WO2023173720A1 - 应用访问方法、云端代理及节点代理组件、设备、介质 - Google Patents

应用访问方法、云端代理及节点代理组件、设备、介质 Download PDF

Info

Publication number
WO2023173720A1
WO2023173720A1 PCT/CN2022/121851 CN2022121851W WO2023173720A1 WO 2023173720 A1 WO2023173720 A1 WO 2023173720A1 CN 2022121851 W CN2022121851 W CN 2022121851W WO 2023173720 A1 WO2023173720 A1 WO 2023173720A1
Authority
WO
WIPO (PCT)
Prior art keywords
node
application
routing table
application access
global routing
Prior art date
Application number
PCT/CN2022/121851
Other languages
English (en)
French (fr)
Inventor
杨勇
Original Assignee
苏州浪潮智能科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 苏州浪潮智能科技有限公司 filed Critical 苏州浪潮智能科技有限公司
Publication of WO2023173720A1 publication Critical patent/WO2023173720A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/12—Discovery or management of network topologies
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00—Routing or path finding of packets in data switching networks
    • H04L45/02—Topology update or discovery
    • H04L45/021—Ensuring consistency of routing table updates, e.g. by using epoch numbers
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L61/00—Network arrangements, protocols or services for addressing or naming
    • H04L61/09—Mapping addresses
    • H04L61/25—Mapping addresses of the same type
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/10—Protocols in which an application is distributed across nodes in the network
    • H04L67/104—Peer-to-peer [P2P] networks
    • H04L67/1061—Peer-to-peer [P2P] networks using node-based peer discovery mechanisms
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/50—Network services
    • H04L67/56—Provisioning of proxy services
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L69/00—Network arrangements, protocols or services independent of the application payload and not provided for in the other groups of this subclass
    • H04L69/16—Implementation or adaptation of Internet protocol [IP], of transmission control protocol [TCP] or of user datagram protocol [UDP]
    • H04L69/164—Adaptation or special uses of UDP protocol

Definitions

  • This application relates to the field of cloud service technology, and in particular to an application access method, cloud agent and node agent components, equipment, and media.
  • cluster nodes are interconnected, and mutual access between applications is achieved through the service object of k8s (an example of a well-architected distributed system).
  • the service is label (label) matches the backend pod (a group of collocated containers), kube-proxy (a network proxy component) configures local iptables (packet filtering system) rules to forward traffic directed to the service ip to the backend pod ip
  • cni container network interface, container network interface
  • a common situation is that a bearer
  • the control nodes of the edge computing k8s cluster are deployed on a public network and expose a public network IP.
  • the computing nodes are distributed in different sites and in different LANs.
  • the nodes in the same LAN are interconnected and the nodes across the network are interconnected. It is impossible to communicate directly, but all nodes in the LAN can communicate with the public network IP exposed by the control node; in this network topology, application pods across the LAN cannot communicate with each other through the traditional k8s service+cni.
  • the purpose of this application is to provide an application access method, cloud agent and node agent components, equipment, and media that can realize application access across edge nodes and reduce the load of centralized traffic agents.
  • the specific plan is as follows:
  • this application discloses an application access method, which is applied to the cloud proxy component, including:
  • the node discovery result is the node discovery result obtained by the node agent component in the newly registered edge node performing node discovery on each edge node in the current cluster;
  • the node discovery result is that no node can be found, perform a UDP hole drilling operation between the LAN where the newly registered edge node is located and the existing LAN to obtain the interoperability relationship between the LANs, and update the global routing table based on the interoperability relationship;
  • the node information of the newly registered edge node is associated with the target LAN information to update the global routing table, where the target LAN information is the LAN information of the LAN where the successfully discovered node is located;
  • the global routing table is delivered to the node agent component in each edge node, so that the node agent component in each edge node forwards application access traffic based on the global routing table to achieve application access.
  • the cluster node list is delivered to the node agent component in the newly registered edge node, so that the node agent component performs node discovery on each edge node in the current cluster based on the cluster node list and obtains the node discovery result.
  • Optional also includes:
  • the node information includes the service discovery port of the node agent component; accordingly, the node agent component accesses each edge node in the current cluster based on the service discovery port of the node agent component of each edge node in the cluster node list.
  • the node discovery result is that no node was found, it also includes:
  • the node agent component in each edge node determines that the local area network where the local node is located and the LAN where the destination node is located are the same LAN based on the global routing table, it will directly forward the application access traffic to the node agent component in the destination node so that the destination node The node proxy component in forwards application access traffic to the destination application;
  • the node agent component in each edge node determines based on the global routing table that the interoperability relationship between the local area network where the local node is located and the LAN where the destination node is located is capable of tunneling through UDP, then the application access traffic will be forwarded to the destination through the NAT gateway of the LAN where the destination node is located.
  • the node proxy component in the node allows the node proxy component in the destination node to forward application access traffic to the destination application.
  • Optional also includes:
  • the target node agent component is a node agent component that determines based on the global routing table that the interoperability relationship between the local area network where the local node is located and the LAN where the destination node is located cannot communicate through UDP hole drilling.
  • this application discloses an application access method, which is applied to the node agent component, including:
  • the local node When the local node performs node registration, perform node discovery on each edge node in the current cluster and obtain the node discovery results;
  • the cloud proxy component Send the node discovery result to the cloud proxy component so that the cloud proxy component updates the global routing table based on the node discovery result. If the node discovery result is that no node can be found, the cloud proxy component will perform the new registration of the edge node where the local area network is the same as the current one. There is already a UDP hole punching operation between LANs to obtain the interoperability relationship between LANs, and update the global routing table based on the interoperability relationship; if the node discovery result is that the node is successfully discovered, the cloud proxy component will combine the node information of the newly registered edge node with the target LAN information is associated to update the global routing table, and the target LAN information is the LAN information of the LAN where the successfully discovered node is located;
  • this application discloses a cloud proxy component, including:
  • the node discovery result acquisition module is used to obtain the node discovery result; wherein the node discovery result is the node discovery result obtained by the node agent component in the newly registered edge node performing node discovery on each edge node in the current cluster;
  • the global routing table update module is used to perform UDP hole-punching operations between the LAN where the newly registered edge node is located and the existing LAN to obtain the interoperability relationship between the LANs if the node discovery result is that no node is found, and based on interoperability
  • the relationship updates the global routing table; if the node discovery result is a successfully discovered node, associate the node information of the newly registered edge node with the target LAN information to update the global routing table, where the target LAN information is the LAN of the LAN where the successfully discovered node is located. information;
  • the global routing table delivery module is used to deliver the global routing table to the node agent component in each edge node, so that the node agent component in each edge node forwards application access traffic based on the global routing table to achieve application access.
  • this application discloses a node agent component, including:
  • the node discovery module is used to perform node discovery on each edge node in the current cluster when the local node performs node registration, and obtain the node discovery results;
  • the node discovery result sending module is used to send the node discovery result to the cloud proxy component so that the cloud proxy component updates the global routing table based on the node discovery result. If the node discovery result is that no node can be found, the cloud proxy component updates the node discovery result. UDP hole-punching operation between the LAN where the registered edge node is located and the existing LAN to obtain the interoperability relationship between the LANs, and update the global routing table based on the interoperability relationship; if the node discovery result is that the node is successfully discovered, the cloud proxy component will be newly registered
  • the node information of the edge node is associated with the target LAN information to update the global routing table.
  • the target LAN information is the LAN information of the LAN where the successfully discovered node is located;
  • the global routing table acquisition module is used to obtain the global routing table issued by the cloud proxy component, and forward application access traffic based on the global routing table to achieve application access.
  • this application discloses an electronic device including a processor and a memory; wherein,
  • Memory used to hold computer programs
  • the processor is used to execute the computer program in the aforementioned application access method.
  • the present application discloses a computer non-volatile readable storage medium for storing a computer program, wherein when the computer program is executed by a processor, the aforementioned application access method is implemented.
  • FIG. 1 is a schematic diagram of an edge computing network provided by this application.
  • Figure 2 is a flow chart of an application access method provided by this application.
  • Figure 3 is a schematic diagram of a component deployment architecture provided by this application.
  • Figure 4 is a schematic diagram of node discovery of a new registered node provided by this application.
  • FIG. 5 is a schematic diagram of a NAT provided by this application.
  • Figure 6 is a schematic diagram of UDP hole drilling for a new edge LAN provided by this application.
  • Figure 7 is a global routing diagram provided by this application.
  • Figure 8 is a schematic diagram of application access within the same local area network provided by this application.
  • Figure 9 is a schematic diagram of a cross-LAN application provided by this application to access through UDP holes;
  • Figure 10 is a schematic diagram of a cross-LAN application provided by this application accessed through the cloud proxy component;
  • Figure 11 is a flow chart of an application access method provided by this application.
  • Figure 12 is a schematic structural diagram of a cloud proxy component provided by this application.
  • Figure 13 is a schematic structural diagram of a node agent component provided by this application.
  • Figure 14 is a structural diagram of an electronic device provided by this application.
  • each Pod has an independent IP address, and it is assumed that all Pods can be in a directly connected, flat network space.
  • connecting the container network plane between the cloud center and the edge data center is expensive and has low performance.
  • the access links between the services deployed in the edge data center and the cloud edge are limited and specific, and the demand is smaller than the demand for Pod mutual access within the edge data center.
  • the container network converges in each data center, and uses high-availability public network gateway + public network IP to expose edge services.
  • Another factor that determines the plane scope of the container network is the node network.
  • the container network is based on the node network.
  • FIG. 1 is a schematic diagram of an edge computing network provided by an embodiment of the present application.
  • api-server is the module used to provide external API services in the k8s cluster.
  • this application provides an application access solution that can realize application access across edge nodes and reduce the load of the centralized traffic proxy.
  • this embodiment of the present application discloses an application access method, which is applied to the cloud proxy component, including:
  • Step S11 Obtain the node discovery result; wherein, the node discovery result is the node discovery result obtained by the node agent component in the newly registered edge node performing node discovery on each edge node in the current cluster.
  • the cluster node list can be delivered to the node agent component in the newly registered edge node, so that the node agent component performs node discovery on each edge node in the current cluster based on the cluster node list and obtains the node discovery result; Get the node discovery results returned by the node agent component in the newly registered edge node.
  • the embodiment of the present application can also obtain the node information of the newly registered edge node sent by the node agent component in the newly registered edge node; add the node information to the cluster node list.
  • the node information includes the service discovery port of the node agent component; accordingly, the node agent component accesses each edge node in the current cluster based on the service discovery port of the node agent component of each edge node in the cluster node list. Further, the node information also includes the unique identifier of the node.
  • the node proxy component in the newly registered edge node listens to the configured local service discovery port, and configures local iptables rules to direct the traffic of the K8S service network segment to the local node proxy component, and at the same time connects to the exposed cloud proxy component.
  • Public network IP establish a long link channel to the cloud proxy component, generate a unique identifier of this node, and report this node information to the cloud proxy component.
  • the cloud agent component discovers that there is a new node registration, records the new node information, and at the same time delivers the node list of existing nodes in the cluster to the node agent component, triggering the automatic discovery process of the node.
  • the node agent component of the newly registered node receives the current cluster node list, enters the automatic discovery process, and attempts to use the IP of each node + the service discovery port of the edge agent component to initiate direct access. At this time, the nodes in the same LAN can return to normal Response, nodes that are not within a LAN cannot be directly connected.
  • k8s Korean
  • Google Open source container cluster management system. Based on Docker technology, it provides a series of complete functions such as deployment and operation, resource scheduling, service discovery and dynamic scaling for containerized applications, improving the convenience of large-scale container cluster management.
  • Service is a resource type in the k8s cluster.
  • the service type resource can provide domain name-based access addresses for pods in the k8s cluster, and can provide simple four-layer load balancing capabilities.
  • FIG. 3 is a schematic diagram of a component deployment architecture provided by an embodiment of the present application.
  • edge LAN 1 and edge LAN 2 as examples, but the number of LANs is not limited.
  • This application mainly includes two types of components, which run on the cloud nodes and edge computing nodes of the k8s cluster that carry edge computing services. The details are as follows: proxy-server (i.e., cloud proxy component), which runs on the cloud of the k8s cluster in the form of deployment. On the node, it communicates with the cluster api-server and exposes services to the outside world through the public IP.
  • proxy-server i.e., cloud proxy component
  • Its main functions include maintaining node information, generating and updating the global routing table based on the node discovery results, and delivering the global routing table to the proxy-agent (i.e., the node agent component). ), perform UDP hole punching between LANs, and relay traffic forwarding.
  • Proxy-agent runs on edge computing nodes in the form of a daemonset, with one instance running on each edge node. Its main functions include node information registration, LAN node discovery, local iptables rule configuration, and classification and forwarding based on the global routing table. The edge node is powered on.
  • the proxy-agent on the node After joining the cluster, the proxy-agent on the node automatically starts, listens to the configured local service discovery port, configures local iptables rules, directs the traffic of the service network segment to the local proxy-agent, and connects to the cloud proxy-agent at the same time.
  • the public network IP exposed by the server establishes a long link channel to the proxy-server, generates a unique ID of the node, and reports the node information to the cloud proxy-server.
  • the cloud proxy-server discovers that a new node has been registered, records the new node information, and sends a list of existing nodes in the cluster to the node's proxy-agent, triggering the automatic discovery process of the node.
  • the proxy-agent of the new node receives the current global node list, enters the automatic discovery process, and attempts to use the IP of each node + the service discovery port of the proxy-agent to initiate direct access. Nodes in the same LAN can return normal responses, but nodes not in the same LAN cannot be directly connected. You can get a view of the current new node and which existing nodes are in a network reachable area, as shown in Figure 4.
  • Figure 4 is a schematic diagram of node discovery of a new registered node provided by an embodiment of the present application. Take edge LAN 1 and edge LAN 2 as examples, but the number of LANs is not limited.
  • Step S12 If the node discovery result is that no node can be found, perform a UDP hole drilling operation between the LAN where the newly registered edge node is located and the existing LAN to obtain the interoperability relationship between the LANs, and update the global routing table based on the interoperability relationship. .
  • the embodiment of the present application can also obtain the public network NAT (i.e. Network Address Translation, Network Address Translation) gateway address of the local area network where the newly registered edge node is located; based on the public network NAT gateway address and the newly registered The node information of the edge node creates the LAN information of the LAN. Furthermore, based on the public network NAT gateway address, a UDP hole punching operation is performed between the LAN where the newly registered edge node is located and the currently existing LAN to obtain the interoperability relationship between the LANs.
  • NAT Network Address Translation, Network Address Translation
  • the global routing table includes the LAN information of each LAN in the cluster, as well as the interworking relationship between each LAN.
  • the LAN information includes the information of each node in this LAN.
  • NAT is an IETF (Internet Engineering Task Force) standard that allows an entire organization to appear on the Internet with a public IP (Internet Protocol) address. It is a method of converting an internal private network address (IP address ) into a legitimate network IP address. IP address ) into a legitimate network IP address. IP address ) into IP address. IP address ) into IP address. IP address ) into IP address. IP address ) into IP address. IP address ) into IP address.
  • IP address IP address
  • NAT Internet Protocol
  • IP address Internet Protocol
  • IP address Internet Protocol
  • Step S13 If the node discovery result is that the node is successfully discovered, associate the node information of the newly registered edge node with the target LAN information to update the global routing table, where the target LAN information is the LAN information of the LAN where the successfully discovered node is located.
  • the proxy-agent sends the node discovery result to the cloud proxy-server.
  • the end proxy-server receives the node discovery result. If it is a new node in an existing LAN, it associates the node information with the existing LAN information, updates the global routing table, and sends it to all proxy-agents.
  • the cloud proxy-server receives the node discovery result. If it is a new LAN node, it creates a new LAN information, obtains the public network NAT gateway address of the LAN, and communicates with the proxy-agent on other existing LAN nodes through the public network.
  • the network NAT gateway attempts to perform UDP hole punching operation.
  • an embodiment of the present application provides a schematic diagram of UDP hole drilling for a new edge LAN. Added edge LAN 1 and edge LAN 2 for UDP hole punching. If the newly added LAN cannot be accessed through UDP hole drilling, the embodiment of this application can maintain the long link channel of the node's proxy-agent in the cloud proxy-server, and subsequent traffic to the LAN is forwarded to the corresponding node through the cloud proxy-server. Within the long link channel.
  • the cloud proxy-server maintains the interoperability information between the newly added LAN and other LANs and the included node information, generates a global routing table, and distributes it to all proxy-agents.
  • Figure 7 is a global routing diagram provided by an embodiment of the present application. Taking LAN 1, LAN 2, and LAN 3 as examples, the global routing table includes LAN information of each LAN, as well as the routing information between the LANs. mutual relationship. The LAN information includes the information of each node in this LAN and the NAT gateway address.
  • the final generated global routing table describes the interoperability information between nodes in the edge computing cluster, including direct interoperability, interoperability through UDP holes, inability to interoperate through UDP holes, and nodes in the same LAN can directly interoperate.
  • Step S14 Deliver the global routing table to the node agent component in each edge node, so that the node agent component in each edge node forwards application access traffic based on the global routing table to implement application access.
  • the node agent component in each edge node determines that the local area network where the local node is located and the LAN where the destination node is located are the same LAN based on the global routing table, it will directly forward the application access traffic to the node agent component in the destination node so that the node agent in the destination node The component forwards application access traffic to the destination application.
  • the node agent component in each edge node determines based on the global routing table that the interoperability relationship between the local area network where the local node is located and the LAN where the destination node is located is capable of interoperability through UDP holes, then through the NAT gateway of the LAN where the destination node is located Forward the application access traffic to the node proxy component in the destination node, so that the node proxy component in the destination node forwards the application access traffic to the destination application.
  • the embodiments of the present application also include: maintaining long links with the node proxy components in each edge node; obtaining application access traffic sent by the target node proxy component based on the long links, and forwarding the application access traffic to the node in the destination node Proxy component, so that the node proxy component in the destination node forwards the application access traffic to the destination application; among them, the destination node proxy component determines based on the global routing table that the interoperability relationship between the local area network where the local node is located and the LAN where the destination node is located is unable to pass UDP Node proxy component for hole-punching and interoperability.
  • the embodiment of this application uses the agent on the edge node to proxy the traffic between edge node applications, uses the edge node agent to automatically detect the network reachability between nodes, and generates global routing information.
  • Use the cloud proxy-server to perform cross-network UDP hole punching to achieve penetration of application traffic.
  • UDP hole punching cannot be achieved, use the cloud proxy-server to relay the cross-network traffic of edge applications.
  • the cloud proxy-server monitors the new/deleted information of nodes, updates the node global routing table in a timely manner, and sends it to the edge agent.
  • the edge agent selects the optimal routing path to forward application traffic based on the node global routing table.
  • Application A normally requests application B through k8s serivce.
  • the node where application A is located and the node where application B is located are in the same LAN and can communicate directly: In this case, the request is sent from the pod of application A and forwarded through iptables.
  • the proxy-agent on the node where A is located the proxy-agent determines that the node where the target application B is located can be directly reached based on the global routing table, and directly forwards the traffic to the proxy-agent port on the node where B is located.
  • the agent forwards the traffic to application B, and the response from application B is also returned along this path.
  • Figure 8 is a schematic diagram of application access within the same local area network provided by an embodiment of the present application.
  • Edge application A and edge application B are both in edge LAN 1.
  • Application A normally requests application B through k8s serivce.
  • the node where application A is located and the node where application B is located are not in the same LAN, but the two LANs can communicate with each other through UDP holes: In this case, the request is from A
  • the application pod is sent out and forwarded through iptables to the proxy-agent on the node where A is located.
  • the proxy-agent determines that the node where the target application B is located can be accessed through UDP holes, and forwards the traffic to the LAN where B is located.
  • the public IP and port of the corresponding node exposed on the NAT gateway forward the traffic to the proxy-agent port of node B through the NAT gateway of the LAN where B is located.
  • Application A normally requests application B through k8s serivce.
  • the node where application A is located and the node where application B is located are not in the same LAN.
  • the two LANs cannot communicate with each other through UDP holes: in this case, the request is from
  • the pod of application A is sent out and forwarded to the proxy-agent on the node where A is located through iptables.
  • the proxy-agent determines that the node where the target application B is located cannot be accessed through UDP holes and forwards the traffic to the cloud proxy-server.
  • FIG. 10 is a schematic diagram of a cross-LAN application accessed through a cloud proxy component according to an embodiment of the present application.
  • Edge application A is in edge LAN 1
  • edge application B is in edge LAN 2.
  • Cross-LAN applications can be mutually accessed through the native k8s service, node-based global routing information can be automatically generated, and the most optimal route can be automatically selected through distributed node agents.
  • Optimal routing paths reduce the load on centralized traffic agents.
  • Centralized agents can also be used for traffic forwarding in scenarios where edge nodes are located behind strict firewalls, improving compatibility with edge computing scenarios and having high practical value. .
  • the node discovery result is the node discovery result obtained by the node agent component in the newly registered edge node performing node discovery on each edge node in the current cluster. If the node discovery result is not If any node can be found, perform a UDP hole drilling operation between the LAN where the newly registered edge node is located and the existing LAN to obtain the interoperability relationship between the LANs, and update the global routing table based on the interoperability relationship.
  • the node discovery result is successful node discovery , then associate the node information of the newly registered edge node with the target LAN information to update the global routing table, where the target LAN information is the LAN information of the LAN where the successfully discovered node is located, and the global routing table is distributed to each edge node.
  • the node proxy component allows the node proxy component in each edge node to forward application access traffic based on the global routing table to achieve application access.
  • the node discovery result of the edge node can be obtained when the edge node is registered, and the global routing table is updated based on the node discovery result, and then distributed to the node agent component of each edge node, so that the When an application accesses across nodes, the node proxy component forwards application traffic based on the global routing table, which not only enables application access across edge nodes, but also reduces the load on the centralized traffic proxy.
  • this embodiment of the present application discloses an application access method, which is applied to the cloud proxy component, including:
  • Step S21 When performing node registration on the local node, perform node discovery on each edge node in the current cluster to obtain the node discovery result.
  • Step S22 Send the node discovery result to the cloud proxy component so that the cloud proxy component updates the global routing table based on the node discovery result. If the node discovery result is that no node can be found, the cloud proxy component will locate the newly registered edge node. The UDP hole-punching operation between the LAN and the existing LAN is performed to obtain the interoperability relationship between the LANs, and the global routing table is updated based on the interoperability relationship; if the node discovery result is that the node is successfully discovered, the cloud proxy component will add the node of the newly registered edge node The information is associated with the target LAN information to update the global routing table, and the target LAN information is the LAN information of the LAN where the successfully discovered node is located.
  • Step S23 Obtain the global routing table delivered by the cloud proxy component, and forward application access traffic based on the global routing table to achieve application access.
  • this embodiment of the present application discloses a cloud proxy component, which includes:
  • the node discovery result acquisition module 11 is used to obtain the node discovery result; wherein the node discovery result is the node discovery result obtained by the node agent component in the newly registered edge node performing node discovery on each edge node in the current cluster;
  • the global routing table update module 12 is used to perform a UDP hole drilling operation between the local area network where the newly registered edge node is located and the currently existing local area network to obtain the interoperability relationship between the local area networks if the node discovery result is that no node is found, and based on The interoperability relationship updates the global routing table; if the node discovery result is that the node is successfully discovered, the node information of the newly registered edge node is associated with the target LAN information to update the global routing table, where the target LAN information is the LAN where the successfully discovered node is located.
  • LAN information is used to perform a UDP hole drilling operation between the local area network where the newly registered edge node is located and the currently existing local area network to obtain the interoperability relationship between the local area networks if the node discovery result is that node is found, and based on The interoperability relationship updates the global routing table; if the node discovery result is that the node is successfully discovered, the node information of the newly registered edge node is associated with
  • the global routing table delivery module 13 is used to deliver the global routing table to the node agent component in each edge node, so that the node agent component in each edge node forwards application access traffic based on the global routing table to implement application access.
  • the node discovery result is the node discovery result obtained by the node agent component in the newly registered edge node performing node discovery on each edge node in the current cluster. If the node discovery result is not If any node can be found, perform a UDP hole drilling operation between the LAN where the newly registered edge node is located and the existing LAN to obtain the interoperability relationship between the LANs, and update the global routing table based on the interoperability relationship.
  • the node information of the newly registered edge node is associated with the target LAN information to update the global routing table, where the target LAN information is the LAN information of the LAN where the successfully discovered node is located, and the global routing table is delivered to each edge node.
  • the node proxy component allows the node proxy component in each edge node to forward application access traffic based on the global routing table to achieve application access.
  • the node discovery result of the edge node can be obtained when the edge node is registered, and the global routing table is updated based on the node discovery result, and then distributed to the node agent component of each edge node, so that the When an application accesses across nodes, the node proxy component forwards application traffic based on the global routing table, which not only enables application access across edge nodes, but also reduces the load on the centralized traffic proxy.
  • the cloud proxy component also includes:
  • the cluster node list delivery module is used to deliver the cluster node list to the node agent component in the newly registered edge node, so that the node agent component can perform node discovery on each edge node in the current cluster based on the cluster node list and obtain the node discovery results. .
  • the device also includes:
  • the node information acquisition module is used to obtain the node information of the newly registered edge node sent by the node agent component in the newly registered edge node;
  • the cluster node list maintenance module is used to add node information to the cluster node list.
  • the node information includes the service discovery port of the node agent component; accordingly, the node agent component accesses each edge node in the current cluster based on the service discovery port of the node agent component of each edge node in the cluster node list.
  • the device further includes:
  • the public network NAT gateway address acquisition module is used to obtain the public network NAT gateway address of the LAN where the newly registered edge node is located;
  • the LAN information creation module is used to create the LAN information of the LAN based on the public network NAT gateway address and the node information of the newly registered edge node.
  • the node agent component in each edge node determines based on the global routing table that the local area network where the local node is located and the LAN where the destination node is located are the same LAN, it will directly forward the application access traffic to the node agent in the destination node.
  • the node agent component in each edge node determines based on the global routing table that the interoperability relationship between the local area network where the local node is located and the LAN where the destination node is located is able to be opened through UDP hole interworking, the application access traffic is forwarded to the node agent component in the destination node through the NAT gateway of the LAN where the destination node is located, so that the node agent component in the destination node forwards the application access traffic to the destination application.
  • the device also includes:
  • the long link maintenance module is used to maintain long links with the node agent components in each edge node;
  • the application access traffic acquisition module is used to obtain the application access traffic sent by the target node proxy component based on the long link, and forward the application access traffic to the node proxy component in the destination node, so that the node proxy component in the destination node forwards the application access traffic.
  • the target node agent component is a node agent component that determines based on the global routing table that the interoperability relationship between the local area network where the local node is located and the LAN where the destination node is located cannot communicate through UDP hole drilling.
  • this embodiment of the present application discloses a node agent component, which includes:
  • the node discovery module 21 is used to perform node discovery on each edge node in the current cluster when the local node performs node registration, and obtain the node discovery result;
  • the node discovery result sending module 22 is used to send the node discovery result to the cloud proxy component so that the cloud proxy component updates the global routing table based on the node discovery result. If the node discovery result is that no node can be found, the cloud proxy component performs The UDP hole-punching operation between the LAN where the newly registered edge node is located and the existing LAN is performed to obtain the interoperability relationship between the LANs, and the global routing table is updated based on the interoperability relationship; if the node discovery result is that the node is successfully discovered, the cloud proxy component will new The node information of the registered edge node is associated with the target LAN information to update the global routing table. The target LAN information is the LAN information of the LAN where the successfully discovered node is located;
  • the global routing table acquisition module 23 is used to acquire the global routing table issued by the cloud proxy component, and forward application access traffic based on the global routing table to achieve application access.
  • the local node when the local node performs node registration, it performs node discovery on each edge node in the current cluster, obtains the node discovery result, and then sends the node discovery result to the cloud proxy component, so that the cloud proxy component can perform node discovery based on the node discovery result. Update the global routing table.
  • the cloud proxy component performs a UDP hole drilling operation between the LAN where the newly registered edge node is located and the current existing LAN to obtain the interoperability relationship between the LANs, and Update the global routing table based on the interoperability relationship; if the node discovery result is a successfully discovered node, the cloud proxy component associates the node information of the newly registered edge node with the target LAN information to update the global routing table, and the target LAN information is the location of the successfully discovered node. The local area network information of the local area network is obtained, and finally the global routing table issued by the cloud proxy component is obtained, and application access traffic is forwarded based on the global routing table to achieve application access.
  • edge node discovery can be performed when the edge node is registered.
  • the cloud proxy component updates the global routing table based on the node discovery results, and then sends it to the node proxy component of each edge node to facilitate applications in each edge node.
  • the node proxy component forwards application traffic based on the global routing table, which not only enables application access across edge nodes, but also reduces the load on the centralized traffic proxy.
  • the embodiment of the present application discloses an electronic device 30, which includes a processor 31 and a memory 32; the memory 32 is used to save the computer program; the processor 31 is used to execute the computer program.
  • the memory 32 as a carrier for resource storage, may be a read-only memory, a random access memory, a magnetic disk or an optical disk, etc., and the storage method may be short-term storage or permanent storage.
  • the electronic device 30 also includes a power supply 33, a communication interface 34, an input and output interface 35 and a communication bus 36; the power supply 33 is used to provide operating voltage for each hardware device on the electronic device 30; the communication interface 34 can provide the electronic device 30 with working voltage.
  • the communication protocol it follows is any communication protocol that can be applied to the technical solution of this application, which is not specifically limited here; the input and output interface 35 is used to obtain external input data or To output data to the outside world, the specific interface type can be selected according to specific application needs, and is not specifically limited here.
  • embodiments of the present application also disclose a computer non-volatile readable storage medium for storing a computer program, wherein when the computer program is executed by a processor, the application access method disclosed in the foregoing embodiments is implemented.
  • RAM random access memory
  • ROM read-only memory
  • electrically programmable ROM electrically erasable programmable ROM
  • registers hard disks, removable disks, CD-ROMs, or anywhere in the field of technology. any other known form of storage media.

Landscapes

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

Abstract

本申请公开了一种应用访问方法、云端代理及节点代理组件、设备、介质，包括：获取节点发现结果；若节点发现结果为未能发现任何节点，则进行新注册的边缘节点所在局域网与当前已存在局域网间的UDP打洞操作以得到局域网间的互通关系，并基于互通关系更新全局路由表；若节点发现结果为成功发现节点，则将新注册的边缘节点的节点信息与目标局域网信息关联以更新全局路由表，目标局域网信息为成功发现的节点所在局域网的局域网信息；将全局路由表下发至各边缘节点中的节点代理组件，以便各边缘节点中的节点代理组件基于全局路由表进行应用访问流量转发，实现应用访问。

Description

应用访问方法、云端代理及节点代理组件、设备、介质
相关申请的交叉引用
本申请要求于2022年3月17日提交中国专利局，申请号为202210260957.4，申请名称为“应用访问方法、云端代理及节点代理组件、设备、介质”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及云服务技术领域，特别涉及一种应用访问方法、云端代理及节点代理组件、设备、介质。
背景技术
在传统的k8s集群中，集群节点之间是互联互通的，应用间的互相访问是通过k8s(是一种架构良好的分布式***的例子)的service(服务)对象实现的，service创建后通过label(标签)匹配后端pod(一组并置的容器)，kube-proxy(一种网络代理组件)配置本地iptables(信息包过滤***)规则，将指向service ip的流量转发到后端的pod ip上，当pod跨节点时就依赖于cni(即(container network interface，容器网络接口)构建基于节点网络之上的pod网络；而在基于k8s的边缘计算场景中，一种常见的情况是一个承载边缘计算k8s集群的控制节点部署在一个公共网络上，对外暴露一个公网ip，计算节点分布在不同的现场，不同的局域网络中，同一个局域网内的节点是互联互通的，跨网络的节点无法直接互通，但是所有局域网内的节点都可以跟控制节点暴露的公网ip通信；在这种网络拓扑中，跨局域网的应用pod之间就无法通过传统的k8s service+cni进行相互通信。
发明内容
有鉴于此，本申请的目的在于提供一种应用访问方法、云端代理及节点代理组件、设备、介质，能够实现跨边缘节点的应用访问，并降低集中式流量代理的负载。其具体方案如下：
第一方面，本申请公开了一种应用访问方法，应用于云端代理组件，包括：
获取节点发现结果；其中，节点发现结果为新注册的边缘节点中的节点代理组件对当前集群中的各边缘节点进行节点发现，得到的节点发现结果；
若节点发现结果为未能发现任何节点，则进行新注册的边缘节点所在局域网与当前已存在局域网间的UDP打洞操作以得到局域网间的互通关系，并基于互通关系更新全局路由表；
若节点发现结果为成功发现节点，则将新注册的边缘节点的节点信息与目标局域网信息关联以更新全局路由表，其中，目标局域网信息为成功发现的节点所在局域网的局域网信息；
将全局路由表下发至各边缘节点中的节点代理组件，以便各边缘节点中的节点代理组件基于全局路由表进行应用访问流量转发，实现应用访问。
可选的，获取节点发现结果之前，还包括：
向新注册的边缘节点中的节点代理组件下发集群节点列表，以便节点代理组件基于集群节点列表对当前集群中的各边缘节点进行节点发现，得到节点发现结果。
可选的，还包括：
获取新注册的边缘节点中的节点代理组件发送的新注册的边缘节点的节点信息；
将节点信息添加至集群节点列表。
可选的，节点信息包括节点代理组件的服务发现端口；相应的，节点代理组件基于集群节点列表中各边缘节点的节点代理组件的服务发现端口对当前集群中的各边缘节点进行访问。
可选的，若节点发现结果为未能发现任何节点，还包括：
获取新注册的边缘节点所在局域网的公网NAT网关地址；
基于该公网NAT网关地址以及新注册的边缘节点的节点信息创建该局域网的局域网信息。
可选的，若各边缘节点中的节点代理组件基于全局路由表确定本地节点所在局域网与目的节点所在局域网为同一局域网，则直接将应用访问流量转发至目的节点中的节点代理组件，以便目的节点中的节点代理组件将应用访问流量转发至目的应用；
若各边缘节点中的节点代理组件基于全局路由表确定本地节点所在局域网与目的节点所在局域网的互通关系为能够通过UDP打洞互通，则通过目的节点所在局域网的NAT网关将应用访问流量转发至目的节点中的节点代理组件，以便目的节点中的节点代理组件将应用访问流量转发至目的应用。
可选的，还包括：
维护与各边缘节点中节点代理组件之间的长链接；
基于长链接获取目标节点代理组件发送的应用访问流量，并将应用访问流量转发至目的节点中的节点代理组件，以便目的节点中的节点代理组件将应用访问流量转发至目的应用；
其中，目标节点代理组件为基于全局路由表确定本地节点所在局域网与目的节点所在局域网之间的互通关系为无法通过UDP打洞互通的节点代理组件。
第二方面，本申请公开了一种应用访问方法，应用于节点代理组件，包括：
在本地节点进行节点注册时，对当前集群中的各边缘节点进行节点发现，得到节点发现结果；
将节点发现结果发送至云端代理组件，以便云端代理组件基于节点发现结果更新全局路由表，其中，若节点发现结果为未能发现任何节点，则云端代理组件进行新注册的边缘节点所在局域网与当前已存在局域网间的UDP打洞操作以得到局域网间的互通关系，并基于互通关系更新全局路由表；若节点发现结果为成功发现节点，则云端代理组件将新注册的边缘节点的节点信息与目标局域网信息关联以更新全局路由表，目标局域网信息为成功发现的节点所在局域网的局域网信息；
获取云端代理组件下发的全局路由表，并基于全局路由表进行应用访问流量转发，实现应用访问。
第三方面，本申请公开了一种云端代理组件，包括：
节点发现结果获取模块，用于获取节点发现结果；其中，节点发现结果为新注册的边缘节点中的节点代理组件对当前集群中的各边缘节点进行节点发现，得到的节点发现结果；
全局路由表更新模块，用于若节点发现结果为未能发现任何节点，则进行新注册的边缘节点所在局域网与当前已存在局域网间的UDP打洞操作以得到局域网间的互通关系，并基于互通关系更新全局路由表；若节点发现结果为成功发现节点，则将新注册的边缘节点的节点信息与目标局域网信息关联以更新全局路由表，其中，目标局域网信息为成功发现的节点所在局域网的局域网信息；
全局路由表下发模块，用于将全局路由表下发至各边缘节点中的节点代理组件，以便各边缘节点中的节点代理组件基于全局路由表进行应用访问流量转发，实现应用访问。
第四方面，本申请公开了一种节点代理组件，包括：
节点发现模块，用于在本地节点进行节点注册时，对当前集群中的各边缘节点进行节点发现，得到节点发现结果；
节点发现结果发送模块，用于将节点发现结果发送至云端代理组件，以便云端代理组件 基于节点发现结果更新全局路由表，其中，若节点发现结果为未能发现任何节点，则云端代理组件进行新注册的边缘节点所在局域网与当前已存在局域网间的UDP打洞操作以得到局域网间的互通关系，并基于互通关系更新全局路由表；若节点发现结果为成功发现节点，则云端代理组件将新注册的边缘节点的节点信息与目标局域网信息关联以更新全局路由表，目标局域网信息为成功发现的节点所在局域网的局域网信息；
全局路由表获取模块，用于获取云端代理组件下发的全局路由表，并基于全局路由表进行应用访问流量转发，实现应用访问。
第五方面，本申请公开了一种电子设备，包括处理器和存储器；其中，
存储器，用于保存计算机程序；
处理器，用于执行计算机程序以前述的应用访问方法。
第六方面，本申请公开了一种计算机非易失性可读存储介质，用于保存计算机程序，其中，计算机程序被处理器执行时实现前述的应用访问方法。
附图说明
为了更清楚地说明本申请实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据提供的附图获得其他的附图。
图1为本申请提供的一种边缘计算网络示意图；
图2为本申请提供的一种应用访问方法流程图；
图3为本申请提供的一种组件部署架构示意图；
图4为本申请提供的一种新注册节点的节点发现示意图；
图5为本申请提供的一种NAT示意图；
图6为本申请提供的一种新增边缘局域网的UDP打洞示意图；
图7为本申请提供的一种全局路由表示意图；
图8为本申请提供的一种同一局域网内应用访问示意图；
图9为本申请提供的一种跨局域网应用通过UDP打洞访问示意图；
图10为本申请提供的一种跨局域网应用通过云端代理组件访问示意图；
图11为本申请提供的一种应用访问方法流程图；
图12为本申请提供的一种云端代理组件结构示意图；
图13为本申请提供的一种节点代理组件结构示意图；
图14为本申请提供的一种电子设备结构图。
具体实施方式
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
原生Kubernetes网络中，每个Pod都有一个独立的IP地址，而且假定所有的Pod都可以在一个直接连通的、扁平的网络空间中。而在边缘场景下，打通云中心与边缘数据中心的容器网络平面，开销大，性能低。而边缘数据中心部署的服务与云边间的访问链路是有限而特定的，在需求上，也小于边缘数据中心内部的Pod互访需求。在边缘场景下，容器网络各自收敛于每个数据中心，并使用高可用公网网关+公网IP暴露边缘服务。另一个决定容器网络平面范围的因素是节点网络，容器网络基于节点网络之上，对于100+广泛分布全国的数据中心，打通跨地域的节点网络，无论是网络专线或VPN都是开销巨大的。因此，通常情况下，各数据中心(云+边缘)内部节点网络可靠且联通，数据中心间节点网络不可靠，不直连，数据中心间的互访需求依赖各数据中心的公网网关。例如，参见图1所示，图1为本申请实施例提供的一种边缘计算网络示意图。其中，api-server是k8s集群中用于提供对外api服务的模块。
目前，如何实现跨边缘节点的应用访问，并降低集中式流量代理的负载是目前需要解决的问题。为此，本申请提供了一种应用访问方案，能够实现跨边缘节点的应用访问，并降低集中式流量代理的负载。
参见图2所示，本申请实施例公开了一种应用访问方法，应用于云端代理组件，包括：
步骤S11：获取节点发现结果；其中，节点发现结果为新注册的边缘节点中的节点代理组件对当前集群中的各边缘节点进行节点发现，得到的节点发现结果。
在具体的实施方式中，可以向新注册的边缘节点中的节点代理组件下发集群节点列表，以便节点代理组件基于集群节点列表对当前集群中的各边缘节点进行节点发现，得到节点发现结果；获取新注册的边缘节点中的节点代理组件返回的节点发现结果。
并且，本申请实施例还可以获取新注册的边缘节点中的节点代理组件发送的新注册的边 缘节点的节点信息；将节点信息添加至集群节点列表。其中，节点信息包括节点代理组件的服务发现端口；相应的，节点代理组件基于集群节点列表中各边缘节点的节点代理组件的服务发现端口对当前集群中的各边缘节点进行访问。进一步的，节点信息还包括节点唯一标识。
其中，新注册的边缘节点中的节点代理组件，监听配置的本地服务发现端口，并且，配置本地iptables规则，将K8S service网段的流量均导向本地的节点代理组件，同时连接云端代理组件暴露的公网IP，建立到云端代理组件的长链接通道，生成本节点的唯一标识，将本节点信息上报至云端代理组件。云端代理组件发现有新的节点注册，记录新的节点信息，同时向该节点的节点代理组件下发集群中已有的节点列表，触发该节点的自动发现流程。新注册节点的节点代理组件收到当前的集群节点列表，进入自动发现流程，尝试使用各节点的IP+边缘代理组件的服务发现端口，发起直接访问，此时在同一个局域网内的节点可以返回正常响应，而不在一个局域网内的节点无法直接连通。
需要指出的是，k8s即(Kubernetes)是Google开源的容器集群管理***。在Docker技术的基础上，为容器化的应用提供部署运行、资源调度、服务发现和动态伸缩等一系列完整功能，提高了大规模容器集群管理的便捷性。service是k8s集群中的一种资源类型，通过service类型的资源可以为k8s集群中的pod提供基于域名的访问地址，并可提供简单的四层负载均衡能力。
例如，参见图3所示，图3为本申请实施例提供的一种组件部署架构示意图。以边缘局域网1和边缘局域网2示例，但不限制局域网的数量。本申请主要包括两类组件，分别运行在承载边缘计算业务的k8s集群的云端节点和边端计算节点上，具体如下：proxy-server(即云端代理组件)，以deployment形式运行在k8s集群的云端节点上，与集群api-server通信，通过公网IP对外暴露服务，主要功能包括维护节点信息，根据节点发现结果生成并更新全局路由表，下发全局路由表到proxy-agent(即节点代理组件)，进行局域网间的UDP打洞，进行中继流量转发。proxy-agent，以daemonset的形式运行在边缘计算节点上，每个边缘节点上运行一个实例，主要功能包括节点信息注册、局域网节点发现、本地iptables规则配置、根据全局路由表进行分类转发。边缘节点开机启动，加入集群后，节点上的proxy-agent自动启动，监听配置的本地服务发现端口，配置本地iptables规则，将service网段的流量都导向本地的proxy-agent，同时连接云端proxy-server暴露的公网IP，建立到proxy-server的长链接通道，生成本节点的唯一id标识，将本节点信息上报至云端proxy-server。云端proxy-server发现有新的节点注册，记录新的节点信息，同时向该节点的proxy-agent下发 集群中已有的节点列表，触发该节点的自动发现流程。新节点的proxy-agent收到当前的全局节点列表，进入自动发现流程，尝试使用各节点的IP+proxy-agent的服务发现端口，发起直接访问。在同一个局域网内的节点可以返回正常响应，而不在一个局域网内的节点无法直接连通，即可得到当前新节点和哪些已有的节点在一个网络可达域内的视图，参见图4所示，图4为本申请实施例提供的一种新注册节点的节点发现示意图。以边缘局域网1和边缘局域网2示例，但不限制局域网的数量。
步骤S12：若节点发现结果为未能发现任何节点，则进行新注册的边缘节点所在局域网与当前已存在局域网间的UDP打洞操作以得到局域网间的互通关系，并基于互通关系更新全局路由表。
在具体的实施方式中，进一步的，本申请实施例还可以获取新注册的边缘节点所在局域网的公网NAT(即NetworkAddressTranslation，网络地址转换)网关地址；基于该公网NAT网关地址以及新注册的边缘节点的节点信息创建该局域网的局域网信息。并且，基于该公网NAT网关地址进行新注册的边缘节点所在局域网与当前已存在局域网间的UDP打洞操作以得到局域网间的互通关系。
其中，全局路由表包括集群中各局域网的局域网信息，以及各局域网之间的互通关系。局域网信息包括本局域网中的各节点信息。
需要指出的是，NAT是一个IETF(即InternetEngineeringTaskForce，互联网工程任务组)标准，允许一个整体机构以一个公用IP(即InternetProtocol)地址出现在Internet上，是一种把内部私有网络地址(即IP地址)翻译成合法网络IP地址的技术。NAT解决了IPv4地址不够用的问题，另外NAT屏蔽了私网用户真实地址，提高了私网用户的安全性。典型的NAT组网模型，网络通常是被划分为私网和公网两部分，各自使用独立的地址空间。如图5所示，私网使用私有地址，而公网使用公网地址。为了让私网中的主机A-01。A-02访问互联网上的服务器Server，需要在网络边界部署一台NAT设备用于执行地址转换。NAT设备通常是路由器或防火墙。
步骤S13：若节点发现结果为成功发现节点，则将新注册的边缘节点的节点信息与目标局域网信息关联以更新全局路由表，其中，目标局域网信息为成功发现的节点所在局域网的局域网信息。
需要指出的是，如果新注册节点不和其他任何节点在同一个局域网内，说明此时出现了一个新的局域网及节点，proxy-agent向云端proxy-server发送节点发现结果。端proxy-server收到节点发现结果，如果是已存在的局域网新增节点，则将该节点信息与已存在的局 域网信息关联，更新全局路由表，下发至所有proxy-agent。云端proxy-server收到节点发现结果，如果是全新的局域网新增节点，则新建一个局域网信息，获取该局域网的公网NAT网关地址，并与其他已存在的局域网节点上的proxy-agent通过公网NAT网关尝试进行UDP打洞操作。例如参见图6所示，本申请实施例提供了一种新增边缘局域网的UDP打洞示意图。新增边缘局域网1与边缘局域网2进行UDP打洞。如果新增局域网不可通过UDP打洞访问，本申请实施例可以在云端proxy-server内维护该节点的proxy-agent的长链接通道，后续到该局域网的流量通过云端proxy-server转发到对应节点的长链接通道内。云端proxy-server根据UDP打洞结果，维护新增局域网与其他局域网的互通信息和包含的节点信息，生成全局路由表，下发给所有proxy-agent。参见图7所示，图7为本申请实施例提供的一种全局路由表示意图，以局域网1、局域网2、局域网3为实例，全局路由表包括各局域网的局域网信息，以及各局域网之间的互通关系。局域网信息包括本局域网中的各节点信息以及NAT网关地址。也即，最终生成的全局路由表描述了边缘计算集群的节点与节点之间的互通信息，包括可直接互通、可通过UDP打洞互通、无法通过UDP打洞，同一局域网的节点可直接互通。
步骤S14：将全局路由表下发至各边缘节点中的节点代理组件，以便各边缘节点中的节点代理组件基于全局路由表进行应用访问流量转发，实现应用访问。
若各边缘节点中的节点代理组件基于全局路由表确定本地节点所在局域网与目的节点所在局域网为同一局域网，则直接将应用访问流量转发至目的节点中的节点代理组件，以便目的节点中的节点代理组件将应用访问流量转发至目的应用。
在具体的实施方式中，若各边缘节点中的节点代理组件基于全局路由表确定本地节点所在局域网与目的节点所在局域网的互通关系为能够通过UDP打洞互通，则通过目的节点所在局域网的NAT网关将应用访问流量转发至目的节点中的节点代理组件，以便目的节点中的节点代理组件将应用访问流量转发至目的应用。
并且，本申请实施例还包括：维护与各边缘节点中节点代理组件之间的长链接；基于长链接获取目标节点代理组件发送的应用访问流量，并将应用访问流量转发至目的节点中的节点代理组件，以便目的节点中的节点代理组件将应用访问流量转发至目的应用；其中，目标节点代理组件为基于全局路由表确定本地节点所在局域网与目的节点所在局域网之间的互通关系为无法通过UDP打洞互通的节点代理组件。
也即，本申请实施例使用边缘节点上的agent代理边缘节点应用间互访的流量，使用边缘节点agent自动探测节点间网络可达性，生成全局路由信息。使用云端proxy-server进行跨网络的UDP打洞，实现应用流量的穿透，当无法实现UDP打洞时，使用云端proxy-server中 继代理边缘应用跨网络的流量。云端proxy-server监听节点的新增/删除信息，及时更新节点全局路由表，下发至边端agent，边端agent根据节点全局路由表，选择最优的路由路径转发应用流量。通过本申请实施例提供的方案，边缘计算集群中跨越局域网的应用之间可以实现通过k8s service进行服务发现和互访，体验与普通全连接集群相同，下面，根据前述生成的全局路由表举例对应用访问的三种场景进行详细介绍。
(1)A应用通过k8s serivce正常请求B应用，A应用所在的节点和B应用所在的节点在同一个局域网内，可以直接通信：这种情况下请求从A应用的pod发出，通过iptables转发进入A所在的节点上的proxy-agent，proxy-agent根据全局路由表，判断目标应用B所在的节点可以直达，直接将流量转发到B所在的节点上的proxy-agent端口，由B所在节点的proxy-agent将流量转发给应用B，应用B的响应也沿此条路径返回。参见图8所示，图8为本申请实施例提供的一种同一局域网内应用访问示意图。边端应用A和边端应用B均在边缘局域网1中。
(2)A应用通过k8s serivce正常请求B应用，A应用所在的节点和B应用所在的节点不在同一个局域网内，但两个局域网间可以通过UDP打洞互访：这种情况下请求从A应用的pod发出，通过iptables转发进入A所在的节点上的proxy-agent，proxy-agent根据全局路由表，判断目标应用B所在的节点可以通过UDP打洞访问，将流量转发到B所在的局域网的NAT网关上暴露的对应节点的公网IP和端口，通过B所在局域网的NAT网关将流量转发到B节点的proxy-agent端口，由B所在节点的proxy-agent将流量转发给应用B，应用B的响应也沿此条路径返回。这种情况下AB应用间的流量通过双方NAT网关直接互相转发，无需通过云端proxy-server，可以避免流量通过云端带来的转发性能瓶颈，参见图9所示，图9为本申请实施例提供的一种跨局域网应用通过UDP打洞访问示意图。边端应用A在边缘局域网1中、边端应用B在边缘局域网2。
(3)A应用通过k8s serivce正常请求B应用，A应用所在的节点和B应用所在的节点不在同一个局域网内，两个局域网间也不可以通过UDP打洞互访：这种情况下请求从A应用的pod发出，通过iptables转发进入A所在的节点上的proxy-agent，proxy-agent根据全局路由表，判断目标应用B所在的节点无法通过UDP打洞访问，将流量转发到云端proxy-server，由proxy-server将流量转发到B所在节点的proxy-agent对应的长链接内，由B所在节点的proxy-agent将流量转发给应用B，应用B的响应也沿此条路径返回。这种情况下流量需要通过云端代理服务器中转，可以解决NAT防火墙规则导致的无法实现UDP打洞场景下的跨网络通信问题。参见图10所示，图10为本申请实施例提供的一种跨局域网应用通过云端代理组件访问示意图。边端应用A在边缘局域网1中、边端应用B在边缘局域网2。
这样，解决了边缘计算场景中应用跨局域网相互访问的问题，可以通过原生的k8s service进行跨局域网应用的互相访问，可以自动生成基于节点的全局路由信息，可以通过分布式的节点代理自动选择最优的路由路径，降低集中式流量代理的负载，也可以在边缘节点位于较严格的防火墙之后的场景下使用集中式代理进行流量转发，提高对边缘计算场景的兼容性，具有较高的实用价值。
可见，本申请实施例先获取节点发现结果，节点发现结果为新注册的边缘节点中的节点代理组件对当前集群中的各边缘节点进行节点发现，得到的节点发现结果，若节点发现结果为未能发现任何节点，则进行新注册的边缘节点所在局域网与当前已存在局域网间的UDP打洞操作以得到局域网间的互通关系，并基于互通关系更新全局路由表，若节点发现结果为成功发现节点，则将新注册的边缘节点的节点信息与目标局域网信息关联以更新全局路由表，其中，目标局域网信息为成功发现的节点所在局域网的局域网信息，将全局路由表下发至各边缘节点中的节点代理组件，以便各边缘节点中的节点代理组件基于全局路由表进行应用访问流量转发，实现应用访问。也即，本申请中，可以在边缘节点注册时，获取边缘节点的节点发现结果，并基于节点发现结果更新全局路由表，然后下发至各边缘节点的节点代理组件，以便各边缘节点中的应用进行跨节点访问时，节点代理组件基于全局路由表进行应用流量转发，在实现跨边缘节点的应用访问的同时，降低了集中式流量代理的负载。
参见图11所示，本申请实施例公开了一种应用访问方法，应用于云端代理组件，包括：
步骤S21：在本地节点进行节点注册时，对当前集群中的各边缘节点进行节点发现，得到节点发现结果。
步骤S22：将节点发现结果发送至云端代理组件，以便云端代理组件基于节点发现结果更新全局路由表，其中，若节点发现结果为未能发现任何节点，则云端代理组件进行新注册的边缘节点所在局域网与当前已存在局域网间的UDP打洞操作以得到局域网间的互通关系，并基于互通关系更新全局路由表；若节点发现结果为成功发现节点，则云端代理组件将新注册的边缘节点的节点信息与目标局域网信息关联以更新全局路由表，目标局域网信息为成功发现的节点所在局域网的局域网信息。
步骤S23：获取云端代理组件下发的全局路由表，并基于全局路由表进行应用访问流量转发，实现应用访问。
关于上述步骤S21至步骤S23的具体实施过程可以参考前述实施例公开的内容，在此不再进行赘述。
参见图12所示，本申请实施例公开了一种云端代理组件，包括：
节点发现结果获取模块11，用于获取节点发现结果；其中，节点发现结果为新注册的边缘节点中的节点代理组件对当前集群中的各边缘节点进行节点发现，得到的节点发现结果；
全局路由表更新模块12，用于若节点发现结果为未能发现任何节点，则进行新注册的边缘节点所在局域网与当前已存在局域网间的UDP打洞操作以得到局域网间的互通关系，并基于互通关系更新全局路由表；若节点发现结果为成功发现节点，则将新注册的边缘节点的节点信息与目标局域网信息关联以更新全局路由表，其中，目标局域网信息为成功发现的节点所在局域网的局域网信息；
全局路由表下发模块13，用于将全局路由表下发至各边缘节点中的节点代理组件，以便各边缘节点中的节点代理组件基于全局路由表进行应用访问流量转发，实现应用访问。
可见，本申请实施例先获取节点发现结果，节点发现结果为新注册的边缘节点中的节点代理组件对当前集群中的各边缘节点进行节点发现，得到的节点发现结果，若节点发现结果为未能发现任何节点，则进行新注册的边缘节点所在局域网与当前已存在局域网间的UDP打洞操作以得到局域网间的互通关系，并基于互通关系更新全局路由表，若节点发现结果为成功发现节点，则将新注册的边缘节点的节点信息与目标局域网信息关联以更新全局路由表，其中，目标局域网信息为成功发现的节点所在局域网的局域网信息，将全局路由表下发至各边缘节点中的节点代理组件，以便各边缘节点中的节点代理组件基于全局路由表进行应用访问流量转发，实现应用访问。也即，本申请中，可以在边缘节点注册时，获取边缘节点的节点发现结果，并基于节点发现结果更新全局路由表，然后下发至各边缘节点的节点代理组件，以便各边缘节点中的应用进行跨节点访问时，节点代理组件基于全局路由表进行应用流量转发，在实现跨边缘节点的应用访问的同时，降低了集中式流量代理的负载。
进一步的，云端代理组件，还包括：
集群节点列表下发模块，用于向新注册的边缘节点中的节点代理组件下发集群节点列表，以便节点代理组件基于集群节点列表对当前集群中的各边缘节点进行节点发现，得到节点发现结果。
进一步的，装置还包括：
节点信息获取模块，用于获取新注册的边缘节点中的节点代理组件发送的新注册的边缘节点的节点信息；
集群节点列表维护模块，用于将节点信息添加至集群节点列表。
其中，节点信息包括节点代理组件的服务发现端口；相应的，节点代理组件基于集群节点列表中各边缘节点的节点代理组件的服务发现端口对当前集群中的各边缘节点进行访问。
在具体的实施方式中，若节点发现结果为未能发现任何节点，装置还包括：
公网NAT网关地址获取模块，用于获取新注册的边缘节点所在局域网的公网NAT网关地址；
局域网信息创建模块，用于基于该公网NAT网关地址以及新注册的边缘节点的节点信息创建该局域网的局域网信息。
并且，在具体的实施方式中，若各边缘节点中的节点代理组件基于全局路由表确定本地节点所在局域网与目的节点所在局域网为同一局域网，则直接将应用访问流量转发至目的节点中的节点代理组件，以便目的节点中的节点代理组件将应用访问流量转发至目的应用；若各边缘节点中的节点代理组件基于全局路由表确定本地节点所在局域网与目的节点所在局域网的互通关系为能够通过UDP打洞互通，则通过目的节点所在局域网的NAT网关将应用访问流量转发至目的节点中的节点代理组件，以便目的节点中的节点代理组件将应用访问流量转发至目的应用。
进一步的，装置还包括：
长链接维护模块，用于维护与各边缘节点中节点代理组件之间的长链接；
应用访问流量获取模块，用于基于长链接获取目标节点代理组件发送的应用访问流量，并将应用访问流量转发至目的节点中的节点代理组件，以便目的节点中的节点代理组件将应用访问流量转发至目的应用；
其中，目标节点代理组件为基于全局路由表确定本地节点所在局域网与目的节点所在局域网之间的互通关系为无法通过UDP打洞互通的节点代理组件。
参见图13所示，本申请实施例公开一种节点代理组件，包括：
节点发现模块21，用于在本地节点进行节点注册时，对当前集群中的各边缘节点进行节点发现，得到节点发现结果；
节点发现结果发送模块22，用于将节点发现结果发送至云端代理组件，以便云端代理组件基于节点发现结果更新全局路由表，其中，若节点发现结果为未能发现任何节点，则云端代理组件进行新注册的边缘节点所在局域网与当前已存在局域网间的UDP打洞操作以得到局域网间的互通关系，并基于互通关系更新全局路由表；若节点发现结果为成功发现节点，则云端代理组件将新注册的边缘节点的节点信息与目标局域网信息关联以更新全局路由表，目 标局域网信息为成功发现的节点所在局域网的局域网信息；
全局路由表获取模块23，用于获取云端代理组件下发的全局路由表，并基于全局路由表进行应用访问流量转发，实现应用访问。
可见，本申请实施例在本地节点进行节点注册时，对当前集群中的各边缘节点进行节点发现，得到节点发现结果，然后将节点发现结果发送至云端代理组件，以便云端代理组件基于节点发现结果更新全局路由表，其中，若节点发现结果为未能发现任何节点，则云端代理组件进行新注册的边缘节点所在局域网与当前已存在局域网间的UDP打洞操作以得到局域网间的互通关系，并基于互通关系更新全局路由表；若节点发现结果为成功发现节点，则云端代理组件将新注册的边缘节点的节点信息与目标局域网信息关联以更新全局路由表，目标局域网信息为成功发现的节点所在局域网的局域网信息，最后获取云端代理组件下发的全局路由表，并基于全局路由表进行应用访问流量转发，实现应用访问。也即，本申请中，可以在边缘节点注册时，进行边缘节点发现，云端代理组件基于节点发现结果更新全局路由表，然后下发至各边缘节点的节点代理组件，以便各边缘节点中的应用进行跨节点访问时，节点代理组件基于全局路由表进行应用流量转发，在实现跨边缘节点的应用访问的同时，降低了集中式流量代理的负载。
参见图14所示，本申请实施例公开了一种电子设备30，包括处理器31和存储器32；其中，存储器32，用于保存计算机程序；处理器31，用于执行计算机程序，前述实施例公开的应用访问方法。
关于上述应用访问方法的具体过程可以参考前述实施例中公开的相应内容，在此不再进行赘述。
并且，存储器32作为资源存储的载体，可以是只读存储器、随机存储器、磁盘或者光盘等，存储方式可以是短暂存储或者永久存储。
另外，电子设备30还包括电源33、通信接口34、输入输出接口35和通信总线36；其中，电源33用于为电子设备30上的各硬件设备提供工作电压；通信接口34能够为电子设备30创建与外界设备之间的数据传输通道，其所遵循的通信协议是能够适用于本申请技术方案的任意通信协议，在此不对其进行具体限定；输入输出接口35，用于获取外界输入数据或向外界输出数据，其具体的接口类型可以根据具体应用需要进行选取，在此不进行具体限定。
进一步的，本申请实施例还公开了一种计算机非易失性可读存储介质，用于保存计算机 程序，其中，计算机程序被处理器执行时实现前述实施例公开的应用访问方法。
关于上述应用访问方法的具体过程可以参考前述实施例中公开的相应内容，在此不再进行赘述。
本说明书中各个实施例采用递进的方式描述，每个实施例重点说明的都是与其它实施例的不同之处，各个实施例之间相同或相似部分互相参见即可。对于实施例公开的装置而言，由于其与实施例公开的方法相对应，所以描述的比较简单，相关之处参见方法部分说明即可。
结合本文中所公开的实施例描述的方法或算法的步骤可以直接用硬件、处理器执行的软件模块，或者二者的结合来实施。软件模块可以置于随机存储器(RAM)、内存、只读存储器(ROM)、电可编程ROM、电可擦除可编程ROM、寄存器、硬盘、可移动磁盘、CD-ROM、或技术领域内所公知的任意其它形式的存储介质中。
以上对本申请所提供的应用访问方法、云端代理及节点代理组件、设备、介质进行了详细介绍，本文中应用了具体个例对本申请的原理及实施方式进行了阐述，以上实施例的说明只是用于帮助理解本申请的方法及其核心思想；同时，对于本领域的一般技术人员，依据本申请的思想，在具体实施方式及应用范围上均会有改变之处，综上，本说明书内容不应理解为对本申请的限制。

Claims (20)

  1. 一种应用访问方法，其特征在于，应用于云端代理组件，包括：
    获取节点发现结果；其中，所述节点发现结果为新注册的边缘节点中的节点代理组件对当前集群中的各边缘节点进行节点发现，得到的节点发现结果；
    若所述节点发现结果为未能发现任何节点，则进行所述新注册的边缘节点所在局域网与当前已存在局域网间的UDP打洞操作以得到局域网间的互通关系，并基于所述互通关系更新全局路由表；
    若所述节点发现结果为成功发现节点，则将所述新注册的边缘节点的节点信息与目标局域网信息关联以更新全局路由表，其中，所述目标局域网信息为成功发现的节点所在局域网的局域网信息；
    将所述全局路由表下发至各边缘节点中的节点代理组件，以便各边缘节点中的节点代理组件基于所述全局路由表进行应用访问流量转发，实现应用访问。
  2. 根据权利要求1所述的应用访问方法，其特征在于，所述获取节点发现结果之前，还包括：
    向新注册的边缘节点中的节点代理组件下发集群节点列表，以便所述节点代理组件基于所述集群节点列表对当前集群中的各边缘节点进行节点发现，得到节点发现结果。
  3. 根据权利要求2所述的应用访问方法，其特征在于，还包括：
    获取所述新注册的边缘节点中的节点代理组件发送的所述新注册的边缘节点的节点信息；
    将所述节点信息添加至所述集群节点列表。
  4. 根据权利要求3所述的应用访问方法，其特征在于，所述节点信息包括节点代理组件的服务发现端口；相应的，所述节点代理组件基于所述集群节点列表中各边缘节点的节点代理组件的服务发现端口对当前集群中的各边缘节点进行访问。
  5. 根据权利要求1所述的应用访问方法，其特征在于，若所述节点发现结果为未能发现任何节点，还包括：
    获取所述新注册的边缘节点所在局域网的公网NAT网关地址；
    基于该公网NAT网关地址以及所述新注册的边缘节点的节点信息创建该局域网的局域网信息。
  6. 根据权利要求1所述的应用访问方法，其特征在于，若各边缘节点中的节点代理组件基于所述全局路由表确定本地节点所在局域网与目的节点所在局域网为同一局域网，则直接将所述应用访问流量转发至所述目的节点中的节点代理组件，以便目的节点中的节点代理组件将所述应用访问流量转发至目的应用；
    若各边缘节点中的节点代理组件基于所述全局路由表确定本地节点所在局域网与所述目的节点所在局域网的互通关系为能够通过UDP打洞互通，则通过目的节点所在局域网的NAT网关将所述应用访问流量转发至所述目的节点中的节点代理组件，以便目的节点中的节点代理组件将所述应用访问流量转发至目的应用。
  7. 根据权利要求1至6任一项所述的应用访问方法，其特征在于，还包括：
    维护与各边缘节点中节点代理组件之间的长链接。
  8. 根据权利要求7所述的应用访问方法，其特征在于，还包括：
    基于所述长链接获取目标节点代理组件发送的应用访问流量，并将所述应用访问流量转发至目的节点中的节点代理组件，以便目的节点中的节点代理组件将所述应用访问流量转发至目的应用；
    其中，所述目标节点代理组件为基于所述全局路由表确定本地节点所在局域网与所述目的节点所在局域网之间的互通关系为无法通过UDP打洞互通的节点代理组件。
  9. 根据权利要求1所述的应用访问方法，其特征在于，还包括：
    监听各节点的新增/删除信息，更新节点全局路由表；
    将更新后的节点全局路由表下发至各节点代理组件。
  10. 一种应用访问方法，其特征在于，应用于节点代理组件，包括：
    在本地节点进行节点注册时，对当前集群中的各边缘节点进行节点发现，得到节点发现结果；
    将所述节点发现结果发送至云端代理组件，以便所述云端代理组件基于所述节点发现结果更新全局路由表，其中，若所述节点发现结果为未能发现任何节点，则所述云端代理组件进行新注册的边缘节点所在局域网与当前已存在局域网间的UDP打洞操作以得到局域网间的互通关系，并基于所述互通关系更新全局路由表；若所述节点发现结果为成功发现节点，则所述云端代理组件将新注册的边缘节点的节点信息与目标局域网信息关联以更新全局路由表，所述目标局域网信息为成功发现的节点所在局域网的局域网信息；
    获取所述云端代理组件下发的全局路由表，并基于所述全局路由表进行应用访问流量转发，实现应用访问。
  11. 根据权利要求10所述的应用访问方法，其特征在于，所述获取所述云端代理组件下发的全局路由表，并基于所述全局路由表进行应用访问流量转发，实现应用访问，包括：
    接收本地第一应用向第二应用发送的应用访问流量；
    基于所述全局路由表选择应用转发方式；
    通过所述转发方式向所述第二应用发送所述应用访问流量。
  12. 根据权利要求11所述的应用访问方法，其特征在于，所述通过所述转发方式向所述第二应用发送所述应用访问流量，包括：
    在所述第二应用所在节点和所述本地第一应用处于同一局域网的情况下，将所述应用访问流量直接转发至所述第二应用所在的节点代理组件，以使得所述第二应用的节点代理组件将所述应用访问流量转发至所述第二应用。
  13. 根据权利要求11所述的应用访问方法，其特征在于，所述通过所述转发方式向所述第二应用发送所述应用访问流量，包括：
    在所述第二应用所在节点和所述本地第一应用不处于同一局域网，且两个局域网间通过UDP打洞互访的情况下，通过所述第二应用所在局域网的NAT网关将所述应用访问流量转发至所述第二应用的节点代理组件，以使得所述以使得所述第二应用的节点代理组件将所述应用访问流量转发至所述第二应用。
  14. 根据权利要求11所述的应用访问方法，其特征在于，所述通过所述转发方式向所述第二应用发送所述应用访问流量，包括：
    在所述第二应用所在节点和所述本地第一应用不处于同一局域网，且两个局域网间不可通过UDP打洞互访的情况下，将所述应用访问流量转发至云端代理组件，以使得所述云端代理组件将所述应用访问流量转发至所述第二应用所在节点的节点代理组件，由所述第二应用的节点代理组件将所述应用访问流量发送至所述第二应用。
  15. 根据权利要求10所述的应用访问方法，其特征在于，还包括：
    监听配置的本地服务发现端口，将所述本地服务发现端口产生的流量均导向本地的节点代理组件。
  16. 根据权利要求10所述的应用访问方法，其特征在于，还包括：
    连接云端代理组件暴露的公网IP，建立到云端代理组件的长链接通道；
    将本节点信息上报至云端代理组件。
  17. 一种云端代理组件，其特征在于，包括：
    节点发现结果获取模块，用于获取节点发现结果；其中，所述节点发现结果为新注册的边缘节点中的节点代理组件对当前集群中的各边缘节点进行节点发现，得到的节点发现结果；
    全局路由表更新模块，用于若所述节点发现结果为未能发现任何节点，则进行所述新注册的边缘节点所在局域网与当前已存在局域网间的UDP打洞操作以得到局域网间的互通关系，并基于所述互通关系更新全局路由表；若所述节点发现结果为成功发现节点，则将所述新注册的边缘节点的节点信息与目标局域网信息关联以更新全局路由表，其中，所述目标局域网信息为成功发现的节点所在局域网的局域网信息；
    全局路由表下发模块，用于将所述全局路由表下发至各边缘节点中的节点代理组件，以便各边缘节点中的节点代理组件基于所述全局路由表进行应用访问流量转发，实现应用访问。
  18. 一种节点代理组件，其特征在于，包括：
    节点发现模块，用于在本地节点进行节点注册时，对当前集群中的各边缘节点进行节点发现，得到节点发现结果；
    节点发现结果发送模块，用于将所述节点发现结果发送至云端代理组件，以便所述云端代理组件基于所述节点发现结果更新全局路由表，其中，若所述节点发现结果为未能发现任何节点，则所述云端代理组件进行新注册的边缘节点所在局域网与当前已存在局域网间的UDP打洞操作以得到局域网间的互通关系，并基于所述互通关系更新全局路由表；若所述节点发现结果为成功发现节点，则所述云端代理组件将新注册的边缘节点的节点信息与目标局域网信息关联以更新全局路由表，所述目标局域网信息为成功发现的节点所在局域网的局域网信息；
    全局路由表获取模块，用于获取所述云端代理组件下发的全局路由表，并基于所述全局路由表进行应用访问流量转发，实现应用访问。
  19. 一种电子设备，其特征在于，包括处理器和存储器；其中，
    所述存储器，用于保存计算机程序；
    所述处理器，用于执行所述计算机程序以实现如权利要求1至15任一项所述的应用访问方法。
  20. 一种计算机非易失性可读存储介质，其特征在于，用于保存计算机程序，其中，所述计算机程序被处理器执行时实现如权利要求1至15任一项所述的应用访问方法。
PCT/CN2022/121851 2022-03-17 2022-09-27 应用访问方法、云端代理及节点代理组件、设备、介质 WO2023173720A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210260957.4A CN114363410B (zh) 2022-03-17 2022-03-17 应用访问方法、云端代理及节点代理组件、设备、介质
CN202210260957.4 2022-03-17

Publications (1)

Publication Number Publication Date
WO2023173720A1 true WO2023173720A1 (zh) 2023-09-21

Family

ID=81094558

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/121851 WO2023173720A1 (zh) 2022-03-17 2022-09-27 应用访问方法、云端代理及节点代理组件、设备、介质

Country Status (2)

Country Link
CN (1) CN114363410B (zh)
WO (1) WO2023173720A1 (zh)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114363410B (zh) * 2022-03-17 2022-06-07 苏州浪潮智能科技有限公司 应用访问方法、云端代理及节点代理组件、设备、介质
CN116743738B (zh) * 2023-07-20 2024-04-05 北京道迩科技有限公司 日志传输方法、装置及电子设备

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140029610A1 (en) * 2012-07-30 2014-01-30 Cisco Technology, Inc. Managing grey zones of unreachable nodes in computer networks
CN112751913A (zh) * 2020-12-22 2021-05-04 联奕科技股份有限公司 一种跨Kubernetes集群的网络通信方法及***
CN113055469A (zh) * 2021-03-11 2021-06-29 网宿科技股份有限公司 云容器存储控制方法、***、电子设备及存储介质
CN114363410A (zh) * 2022-03-17 2022-04-15 苏州浪潮智能科技有限公司 应用访问方法、云端代理及节点代理组件、设备、介质

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2013120512A1 (en) * 2012-02-14 2013-08-22 Telefonaktiebolaget L M Ericsson (Publ) Method and apparatus for routing in a multi-area communication network
EP3440824B1 (en) * 2016-04-25 2020-06-03 Huawei Technologies Co., Ltd. Edge server and method of operating an edge server
CN112887452B (zh) * 2021-03-26 2022-11-04 中星科源(北京)信息技术有限公司 局域网间通信方法与***、nat网关
CN113726843B (zh) * 2021-07-13 2024-03-12 阿里巴巴新加坡控股有限公司 边缘云***、数据传输方法、设备及存储介质
CN114064206A (zh) * 2021-11-05 2022-02-18 郑州云海信息技术有限公司 一种访问边缘节点的pod方法、***、设备及存储介质

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140029610A1 (en) * 2012-07-30 2014-01-30 Cisco Technology, Inc. Managing grey zones of unreachable nodes in computer networks
CN112751913A (zh) * 2020-12-22 2021-05-04 联奕科技股份有限公司 一种跨Kubernetes集群的网络通信方法及***
CN113055469A (zh) * 2021-03-11 2021-06-29 网宿科技股份有限公司 云容器存储控制方法、***、电子设备及存储介质
CN114363410A (zh) * 2022-03-17 2022-04-15 苏州浪潮智能科技有限公司 应用访问方法、云端代理及节点代理组件、设备、介质

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
ANONYMOUS: "The extension and expansion of Service Mesh in edge computing scenarios_Edge Computing Community Blog-CSDN Blog", 15 February 2022 (2022-02-15), XP093089794, Retrieved from the Internet <URL:https://blog.csdn.net/weixin_41033724/article/details/122954940> *

Also Published As

Publication number Publication date
CN114363410A (zh) 2022-04-15
CN114363410B (zh) 2022-06-07

Similar Documents

Publication Publication Date Title
CN107181688B (zh) 一种sdn网络中实现服务器端跨域数据传输优化的***和方法
WO2023173720A1 (zh) 应用访问方法、云端代理及节点代理组件、设备、介质
EP2579544B1 (en) Methods and apparatus for a scalable network with efficient link utilization
JP2020162146A (ja) 仮想ネットワークにおける分散型フロー状態ｐ２ｐ設定のためのシステムおよび方法
US8130766B2 (en) System and method for implementing multimedia calls across a private network boundary
US10469444B2 (en) System and method for direct connections between previously unconnected network devices across one or more unknown networks
US7441270B1 (en) Connectivity in the presence of barriers
US9332068B2 (en) Mechanisms for transparently converting client-server software agents to peer-to-peer software agents
WO2015143802A1 (zh) 业务功能链处理方法及装置
CN101043430B (zh) 一种设备之间网络地址转换的方法
US11784917B2 (en) Layer-2 network extension over Layer-3 network using encapsulation
US20060053485A1 (en) Network connection through NAT routers and firewall devices
CN114172865B (zh) 一种云网络下IPv6双栈的实现方法
WO2009146615A1 (zh) 网络地址转换业务的处理方法和***及处理器
WO2015090087A1 (zh) 路由扩散的方法及装置
US20110235588A1 (en) Method, device, and multi-address space mobile network for sending data and forwarding data
Jeong et al. Lisp controller: a centralized lisp management system for isp networks
US11706185B2 (en) Address resolution protocol (ARP) response to ARP requests
US11552883B1 (en) Session establishment using path change
US10148550B1 (en) Methods and apparatus for a scalable network with efficient link utilization
CN117319140A (zh) 一种虚拟网络互联方法、装置、设备及存储介质
Fu et al. A P4-Based Identifier Mapping Network System
CN116232929A (zh) 基于Wireguard的虚拟私有网络环境构建与管理方法
JP2024517714A (ja) セルラーデータパケットのルーティングの為のｉｐネットワークの改善された使用
Makhijani et al. A Scalable and Dynamic Distribution of Tenant Networks across Multiple Provider Domains using Cloudcasting

Legal Events

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

Ref document number: 22931739

Country of ref document: EP

Kind code of ref document: A1