Load balancing method, device and equipment of supernode system southbound network and medium

Abstract

Embodiments of the present application disclose a load balancing method, device and equipment for a southbound network of a supernode system and a medium, comprising: in the southbound network of the supernode system, obtaining a symmetric network segment corresponding to each networking type; wherein each symmetric network segment contains at least one group of source leaf switches and destination leaf switches, and each group of source leaf switches and destination leaf switches is configured with multiple links; using a static routing scheme or a policy routing scheme corresponding to each symmetric network segment, the traffic in each symmetric network segment is evenly distributed to the corresponding links; wherein the static routing scheme is used to route the traffic to different southbound network out-interface groups on the destination leaf switch to different links, and the policy routing scheme is used to route the traffic from different southbound network in-interface groups on the source leaf switch to different links, avoiding data transmission disorder and ensuring the load balancing effect of the southbound network of the supernode system.

Description

Technical Field

[0001] This invention relates to the field of southbound network control, and more particularly to a load balancing method, apparatus, device, and medium for a southbound network of a supernode system. Background Technology

[0002] When the southbound network of a supernode system exceeds one layer, it involves interconnecting multiple links between two or more switching nodes. In this case, load balancing scheduling of these links is necessary. Uneven load distribution among the links can lead to network congestion and low utilization of some link resources, ultimately resulting in a decrease in the overall network transmission efficiency.

[0003] In existing technologies, packet spraying or flow-by-flow hashing are often used to achieve load balancing in the southbound network of a supernode system. Packet spraying distributes individual data packets from the same data stream across multiple equivalent paths for transmission, achieving packet-level load balancing. Flow-by-flow hashing performs hash calculations on the data stream, mapping all data packets from the same stream to the same link for transmission.

[0004] However, in packet spraying, data packets of the same data stream need to be transmitted in parallel through multiple paths, and are prone to out-of-order delivery due to differences in path latency and congestion. Therefore, if the receiving end does not have the ability to reassemble out-of-order data, it will cause data corruption and service anomalies. Secondly, in stream-by-stream hashing, the same data stream always travels through the same link, which can easily lead to uneven load when the number of streams is small. Summary of the Invention

[0005] This invention provides a load balancing method, apparatus, device, and medium for the southbound network of a supernode system. It solves the problem that existing technologies that use packet spraying or flow-by-flow hashing to achieve load balancing in the southbound network of a supernode system are prone to data corruption or uneven link load. It fully utilizes the symmetry of the southbound network of the supernode system, avoids data transmission corruption, and ensures the load balancing effect of the southbound network of the supernode system.

[0006] In a first aspect, embodiments of the present invention provide a load balancing method for a southbound network of a supernode system, comprising: acquiring the southbound network in the supernode system and the network topology adopted by the southbound network; acquiring symmetric network segments corresponding to each network topology in the southbound network; wherein each symmetric network segment contains at least one set of source leaf switches and destination leaf switches, and multiple links are configured between each set of source leaf switches and destination leaf switches; and using a static routing scheme or a policy routing scheme corresponding to each symmetric network segment to evenly distribute traffic within each symmetric network segment to the corresponding links; wherein the static routing scheme is used to route traffic to different southbound network outgoing interface groups on the destination leaf switch to different links, and the policy routing scheme is used to route traffic from different southbound network incoming interface groups on the source leaf switch to different links.

[0007] Secondly, embodiments of the present invention also provide a load balancing device for a southbound network of a supernode system, comprising: a southbound network acquisition module, used to acquire the southbound network in the supernode system and the network topology adopted by the southbound network; a network segment acquisition module, used to acquire symmetrical network segments corresponding to each network topology in the southbound network; wherein each symmetrical network segment includes at least one set of source leaf switches and destination leaf switches, and multiple links are configured between each set of source leaf switches and destination leaf switches; and a traffic balancing allocation module, used to evenly allocate traffic within each symmetrical network segment to the corresponding links using a static routing scheme or a policy routing scheme corresponding to each symmetrical network segment; wherein the static routing scheme is used to route traffic to different southbound network outgoing interface groups on the destination leaf switch to different links, and the policy routing scheme is used to route traffic from different southbound network incoming interface groups on the source leaf switch to different links.

[0008] Thirdly, embodiments of the present invention also provide an electronic device, the electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to execute the load balancing method for the southbound network of the supernode system provided in any embodiment of the present invention.

[0009] Fourthly, embodiments of the present invention also provide a computer-readable storage medium storing computer instructions, which are used to cause a processor to execute and implement the load balancing method of the southbound network of the supernode system of any embodiment of the present invention.

[0010] The technical solution of this invention obtains symmetrical network segments corresponding to each network type in the southbound network, and uses a static routing scheme or a policy routing scheme corresponding to each symmetrical network segment to evenly distribute traffic within each symmetrical network segment to the corresponding links. The static routing scheme routes traffic destined for different southbound network outgoing interface groups on the destination leaf switch to different links, while the policy routing scheme routes traffic from different southbound network incoming interface groups on the source leaf switch to different links. This solves the problem that existing technologies using packet spraying or flow-by-flow hashing to achieve load balancing in the southbound network of a supernode system are prone to data corruption or uneven link load. It fully utilizes the symmetry of the southbound network in the supernode system, avoids data transmission corruption, and ensures the load balancing effect of the southbound network.

[0011] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a flowchart of a load balancing method for a southbound network of a supernode system according to Embodiment 1 of the present invention.

[0014] Figure 2 This is a flowchart of another load balancing method for a southbound network of a supernode system provided in Embodiment 2 of the present invention.

[0015] Figure 3 This is a schematic diagram of a static routing scheme for leaf-to-leaf networking provided in Embodiment 2 of the present invention.

[0016] Figure 4 This is a schematic diagram of a policy routing scheme for leaf-to-leaf networking provided according to Embodiment 2 of the present invention.

[0017] Figure 5 This is a flowchart of another load balancing method for a southbound network of a supernode system provided in Embodiment 3 of the present invention.

[0018] Figure 6 This is a schematic diagram of a static routing scheme for leaf-ridge networking provided in Embodiment 3 of the present invention.

[0019] Figure 7 This is a schematic diagram of a policy routing scheme for leaf-spine networking provided in Embodiment 3 of the present invention.

[0020] Figure 8 This is a schematic diagram of the structure of a load balancing device for a southbound network of a supernode system according to Embodiment 4 of the present invention.

[0021] Figure 9 This is a schematic diagram of the structure of an electronic device provided in Embodiment 5 of the present invention. Detailed Implementation

[0022] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0023] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0024] Example 1 Figure 1 This is a flowchart of a load balancing method for a southbound network of a supernode system according to Embodiment 1 of the present invention. This embodiment is applicable to the case of load balancing for a southbound network of a supernode system. The method can be executed by a load balancing device for the southbound network of the supernode system. The load balancing device for the southbound network of the supernode system can be implemented in hardware and / or software. The load balancing device for the southbound network of the supernode system can be configured in an electronic device such as the management server of the supernode system.

[0025] like Figure 1 As shown in the figure, this embodiment discloses a load balancing method for a southbound network of a supernode system, including steps S110-S130.

[0026] S110. Obtain the southbound network in the supernode system, and the network topology used by the southbound network.

[0027] In this embodiment, the supernode system can be understood as a rack-level computing infrastructure that deeply couples a large number of Artificial Intelligence (AI) accelerator cards into a logically single system, with high-speed interconnection and unified memory at its core. An AI accelerator card can be understood as a hardware acceleration module that integrates an AI chip and corresponding peripheral circuits, specifically designed to perform AI-related computational tasks.

[0028] The southbound network can be understood as a high-speed interconnection network within the supernode system, used to enable data transmission between AI accelerator cards and between AI accelerator cards and switching nodes. Switching nodes receive traffic from upstream AI accelerator cards and forward it to the appropriate downstream AI accelerator cards according to routing policies. There are various types of switching nodes, such as leaf switches and spine switches.

[0029] Network topology can be understood as the interconnection topology between nodes in a southbound network, which can include leaf-leaf networks and leaf-spine networks. A leaf-leaf network can be understood as a network type consisting of multiple leaf switches directly interconnected, without a spine switch. A leaf-spine network can be understood as a two-layer network type including both leaf and spine switches, with all leaf and spine switches fully interconnected.

[0030] Specifically, this step involves obtaining southbound network information from the supernode system, including topology and network type. Then, the network topology used by the southbound network can be determined based on this information.

[0031] S120. In the southbound network, obtain the symmetrical network segments corresponding to each network type; wherein each symmetrical network segment contains at least one set of source leaf switches and destination leaf switches, and multiple links are configured between each set of source leaf switches and destination leaf switches.

[0032] In this embodiment, a symmetrical network segment can be understood as a network segment where the number of southbound network interfaces connected to the source leaf switch and the destination leaf switch are equal, and the number of links corresponding to switches of the same type is also equal. A source leaf switch can be understood as a leaf switch directly connected to the AI ​​accelerator card that sends traffic. A destination leaf switch can be understood as a leaf switch directly connected to the AI ​​accelerator card that receives traffic.

[0033] Specifically, in this step, when the southbound network uses only leaf-leaf networking, a symmetrical network segment of leaf-leaf networking type is obtained. When the southbound network uses only leaf-ridge networking type, a symmetrical network segment of leaf-ridge networking type is obtained. When the southbound network uses both leaf-leaf and leaf-ridge networking types, both symmetrical network segments of leaf-leaf networking type and symmetrical network segments of leaf-ridge networking type are obtained.

[0034] It is worth noting that a symmetric network segment of the leaf-spine networking type contains at least two leaf switches and one spine switch, and each leaf switch is configured with multiple links to each spine switch.

[0035] S130. Using a static routing scheme or a policy routing scheme corresponding to each symmetrical network segment, the traffic within each symmetrical network segment is evenly distributed to the corresponding links. The static routing scheme is used to route traffic to different southbound network outgoing interface groups on the destination leaf switch to different links, and the policy routing scheme is used to route traffic from different southbound network incoming interface groups on the source leaf switch to different links.

[0036] In this embodiment, the southbound network outgoing interface group can be understood as a set of interfaces used to forward traffic out of the southbound network; it is the set of exit points for traffic leaving the destination leaf switch. Each southbound network outgoing interface group includes at least two southbound network outgoing interfaces.

[0037] A southbound network ingress interface group can be understood as a set of interfaces used to receive traffic flowing into the southbound network; it is the set of entry points for traffic into the source leaf switch. Each southbound network ingress interface group includes at least two southbound network ingress interfaces.

[0038] In this step, specifically, a static routing scheme corresponding to each symmetrical network segment can be used to evenly distribute traffic within each symmetrical network segment to the corresponding links based on the southbound network outgoing interface groups corresponding to different traffic types. Alternatively, a policy-based routing scheme corresponding to each symmetrical network segment can be used to evenly distribute traffic within each symmetrical network segment to the corresponding links based on the southbound network incoming interface groups corresponding to different traffic types.

[0039] The advantage of this setup is that, since static routing and policy routing schemes only rely on the basic features commonly supported by each Layer 3 Ethernet switch, there is no need for the switch to have advanced functions such as dynamic load balancing and global load balancing, nor is there a need to introduce an additional complex operation and maintenance platform to achieve good load balancing results. This improves the versatility of the technical solution in this embodiment and reduces the application cost of the technical solution in this embodiment.

[0040] Optionally, after using a static routing scheme or policy-based routing scheme corresponding to each symmetric network segment to evenly distribute traffic within each symmetric network segment to the corresponding links, the method further includes: detecting whether there are faulty links in each symmetric network segment; identifying symmetric network segments with faulty links as faulty network segments, and using a dynamic routing scheme to re-determine the links corresponding to traffic within the faulty network segments; wherein the dynamic routing scheme is a scheme determined based on a dynamic routing protocol. Various dynamic routing protocols can be used, such as the Border Gateway Protocol (BGP) and the Open Shortest Path First (OSPF) protocol.

[0041] Specifically, the connectivity status of each link in each symmetric network segment can be monitored, and links with a disconnected connectivity status can be identified as faulty links. Then, since faulty network segments containing faulty links cannot achieve link load balancing through static or policy-based routing schemes, a dynamic routing scheme can be used to redetermine the links corresponding to traffic within the faulty network segment.

[0042] The advantage of this setup is that when a symmetric network segment fails, a backup routing scheme using dynamic routing can be used to re-determine the links corresponding to traffic within the failed network segment. By leveraging the equal-cost multipath capability of the dynamic routing protocol itself, the system reliability and load balancing effect in failure scenarios can be maximized.

[0043] The technical solution of this embodiment obtains the southbound network in the supernode system and the network topology adopted by the southbound network; in the southbound network, it obtains symmetrical network segments corresponding to each network topology; wherein each symmetrical network segment contains at least one set of source leaf switches and destination leaf switches, and multiple links are configured between each set of source leaf switches and destination leaf switches; and it uses a static routing scheme or a policy routing scheme corresponding to each symmetrical network segment to evenly distribute the traffic in each symmetrical network segment to the corresponding links. The static routing scheme is used to route traffic to different southbound network outgoing interface groups on the destination leaf switch to different links, and the policy routing scheme is used to route traffic from different southbound network incoming interface groups on the source leaf switch to different links. This solves the problem that existing technologies that use packet spraying or flow-by-flow hashing to achieve load balancing of the southbound network of the supernode system are prone to data corruption or uneven link load. It makes full use of the symmetry of the southbound network topology of the supernode system, avoids data transmission corruption, and ensures the load balancing effect of the southbound network of the supernode system.

[0044] Example 2 Figure 2This is a flowchart of another load balancing method for a southbound network of a supernode system according to Embodiment 2 of the present invention. This embodiment is a refinement based on the above embodiments, and specifically describes the case where the southbound network only supports leaf-leaf networking.

[0045] like Figure 2 As shown in the figure, this embodiment discloses a load balancing method for a southbound network of a supernode system, including steps S210-S260.

[0046] S210. Obtain the southbound network in the supernode system, and the network topology used by the southbound network.

[0047] S220. When the southbound network adopts only a leaf-leaf network, obtain a symmetrical network segment with a leaf-leaf network type in the southbound network; wherein each symmetrical network segment contains at least one set of source leaf switches and destination leaf switches, and multiple links are configured between each set of source leaf switches and destination leaf switches.

[0048] In this embodiment, the number of southbound network interfaces on each switch in the symmetrical network segment is a power of 2, and the number of links between each interconnecting switch is also a power of 2. For example, a network segment containing a set of source leaf switches and destination leaf switches, with 64 southbound network interfaces on both the source and destination leaf switches, and 32 links between the source and destination leaf switches, can be considered a symmetrical network segment with a leaf-leaf network type. Alternatively, a network segment containing a set of source and destination leaf switches, with 32 southbound network interfaces on both the source and destination leaf switches, and 16 links between the source and destination leaf switches, can also be considered a symmetrical network segment with a leaf-leaf network type.

[0049] S230. When the routing scheme corresponding to the symmetrical network segment is a static routing scheme, the traffic to be forwarded flowing into any source leaf switch in the symmetrical network segment is parsed to obtain the destination leaf switch to which the traffic to be forwarded needs to be forwarded, and the target southbound network outgoing interface group on the destination leaf switch.

[0050] In this embodiment, the traffic to be forwarded can be understood as a data stream flowing into a symmetric network segment. It may carry information such as the source southbound network interface address and the destination southbound network interface address, as well as information such as the southbound network ingress interface number and the southbound network egress interface number.

[0051] In this step, specifically, the traffic to be forwarded can be parsed to obtain information such as the source and destination southbound network interface addresses, or the southbound inbound and outbound interface numbers carried within the traffic. Then, based on the destination southbound network interface address or the southbound outbound interface number carried within the traffic, the destination leaf switch to which the traffic should be forwarded, and the target southbound outbound interface group on the destination leaf switch to which the traffic should be forwarded, can be determined.

[0052] Optionally, before implementing southbound network load balancing using a static routing scheme corresponding to the symmetrical network segments, the method further includes: establishing a one-to-one correspondence between each southbound network outgoing interface group and each link on the destination leaf switch. Specifically, the number of southbound network outgoing interfaces in each southbound network outgoing interface group is a power of 2, and each southbound network outgoing interface group contains the same number of southbound network outgoing interfaces.

[0053] by Figure 3 Taking the symmetrical network segment shown as an example, this symmetrical network segment includes a set of source leaf switches and destination leaf switches. Each source leaf switch and destination leaf switch has 64 southbound network interfaces, and 32 links are deployed between the source leaf switches and the destination leaf switches. Based on Figure 3 The symmetrical network segment shown can establish a correspondence between the 0th southbound network outgoing interface group and the 0th link, the 1st southbound network outgoing interface group and the 1st link, and so on, until the 31st southbound network outgoing interface group and the 31st link are established.

[0054] S240. Determine a first link corresponding to the target southbound network outgoing interface group, and send the traffic to be forwarded to the target southbound network outgoing interface group along the first link.

[0055] In this step, specifically, since one southbound network outgoing interface group on the destination leaf switch corresponds to one link, a first link corresponding to the target southbound network outgoing interface group can be determined based on the correspondence between each southbound network outgoing interface group and each link, as well as the target southbound network outgoing interface group.

[0056] For example, such as Figure 3 As shown, when the target southbound network outgoing interface group is southbound network outgoing interface group 0, the first link corresponding to the target southbound network outgoing interface group is determined to be link 0. When the target southbound network outgoing interface group is southbound network outgoing interface group 1, the first link corresponding to the target southbound network outgoing interface group is determined to be link 1. When the target southbound network outgoing interface group is southbound network outgoing interface group 31, the first link corresponding to the target southbound network outgoing interface group is determined to be link 31.

[0057] S250. When the routing scheme corresponding to the symmetrical network segment is a policy routing scheme, the traffic to be forwarded flowing into any source leaf switch in the symmetrical network segment is parsed to obtain the first southbound network ingress interface group that the traffic to be forwarded passes through when it enters the source leaf switch.

[0058] In this step, specifically, the traffic to be forwarded can be parsed to obtain information such as the source and destination southbound network interface addresses, or the southbound ingress and egress interface numbers carried within the traffic. Then, based on the source or ingress interface address carried within the traffic, the source leaf switch into which the traffic enters, and the first southbound ingress interface group through which the traffic passes when entering the source leaf switch, can be determined.

[0059] Optionally, before adopting a policy routing scheme corresponding to the symmetrical network segment to achieve southbound network load balancing, the method further includes: establishing a one-to-one correspondence between each southbound network ingress interface group and each link on the source-leaf switch. Specifically, the number of southbound network ingress interfaces in each southbound network ingress interface group is a power of 2, and each southbound network ingress interface group contains the same number of southbound network ingress interfaces.

[0060] by Figure 4 Taking the symmetrical network segment shown as an example, this symmetrical network segment includes a set of source leaf switches and destination leaf switches. Each source leaf switch and destination leaf switch has 64 southbound network interfaces, and 32 links are deployed between the source leaf switches and the destination leaf switches. Based on Figure 4 The symmetrical network segment shown can establish a correspondence between the southbound network ingress interface group 0 and link 0, the southbound network ingress interface group 1 and link 1, and so on, until a correspondence between the southbound network ingress interface group 31 and link 31 is established.

[0061] S260. Determine a second link corresponding to the first southbound network ingress interface group, and send the traffic to be forwarded along the second link to the corresponding destination leaf switch.

[0062] In this step, specifically, since one southbound network ingress interface group on the source leaf switch corresponds to one link, a second link corresponding to the first southbound network ingress interface group can be determined based on the correspondence between each southbound network ingress interface group and each link, as well as the first southbound network ingress interface group.

[0063] For example, such as Figure 4As shown, when the first southbound network ingress interface group is southbound network ingress interface group 0, the second link corresponding to the first southbound network ingress interface group is determined to be link 0. When the first southbound network ingress interface group is southbound network ingress interface group 1, the second link corresponding to the first southbound network ingress interface group is determined to be link 1. When the first southbound network ingress interface group is southbound network ingress interface group 31, the second link corresponding to the first southbound network ingress interface group is determined to be link 31.

[0064] The technical solution of this embodiment obtains symmetrical network segments with leaf-leaf networking type in the southbound network when the southbound network adopts only leaf-leaf networking. Each symmetrical network segment contains at least one set of source leaf switches and destination leaf switches, and multiple links are configured between each set of source leaf switches and destination leaf switches. By adopting a static routing scheme or policy routing scheme corresponding to the symmetrical network segment with leaf-leaf networking type, load balancing of the symmetrical network segment with leaf-leaf networking type can be achieved, which can improve the load balancing efficiency of the southbound network with leaf-leaf networking type while ensuring its load balancing effect.

[0065] Example 3 Figure 5 This is a flowchart of another load balancing method for a southbound network of a supernode system according to Embodiment 3 of the present invention. This embodiment is a further optimization and extension based on the above embodiments and can be combined with various optional technical solutions in the above embodiments.

[0066] Figure 5 This is a flowchart of another load balancing method for a southbound network of a supernode system according to Embodiment 3 of the present invention. This embodiment is a refinement based on the above embodiments, and specifically describes the case where the southbound network only supports leaf-ridge networking.

[0067] like Figure 5 As shown in the figure, this embodiment discloses a load balancing method for a southbound network of a supernode system, including S310-S3100.

[0068] S310. Obtain the southbound network in the supernode system, and the network topology used by the southbound network.

[0069] S320. When the southbound network adopts only leaf-spine networking, obtain symmetrical network segments with leaf-spine networking in the southbound network; wherein each symmetrical network segment contains at least two leaf switches and one spine switch, and multiple links are configured between each leaf switch and each spine switch.

[0070] In this embodiment, the number of southbound network interfaces on each switch in the symmetrical network segment is a power of 2, and the number of links between each interconnecting switch is also a power of 2. For example, a network segment containing two sets of source leaf switches and destination leaf switches, with 64 southbound network interfaces on both source and destination leaf switches, and 16 links between each leaf switch and each spine switch, can be considered a symmetrical network segment of the leaf-spine networking type. Alternatively, a network segment containing two sets of source and destination leaf switches, with 32 southbound network interfaces on both source and destination leaf switches, and 8 links between each leaf switch and each spine switch, can also be considered a symmetrical network segment of the leaf-spine networking type.

[0071] S330. When the routing scheme corresponding to the symmetrical network segment is a static routing scheme, the traffic to be forwarded flowing into any source leaf switch in the symmetrical network segment is parsed to obtain the destination leaf switches to which the traffic to be forwarded needs to be forwarded, and the corresponding first southbound network outgoing interface group on each destination leaf switch.

[0072] In this step, specifically, the traffic to be forwarded can be parsed to obtain information such as the source southbound network interface address and destination southbound network interface address, or the southbound network ingress interface number and southbound network egress interface number carried within the traffic. Then, based on the destination southbound network interface address or southbound network egress interface number carried within the traffic, the destination leaf switch to which the traffic should be forwarded, and the first southbound network egress interface group on the destination leaf switch to which the traffic should be forwarded, can be determined.

[0073] Optionally, before implementing southbound network load balancing using a static routing scheme corresponding to the symmetrical network segments, the following steps are also included: establishing the correspondence between each southbound network outgoing interface group, each spine switch, each incoming link, and each outgoing link. Specifically, the number of southbound network outgoing interfaces within each southbound network outgoing interface group is a power of 2, and each southbound network outgoing interface group contains the same number of southbound network outgoing interfaces. The first half of the southbound network outgoing interface groups of each leaf switch corresponds to the first half of the spine switches, and the second half of the southbound network outgoing interface groups of each leaf switch corresponds to the second half of the spine switches. For example, in the case of 4 leaf switches and 2 spine switches, it can be determined that the first half of the southbound network outgoing interface groups of each leaf switch corresponds to spine switch 0, and the second half of the southbound network outgoing interface groups of each leaf switch corresponds to spine switch 1.

[0074] by Figure 6Taking the symmetrical network segment shown as an example, this symmetrical network segment includes leaf switch 0, leaf switch 1, leaf switch 2, leaf switch 3, spine switch 0, and spine switch 1. The four leaf switches are interconnected with the two spine switches via 16 links. Specifically, the 16 links connecting each leaf switch to spine switch 0 are numbered 0-15, and the 16 links connecting to spine switch 1 are numbered 16-31. Similarly, the 16 links connecting each spine switch to leaf switch 0 are numbered 0-15, the 16 links connecting to leaf switch 1 are numbered 16-31, the 16 links connecting to leaf switch 2 are numbered 32-47, and the 16 links connecting to leaf switch 3 are numbered 48-63. Each leaf switch has 64 southbound network interfaces.

[0075] based on Figure 6 The symmetrical network segments shown indicate that the first half of the southbound network outgoing interface groups of each leaf switch corresponds to spine switch 0, and the second half of the southbound network outgoing interface groups of each leaf switch corresponds to spine switch 1.

[0076] Then, it can be based on Figure 6 The symmetrical network segment shown establishes the correspondence between each southbound network outgoing interface group and each incoming link. Taking leaf switch 0 as an example, traffic entering from leaf switch 0 and destined for outgoing interface group 0 can be routed to link 0. Outgoing interface group 0 includes six southbound network outgoing interfaces: outgoing interfaces 0 and 1 connected to leaf switches 1, 2, and 3. Traffic entering from leaf switch 0 and destined for outgoing interface group 1 can be routed to link 1. Outgoing interface group 1 includes six southbound network outgoing interfaces: outgoing interfaces 2 and 3 connected to leaf switches 1, 2, and 3. This continues until traffic entering from leaf switch 0 and destined for outgoing interface group 31 is routed to link 31. Outgoing interface group 31 includes six southbound network outgoing interfaces: southbound network outgoing interfaces 62 and 63 connected to leaf switches 1, 2, and 3.

[0077] Finally, it can be based on Figure 6 For the symmetrical network segment shown, establish the correspondence between each southbound network outgoing interface group and each outgoing link. Taking the spine switch corresponding to each outgoing link as spine switch 0 as an example, traffic to southbound network outgoing interfaces 0 and 1 of leaf switch 0 can be routed to link 0, traffic to southbound network outgoing interfaces 2 and 3 of leaf switch 0 can be routed to link 1, and so on, traffic to southbound network outgoing interfaces 30 and 31 of leaf switch 0 can be routed to link 15.

[0078] Traffic destined for southbound network interfaces 0 and 1 of leaf switch 1 will be routed to link 16. Traffic destined for southbound network interfaces 2 and 3 of leaf switch 1 will be routed to link 17, and so on. Traffic destined for southbound network interfaces 30 and 31 of leaf switch 1 will be routed to link 31.

[0079] Traffic destined for southbound network interfaces 0 and 1 of leaf switch 2 will be routed to link 32. Traffic destined for southbound network interfaces 2 and 3 of leaf switch 2 will be routed to link 33, and so on. Traffic destined for southbound network interfaces 30 and 31 of leaf switch 2 will be routed to link 47.

[0080] Traffic destined for southbound network interfaces 0 and 1 of leaf switch 3 is routed to link 48; traffic destined for southbound network interfaces 2 and 3 of leaf switch 3 is routed to link 49; and so on, traffic destined for southbound network interfaces 30 and 31 of leaf switch 3 is routed to link 63.

[0081] Taking spine switch 1 as an example, traffic destined for southbound network interfaces 32 and 33 of leaf switch 0 is routed to link 0; traffic destined for southbound network interfaces 34 and 35 of leaf switch 0 is routed to link 1; and so on, traffic destined for southbound network interfaces 62 and 63 of leaf switch 0 is routed to link 15. Similarly, traffic destined for southbound network interfaces 32 and 33 of leaf switch 1 is routed to link 16; traffic destined for southbound network interfaces 34 and 35 of leaf switch 1 is routed to link 17; and so on, traffic destined for southbound network interfaces 62 and 63 of leaf switch 1 is routed to link 31.

[0082] Traffic destined for southbound network interfaces 32 and 33 of leaf switch 2 will be routed to link 32. Traffic destined for southbound network interfaces 34 and 35 of leaf switch 2 will be routed to link 33, and so on. Traffic destined for southbound network interfaces 62 and 63 of leaf switch 2 will be routed to link 47. Traffic destined for southbound network interfaces 32 and 33 of leaf switch 3 will be routed to link 48. Traffic destined for southbound network interfaces 34 and 35 of leaf switch 3 will be routed to link 49, and so on. Traffic destined for southbound network interfaces 62 and 63 of leaf switch 3 will be routed to link 63.

[0083] S340. Based on the first southbound network outgoing interface group, determine the first target spine switch corresponding to the source leaf switch, and the first incoming link corresponding to the first target spine switch.

[0084] In this step, specifically, based on the first southbound network outgoing interface group and the correspondence between each southbound network outgoing interface group, each spine switch, and each incoming link, the first target spine switch corresponding to the source leaf switch, and the first incoming link corresponding to the first target spine switch, can be determined. The first incoming link can be understood as the link through which the traffic to be forwarded flows from the source leaf switch into the first target spine switch.

[0085] S350. Based on the one-to-one correspondence between each outgoing link on the first target spine switch and each southbound network outgoing interface group on each destination leaf switch, and the first southbound network outgoing interface group, determine the first outgoing link corresponding to the first target spine switch.

[0086] The first outgoing link can be understood as the link through which the traffic to be forwarded flows from the first target spine switch into the destination leaf switch.

[0087] S360. Through the source leaf switch, the traffic to be forwarded is sent to the first target spine switch along the first inbound link, and through the first target spine switch, the traffic to be forwarded is sent to the first southbound network outbound interface group along the first outbound link.

[0088] For example, assuming the source leaf switch is leaf switch 0, the destination leaf switches include leaf switches 1, 2, and 3, and the first southbound network outgoing interface group consists of six southbound network outgoing interfaces, namely southbound network outgoing interfaces 0 and 1 connected to leaf switches 1, 2, and 3, it can be determined that the first target spine switch is spine switch 0, the first inbound link is link 0 between leaf switch 0 and spine switch 0, the first outbound link is link 16 between spine switch 0 and leaf switch 1, link 32 between spine switch 0 and leaf switch 2, and link 48 between spine switch 0 and leaf switch 3.

[0089] S370. When the routing scheme corresponding to the symmetrical network segment is a policy routing scheme, the traffic to be forwarded flowing into any source leaf switch in the symmetrical network segment is parsed to obtain the second southbound network ingress interface group that the traffic to be forwarded passes through when it enters the source leaf switch, and the destination leaf switch to which the traffic to be forwarded needs to be forwarded.

[0090] In this step, specifically, the traffic to be forwarded can be parsed to obtain information such as the source and destination southbound network interface addresses, or the southbound ingress and egress interface numbers, carried within the traffic. Then, based on the source or ingress interface address, the source leaf switch into which the traffic enters, and the second-direction ingress interface group it traverses upon entering the source leaf switch, can be determined. Finally, based on the destination southbound interface address or egress interface number, the destination leaf switch to which the traffic should be forwarded can be determined.

[0091] Optionally, the traffic to be forwarded flowing into any source leaf switch within the symmetric network segment is parsed to obtain the second southbound network ingress interface group through which the traffic to be forwarded passes when entering the source leaf switch, and the destination leaf switch to which the traffic to be forwarded needs to be forwarded. This includes: when the southbound network uses Internet Protocol (IP), parsing the traffic to be forwarded flowing into any source leaf switch within the symmetric network segment to obtain the ingress interface IP address and destination IP address corresponding to the traffic to be forwarded; determining the second southbound network ingress interface group through which the traffic to be forwarded passes when entering the source leaf switch based on the ingress interface IP address; and determining the destination leaf switch to which the traffic to be forwarded needs to be forwarded based on the destination IP address.

[0092] Optionally, before implementing southbound network load balancing using a policy routing scheme corresponding to the symmetrical network segments, the following steps are also included: establishing the correspondence between each southbound network ingress interface group, each spine switch, each ingress link, and each outgress link. Specifically, the number of southbound network ingress interfaces within each southbound network ingress interface group is a power of 2, and each southbound network ingress interface group contains the same number of southbound network ingress interfaces. The first half of the southbound network ingress interface groups of each leaf switch corresponds to the first half of the spine switches, and the second half of the southbound network ingress interface groups of each leaf switch corresponds to the second half of the spine switches.

[0093] by Figure 7 Taking the symmetrical network segment shown as an example, it can be determined that the first half of the southbound network ingress interface group of each leaf switch corresponds to spine switch 0, and the second half of the southbound network ingress interface group of each leaf switch corresponds to spine switch 1.

[0094] Then, it can be based on Figure 6 The symmetrical network segment shown establishes a correspondence between each southbound network ingress interface group and each ingress link. Taking leaf switch 0 as an example, traffic coming from southbound network ingress interfaces 0 and 1 connected to leaf switch 0 can be routed to link 0. Traffic coming from southbound network ingress interfaces 2 and 3 connected to leaf switch 0 can be routed to link 1. Similarly, traffic coming from southbound network ingress interfaces 62 and 63 connected to leaf switch 0 can be routed to link 31.

[0095] Finally, it can be based on Figure 7 The symmetrical network segment shown establishes the correspondence between each southbound network ingress interface group and each outgress link. Taking the 0th spine switch as an example, traffic coming from the 0th and 1st southbound network ingress interfaces connected to the 0th leaf switch and destined for the 1st leaf switch can be routed to link 16. Traffic coming from the 2nd and 3rd southbound network ingress interfaces connected to the 0th leaf switch and destined for the 1st leaf switch can be routed to link 17. Similarly, traffic coming from the 30th and 31st southbound network ingress interfaces connected to the 0th leaf switch and destined for the 1st leaf switch can be routed to link 31.

[0096] Traffic originating from southbound network ingress interfaces 0 and 1 connected to leaf switch 0 and destined for leaf switch 2 will be routed to link 32. Traffic originating from southbound network ingress interfaces 2 and 3 connected to leaf switch 0 and destined for leaf switch 2 will be routed to link 33. Similarly, traffic originating from southbound network ingress interfaces 30 and 31 connected to leaf switch 0 and destined for leaf switch 2 will be routed to link 47.

[0097] Traffic originating from southbound network ingress interfaces 0 and 1 connected to leaf switch 0 and destined for leaf switch 3 will be routed to link 48. Traffic originating from southbound network ingress interfaces 2 and 3 connected to leaf switch 0 and destined for leaf switch 3 will be routed to link 49. Similarly, traffic originating from southbound network ingress interfaces 30 and 31 connected to leaf switch 0 and destined for leaf switch 3 will be routed to link 63.

[0098] S380. Based on the second southbound network ingress interface group, determine the second target spine switch corresponding to the source leaf switch, and the second ingress link corresponding to the second target spine switch.

[0099] In this step, specifically, based on the second southbound network ingress interface group and the correspondence between each southbound network ingress interface group, each spine switch, and each ingress link, the second target spine switch corresponding to the source leaf switch, and the second ingress link corresponding to the second target spine switch, can be determined. The second ingress link can be understood as the link through which the traffic to be forwarded flows from the source leaf switch into the second target spine switch.

[0100] S390. Obtain each outgoing link between the second target spine switch and the destination leaf switch, and determine the second outgoing link corresponding to the second target spine switch based on the one-to-one correspondence between each southbound network ingress interface group and each outgoing link on the source leaf switch, as well as the second southbound network ingress interface group.

[0101] The second outgoing link can be understood as the link through which the traffic to be forwarded flows from the second target spine switch into the corresponding destination leaf switch.

[0102] S3100: The traffic to be forwarded is sent to the second target spine switch via the source leaf switch along the second inbound link, and then sent to the destination leaf switch via the second target spine switch along the second outbound link.

[0103] For example, assuming the source leaf switch is leaf switch 0, the second southbound network ingress interface group is southbound network ingress interface group 0, and the destination leaf switch is leaf switch 1, then the second target spine switch can be determined to be spine switch 0, the second ingress link is link 0 between leaf switch 0 and spine switch 0, and the second outgress link is link 16 between spine switch 0 and leaf switch 1.

[0104] The technical solution of this embodiment obtains symmetrical network segments of leaf-spine networking type in the southbound network when the networking type adopted by the southbound network is only leaf-spine networking. Each symmetrical network segment contains at least two leaf switches and one spine switch, and multiple links are configured between each leaf switch and each spine switch. By adopting a static routing scheme or policy routing scheme corresponding to the symmetrical network segment of leaf-spine networking type, load balancing of the symmetrical network segment of leaf-spine networking type can be achieved, which can improve the load balancing effect of the symmetrical network segment of leaf-spine networking type, thereby improving the overall load balancing performance of the southbound network.

[0105] Example 4 Figure 8 This is a schematic diagram of a load balancing device for a southbound network of a supernode system according to Embodiment 4 of the present invention. This embodiment is applicable to the situation of load balancing for a southbound network of a supernode system. The method can be executed by the load balancing device of the southbound network of the supernode system. The load balancing device of the southbound network of the supernode system can be implemented in hardware and / or software. The load balancing device of the southbound network of the supernode system can be configured in electronic devices such as the management server of the supernode system.

[0106] like Figure 8 As shown, the load balancing device for the southbound network of the supernode system disclosed in this embodiment includes a southbound network acquisition module 81, a network segment acquisition module 82, and a traffic balancing distribution module 83.

[0107] Among them, the southbound network acquisition module 81 is used to acquire the southbound network in the supernode system and the networking type adopted by the southbound network.

[0108] The network segment acquisition module 82 is used to acquire symmetrical network segments corresponding to each network type in the southbound network; wherein each symmetrical network segment contains at least one set of source leaf switches and destination leaf switches, and multiple links are configured between each set of source leaf switches and destination leaf switches.

[0109] The traffic balancing and allocation module 83 is used to distribute the traffic within each symmetrical network segment to the corresponding links using a static routing scheme or a policy routing scheme corresponding to each symmetrical network segment.

[0110] The static routing scheme is used to route traffic destined for different southbound network outgoing interface groups on the destination leaf switch to different links, while the policy routing scheme is used to route traffic coming from different southbound network incoming interface groups on the source leaf switch to different links.

[0111] The technical solution in this embodiment, through the cooperation of the southbound network acquisition module 81, the network segment acquisition module 82, and the traffic balancing allocation module 83, solves the problem that existing technologies for achieving load balancing of the southbound network of the supernode system using packet spraying or flow-by-flow hashing are prone to data corruption or uneven link load. It fully utilizes the symmetry of the southbound network of the supernode system in the network topology, avoids data transmission corruption, and ensures the load balancing effect of the southbound network of the supernode system.

[0112] Optionally, for a symmetrical network segment with a leaf-leaf network topology, the traffic balancing distribution module 83 includes a first static routing unit and a first policy routing unit.

[0113] The first static routing unit is used to parse the traffic to be forwarded into any source leaf switch in the symmetrical network segment, to obtain the destination leaf switch to which the traffic to be forwarded needs to be forwarded, and the target southbound network outgoing interface group on the destination leaf switch; to determine a first link corresponding to the target southbound network outgoing interface group, and to send the traffic to be forwarded to the target southbound network outgoing interface group along the first link.

[0114] The first policy routing unit is used to parse the traffic to be forwarded into any source leaf switch within the symmetric network segment, obtain the first southbound network ingress interface group through which the traffic to be forwarded enters the source leaf switch; determine a second link corresponding to the first southbound network ingress interface group, and send the traffic to be forwarded to the corresponding destination leaf switch along the second link.

[0115] Optionally, for a symmetrical network segment with a leaf-spine networking type, the traffic balancing distribution module 83 includes: a second static routing unit and a second policy routing unit.

[0116] The second static routing unit is used to parse the traffic to be forwarded from any source leaf switch within a symmetrical network segment, to obtain the destination leaf switches to which the traffic to be forwarded should be, and the corresponding first southbound network outgoing interface group on each destination leaf switch; based on the first southbound network outgoing interface group, to determine the first target spine switch corresponding to the source leaf switch, and the first incoming link corresponding to the first target spine switch; based on the one-to-one correspondence between each outgoing link on the first target spine switch and each southbound network outgoing interface group on each destination leaf switch, and the first southbound network outgoing interface group, to determine the first outgoing link corresponding to the first target spine switch; and through the source leaf switch, to send the traffic to be forwarded along the first incoming link to the first target spine switch, and through the first target spine switch, to send the traffic to be forwarded along the first outgoing link to the first southbound network outgoing interface group.

[0117] The second policy routing unit is used to parse the traffic to be forwarded into any source leaf switch within the symmetric network segment, obtain the second southbound network ingress interface group through which the traffic to be forwarded enters the source leaf switch, and the destination leaf switch to which the traffic to be forwarded needs to be forwarded; based on the second southbound network ingress interface group, determine the second target spine switch corresponding to the source leaf switch, and the second ingress link corresponding to the second target spine switch; obtain each outgress link between the second target spine switch and the destination leaf switch, and determine the second outgress link corresponding to the second target spine switch based on the one-to-one correspondence between each southbound network ingress interface group and each outgress link on the source leaf switch, and the second southbound network ingress interface group; send the traffic to be forwarded to the second target spine switch along the second ingress link through the source leaf switch, and send the traffic to be forwarded to the destination leaf switch along the second outgress link through the second target spine switch.

[0118] Optionally, the second policy routing unit is specifically used to: parse the traffic to be forwarded into any source leaf switch within the symmetric network segment to obtain the ingress interface IP address and destination IP address corresponding to the traffic to be forwarded; determine the second southbound network ingress interface group that the traffic to be forwarded passes through when entering the source leaf switch based on the ingress interface IP address; and determine the destination leaf switch to which the traffic to be forwarded needs to be forwarded based on the destination IP address.

[0119] Optionally, the device further includes a backup scheme activation module, which is used to: detect whether there is a faulty link in each symmetric network segment; identify the symmetric network segment with the faulty link as a faulty network segment, and use a dynamic routing scheme to re-determine the link corresponding to the traffic in the faulty network segment; wherein, the dynamic routing scheme is a scheme determined based on a dynamic routing protocol.

[0120] The load balancing device for the southbound network of the supernode system provided in this embodiment of the invention can execute the load balancing method for the southbound network of the supernode system provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method execution. Content not described in detail in this embodiment can be referred to the description in any method embodiment of this application.

[0121] Example 5 Figure 9 A schematic diagram of the structure of an electronic device 10 that can be used to implement embodiments of the present invention is shown. For example... Figure 9As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory 12 or a random access memory 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the read-only memory 12 or loaded from storage unit 18 into the random access memory 13. The random access memory 13 may also store various programs and data required for the operation of the electronic device 10. The processor 11, read-only memory 12, and random access memory 13 are interconnected via a bus 14. An input / output interface 15 is also connected to the bus 14.

[0122] Multiple components in electronic device 10 are connected to input / output interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of monitors, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0123] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as load balancing methods for southbound networks in a supernode system.

[0124] In some embodiments, the load balancing method for the southbound network of the supernode system can be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 10 via read-only memory 12 and / or communication unit 19. When the computer program is loaded into random access memory 13 and executed by processor 11, one or more steps of the load balancing method for the southbound network of the supernode system described above can be performed. Alternatively, in other embodiments, processor 11 can be configured to perform the load balancing method for the southbound network of the supernode system by any other suitable means (e.g., by means of firmware).

[0125] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0126] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0127] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory, read-only memory, erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0128] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0129] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0130] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0131] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0132] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A load balancing method for a southbound network of a supernode system, characterized in that, The load balancing method for the southbound network of the supernode system includes: Obtain the southbound network in the supernode system, and the network topology used by the southbound network; In the southbound network, symmetrical network segments corresponding to each network type are obtained; wherein, each symmetrical network segment contains at least one set of source leaf switches and destination leaf switches, and multiple links are configured between each set of source leaf switches and destination leaf switches. The traffic within each symmetrical network segment is evenly distributed to the corresponding links using either a static routing scheme or a policy routing scheme corresponding to each symmetrical network segment. The static routing scheme is used to route traffic destined for different southbound network outgoing interface groups on the destination leaf switch to different links, and the policy routing scheme is used to route traffic coming from different southbound network incoming interface groups on the source leaf switch to different links.

2. The load balancing method for the southbound network of a supernode system according to claim 1, characterized in that, For symmetric network segments with a leaf-leaf topology, a static routing scheme corresponding to the symmetric network segment is adopted to evenly distribute traffic within the symmetric network segment to the corresponding links, including: The traffic to be forwarded into any source leaf switch within a symmetric network segment is parsed to obtain the destination leaf switch to which the traffic to be forwarded needs to be forwarded, and the target southbound network outgoing interface group on the destination leaf switch. A first link corresponding to the target southbound network outgoing interface group is determined, and the traffic to be forwarded is sent to the target southbound network outgoing interface group along the first link.

3. The load balancing method for the southbound network of a supernode system according to claim 1, characterized in that, For symmetric network segments with a leaf-leaf topology, a policy routing scheme corresponding to the symmetric network segment is adopted to evenly distribute traffic within the symmetric network segment to the corresponding links, including: The traffic to be forwarded into any source leaf switch within a symmetrical network segment is analyzed to obtain the first southbound network ingress interface group that the traffic to be forwarded passes through when it enters the source leaf switch. A second link corresponding to the first southbound network ingress interface group is determined, and the traffic to be forwarded is sent to the corresponding destination leaf switch along the second link.

4. The load balancing method for the southbound network of a supernode system according to claim 1, characterized in that, For symmetrical network segments with a leaf-spine topology, a static routing scheme corresponding to the symmetrical network segment is adopted to evenly distribute traffic within the symmetrical network segment to the corresponding links, including: The traffic to be forwarded into any source leaf switch in the symmetric network segment is parsed to obtain the destination leaf switches to which the traffic to be forwarded needs to be forwarded, and the corresponding first southbound network outgoing interface group on each destination leaf switch. Based on the first southbound network outgoing interface group, determine the first target spine switch corresponding to the source leaf switch, and the first incoming link corresponding to the first target spine switch; Based on the one-to-one correspondence between each outgoing link on the first target spine switch and each southbound network outgoing interface group on each destination leaf switch, and the first southbound network outgoing interface group, the first outgoing link corresponding to the first target spine switch is determined. The traffic to be forwarded is sent to the first target spine switch via the source leaf switch along the first inbound link, and then sent to the first southbound network outbound interface group via the first target spine switch along the first outbound link.

5. The load balancing method for the southbound network of a supernode system according to claim 1, characterized in that, For symmetrical network segments with a leaf-spine topology, a policy-based routing scheme corresponding to the symmetrical network segment is adopted to evenly distribute traffic within the symmetrical network segment to the corresponding links, including: The traffic to be forwarded into any source leaf switch within a symmetrical network segment is parsed to obtain the second southbound network ingress interface group that the traffic to be forwarded passes through when it enters the source leaf switch, and the destination leaf switch to which the traffic to be forwarded needs to be forwarded. Based on the second southbound network ingress interface group, determine the second target spine switch corresponding to the source leaf switch, and the second ingress link corresponding to the second target spine switch; Obtain each outgoing link between the second target spine switch and the destination leaf switch, and determine the second outgoing link corresponding to the second target spine switch based on the one-to-one correspondence between each southbound network ingress interface group on the source leaf switch and each outgoing link, as well as the second southbound network ingress interface group. The traffic to be forwarded is sent to the second target spine switch via the source leaf switch along the second inbound link, and then sent to the destination leaf switch via the second target spine switch along the second outbound link.

6. The load balancing method for the southbound network of a supernode system according to claim 5, characterized in that, When the southbound network uses Internet Protocol (IP), the traffic to be forwarded flowing into any source leaf switch within the symmetric network segment is parsed to obtain the second southbound network ingress interface group that the traffic to be forwarded passes through when entering the source leaf switch, and the destination leaf switch to which the traffic to be forwarded needs to be forwarded, including: The traffic to be forwarded into any source leaf switch within a symmetrical network segment is parsed to obtain the ingress interface IP address and destination IP address corresponding to the traffic to be forwarded. Based on the ingress interface IP address, determine the second southbound network ingress interface group that the traffic to be forwarded passes through when entering the source leaf switch; Based on the destination IP address, determine the destination leaf switch to which the traffic to be forwarded should be forwarded.

7. The load balancing method for the southbound network of a supernode system according to claim 1, characterized in that, After employing a static routing scheme or policy routing scheme corresponding to each symmetric network segment to evenly distribute traffic within each symmetric network segment to the corresponding links, the process also includes: Detect whether there are faulty links in each symmetric network segment; Symmetrical network segments with faulty links are identified as faulty network segments, and a dynamic routing scheme is used to redetermine the links corresponding to the traffic within the faulty network segments. The dynamic routing scheme is a scheme determined based on a dynamic routing protocol.

8. A load balancing device for a southbound network of a supernode system, characterized in that, The load balancing device for the southbound network of the supernode system includes: The southbound network acquisition module is used to acquire the southbound network in the supernode system, as well as the network topology adopted by the southbound network; The network segment acquisition module is used to acquire symmetrical network segments corresponding to each networking type in the southbound network; wherein each symmetrical network segment contains at least one set of source leaf switches and destination leaf switches, and multiple links are configured between each set of source leaf switches and destination leaf switches. The traffic balancing and distribution module is used to distribute the traffic within each symmetrical network segment to the corresponding links using a static routing scheme or a policy routing scheme corresponding to each symmetrical network segment. The static routing scheme is used to route traffic destined for different southbound network outgoing interface groups on the destination leaf switch to different links, and the policy routing scheme is used to route traffic coming from different southbound network incoming interface groups on the source leaf switch to different links.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and a memory communicatively connected to said at least one processor; The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the load balancing method for the southbound network of the supernode system according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that are used to cause a processor to execute the load balancing method for the southbound network of the supernode system according to any one of claims 1-7.

Images