the present invention, such as Figure 7As shown, the electronic device may include: a processor 710, a communication interface 720, a memory 730, and a communication bus 740, wherein the processor 710, the communication interface 720, and the memory 730 communicate with each other via the communication bus 740. The processor 710 may invoke logic instructions in the memory 730 to execute a method for determining a network performance status, the method comprising: measuring a first network performance status between a first computing power gateway and at least one second computing power gateway; the first computing power gateway is a computing power gateway connected to a user terminal, and the second computing power gateway is a computing power gateway connected to a computing power resource pool that provides computing power resources; receiving a second network performance status between each second computing power gateway and each computing power service instance, sent by each second computing power gateway; the computing power service instance is deployed in each computing power resource in the computing power resource pool connected to the second computing power gateway; and determining a comprehensive network performance status between the first computing power gateway and each computing power service instance based on the first network performance status and the second network performance status; wherein the comprehensive network performance status is used to determine an optimal computing power route for a user access request.

[0141] Furthermore, the logic instructions in the aforementioned memory 730 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product, stored in a storage medium, includes instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0142] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the network performance status determination method provided by the above methods, the method including: measuring the first network performance status between the first computing power gateway and at least one second computing power gateway respectively; the first computing power gateway is a computing power gateway connected to the user terminal, and the second computing power gateway is a computing power gateway connected to the computing power resource pool that provides computing power resources; receiving the second network performance status between each second computing power gateway and each computing power service instance sent by each second computing power gateway; the computing power service instance is deployed in each computing power resource in the computing power resource pool connected to the second computing power gateway; based on the first network performance status and the second network performance status, determining the comprehensive network performance status between the first computing power gateway and each computing power service instance; wherein the comprehensive network performance status is used to determine the optimal computing power route for the user access request.

[0143] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a method for determining the network performance status provided by the above-mentioned methods, the method comprising: measuring the first network performance status between a first computing power gateway and at least one second computing power gateway respectively; the first computing power gateway is a computing power gateway connected to a user terminal, and the second computing power gateway is a computing power gateway connected to a computing power resource pool that provides computing power resources; receiving the second network performance status between each second computing power gateway and each computing power service instance sent by each second computing power gateway; the computing power service instance is deployed in each computing power resource in the computing power resource pool connected to the second computing power gateway; based on the first network performance status and the second network performance status, determining the comprehensive network performance status between the first computing power gateway and each computing power service instance; wherein the comprehensive network performance status is used to determine the optimal computing power route for the user access request.

[0144] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0145] Those skilled in the art can clearly understand the technical solutions of the various embodiments from the above description of the embodiments, and the various embodiments can be implemented by means of software with the necessary general hardware platforms, and of course, can also be implemented by hardware. Based on such understanding, the above technical solutions, essentially or in other words, the part of the prior art that makes a contribution, can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, and the like, and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0146] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, rather than limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for some technical features therein; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for determining network performance status, applied to a first computing power gateway, characterized in that: include: respectively measuring a first network performance status between the first computing power gateway and at least one second computing power gateway; The first computing power gateway is a computing power gateway connected to the user terminal, and the second computing power gateway is a computing power gateway connected to the computing power resource pool that provides computing power resources; Receiving a second network performance status between each second computing power gateway and each computing power service instance sent by each second computing power gateway; the computing power service instance is deployed in each computing power resource in the computing power resource pool connected to the second computing power gateway; Based on the first network performance status and the second network performance status, determine the comprehensive network performance status between the first computing power gateway and each of the computing power service instances; wherein the comprehensive network performance status is used to determine the optimal computing power routing for user access requests.

2. The method for determining network performance status according to claim 1, wherein: The respectively measuring the first network performance status between the first computing power gateway and at least one second computing power gateway includes: Send at least one detection request message to each of the second computing power gateways; Receive a response packet corresponding to each of the detection request messages sent by each of the second computing power gateways; The first network performance status is determined based on the detection request message and a response packet corresponding to the detection request message.

3. The method for determining network performance status according to claim 1, wherein: The determining, based on the first network performance status and the second network performance status, of a comprehensive network performance status between the first computing power gateway and each of the computing power service instances includes: Determine a target second computing power gateway corresponding to the first computing power gateway and a target computing power service instance corresponding to the target second computing power gateway; Searching for a first path network performance status between the first computing power gateway and the target second computing power gateway from all the first network performance statuses; Searching for a second path network performance status between the target second computing power gateway and the target computing power service instance from all the second network performance statuses; The first path network performance status and the second path network performance status are integrated to obtain the comprehensive network performance status.

4. The method for determining network performance status according to claim 2, wherein: The first network performance status includes: packet loss rate and average delay; Determining the packet loss rate based on the detection request message and a response packet corresponding to the detection request message includes: Determining a first number of the detection request packets and a second number of the response packets; determining the packet loss rate based on the first number and the second number; Determining the average delay based on the detection request message and a response packet corresponding to the detection request message includes: Determining a time difference between sending each of the detection request messages and receiving a response packet corresponding to the detection request message; Based on each of the time differences, the average delay is determined.

5. The method for determining network performance status according to claim 4, wherein: The first network performance status further includes delay jitter; and determining the delay jitter includes: determining a minimum time difference and a maximum time difference from each of the time differences; The delay jitter is determined based on the minimum time difference, the maximum time difference, and the average delay.

6. The method for determining network performance status according to claim 1, wherein: The method further comprises: Based on the first change value, the first network performance state is updated to obtain a third network performance state; the first change value is used to represent the change value between the first network performance state in the current time period and the first network performance state in the previous time period; Receive a fourth network performance state; the fourth network performance state is obtained by the second computing power gateway updating the second network performance state based on the second change value; the second change value is used to represent the change value between the second network performance state in the current time period and the second network performance state in the previous time period; Based on the third network performance status and the fourth network performance status, the comprehensive network performance status from the first computing power gateway to each of the computing power service instances is recalculated to obtain an updated comprehensive network performance status.

7. The method for determining network performance status according to claim 1, wherein: The receiving the second network performance status between each second computing power gateway and each computing power service instance sent by each second computing power gateway to obtain the second network performance status includes: Establishing a neighbor relationship with each of the second computing power gateways respectively; Using each of the neighbor relationships, receiving a second network performance status message sent by the second computing power gateway corresponding to each of the neighbor relationships; Parse the second network performance status message to obtain the second network performance status.

8. A device for determining network performance status, characterized in that: include: A measurement module, configured to respectively measure a first network performance status between the first computing power gateway and at least one second computing power gateway; The first computing power gateway is a computing power gateway connected to the user terminal, and the second computing power gateway is a computing power gateway connected to the computing power resource pool that provides computing power resources; A receiving module, configured to receive a second network performance status between each second computing power gateway and each computing power service instance sent by each second computing power gateway; the computing power service instance is deployed in each computing power resource in the computing power resource pool connected to the second computing power gateway; A determination module is used to determine the comprehensive network performance status between the first computing power gateway and each of the computing power service instances based on the first network performance status and the second network performance status; wherein the comprehensive network performance status is used to determine the optimal computing power route for user access requests.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method for determining the network performance status according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for determining the network performance status according to any one of claims 1 to 7 is implemented.

Images