Distributed load balancing system, health check method, and service node

• US 10,187,459 B2
• Filed: 10/13/2017
• Issued: 01/22/2019
• Est. Priority Date: 12/14/2016
• Status: Active Grant

• Current Assignee: Huawei Cloud Computing Technologies Company Limited (Huawei Investment & Holding Co., Ltd.)
• Original Assignee: Huawei Technologies Co., Ltd. (Huawei Investment & Holding Co., Ltd.)
• Inventors: Luo, Gaowei
• Primary Examiner(s): Abedin, Normin
• Application Number:
  US15/783,519
• Publication Number:
  US 20180167453A1
• Time in Patent Office: 466 Days
• CPC Class Codes: H04L 12/4641 Virtual LANs, VLANs, e.g. virtual private networks [VPN] LAN interconnection over a bridge based backbone H04L12/462; encapsulation techniques H04L12/4633; routing of packets H04L45/00; packet switches H04L49/00; virtual private networks for security H04L63/0272
• CPC Class Codes: H04L 41/5058 Service discovery by the service manager
• CPC Class Codes: H04L 41/5096 wherein the managed service relates to distributed or central networked applications
• CPC Class Codes: H04L 43/0817 by checking functioning
• CPC Class Codes: H04L 43/10 Active monitoring, e.g. heartbeat, ping or trace-route
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers
• CPC Class Codes: H04L 67/1008 based on parameters of servers, e.g. available memory or workload monitoring of computer activity G06F11/30
• CPC Class Codes: H04L 67/1029 using data related to the state of servers by a load balancer

Abstract

A distributed load balancing system includes a cloud management platform, an access node, and at least one service node, where the access node and the at least one service node are physical servers, a first virtual load balancer runs on the access node, and at least one backend server runs on the service node. A detection message is sent by using a detection module on a target service node to detect a status of a target backend server, and the status of the target backend server is shared to a first virtual load balancer by using the detection module. Therefore, not only a quantity of detection messages in the load balancing system is reduced, but also signaling load on the access node is reduced.

Claims

1. A distributed load balancing system, comprising a cloud management platform, an access node, and at least one service node, wherein the access node, a the first node, and the at least one target service nodes are physical servers, a first virtual load balancer runs on the access node, and at least one backend server runs on each of the one or more target service nodes, wherein a first node programmed to:

determine, by the cloud management platform, according to an identifier of a service, multiple target backend servers providing the service, and a network address of each target backend server of the multiple target backend servers;

determine, by the cloud management platform, an identifier of one or more target service nodes on which the multiple target backend servers are located; and

send, by the cloud management platform, a check request message to each target service node, wherein each check request message carries a network address of a target backend server running on a target service node that can receive the check request message, and the identifier of the service;

wherein each target service node is programmed to;

receive the check request message sent to the target service node, and send the network address of the target backend server running on the target service node carried in the check request message to a detection module corresponding to the identifier of the service carried in the check request message;

send, by the detection module, according to the received network address, a detection message to the target backend server corresponding to the received network address to detect a status of the target backend server; and

share, by the detection module, the status of the target backend server to the first virtual load balancer on the access node; and

wherein access node is programmed to;

obtain, by the first virtual load balancer, a status of the target backend server on each target service node, and update, according to the status of the target backend server on each target service node, a recorded status of the backend server providing the service.

2. The system according to claim 1, wherein the detection module on each target service node is deployed in a first namespace on the target service node, and the first virtual load balancer is deployed in a second namespace on the access node, wherein the first namespace and the second namespace have a same virtual Internet Protocol (IP) address.

3. The system according to claim 2, wherein the detection module on each target service node is further programmed to construct a notification packet, and send the notification packet to the first virtual load balancer, wherein a source IP address of the notification packet is the virtual IP address, a destination IP address of the notification packet is a subnet broadcast address corresponding to the virtual IP address, and the notification packet carries the status of the target backend server running on the target service node on which the detection module is located.

4. The system according to claim 3, wherein namespaces corresponding to different services have different virtual local area network (VLAN) identifiers, wherein a VLAN identifier of the first namespace and a VLAN identifier of the second namespace are the same.

5. The system according to claim 3, wherein the detection module on each target service node is further programmed to send the notification packet to the first virtual load balancer through a tunnel between the access node and the target service node on which the detection module is located.

6. The system according to claim 1, wherein when a client accessing the service runs on a target service node, the target service node further comprises a second virtual load balancer;

and the target service node is further programmed to send a by the detection module, the status of the target backend server on the target service node to the second virtual load balancer.

7. The system according to claim 1, wherein the cloud management platform records a correspondence among the identifier of the service, a network address of each backend server running the service, and an identifier of a service node on which the backend server is located.

8. The system according to claim 1, wherein the access node is further programmed to receive a request that is for accessing the service and that is sent by the client, and select, for the client according to the updated status of the backend server, a backend server providing the service.

9. A health check method in a distributed load balancing system, comprising:

determining, by a first node according to an identifier of a service, multiple target backend servers providing the service, and a network address of each target backend server of the multiple target backend servers;

determining, by the first node, an identifier of each target service node on which the multiple target backend servers are located; and

sending, by the first node, a check request message to each target service node, wherein each check request message carries a network address of a target backend server running on a target service node that receives the check request message, and the identifier of the service;

receiving, by each target service node, the check request message, and sending the network address of the target backend server running on the target service node to a detection module corresponding to the identifier of the service carried in the check request message;

sending, by each target service node according to the received network address, a detection message to the target backend server corresponding to the received network address to detect a status of the target backend server; and

sharing, each target service node, the status of the target backend server to a first virtual load balancer on an access node; and

obtaining, by the access node, a status of the target backend server on each target service node, and updating, according to the status of the target backend server on each target service node, a recorded status of the backend server providing the service.

10. The method according to claim 9, wherein the detection module on each target service node is deployed in a first namespace on the target service node, and the first virtual load balancer is deployed in a second namespace on the access node, wherein the first namespace and the second namespace have a same virtual Internet Protocol IP address.

11. The method according to claim 10, comprising:

constructing, by each target service node, a notification packet, and sending the notification packet to the first virtual load balancer, wherein a source IP address of the notification packet is the virtual IP address, a destination IP address of the notification packet is a subnet broadcast address corresponding to the virtual IP address, and the notification packet carries the status of the target backend server running on the target service node on which the detection module is located.

12. The method according to claim 11, wherein namespaces corresponding to different services have different virtual local area network VLAN identifiers, wherein a VLAN identifier of the first namespace and a VLAN identifier of the second namespace are the same.

13. The method according to claim 11, wherein the sending, by each target service node, the notification packet to the first virtual load balancer comprises:

sending, by each target service node, the notification packet to the first virtual load balancer through a tunnel between the access node and the target service node on which the detection module is located.

14. The method according to claim 9, wherein when a client accessing the service runs on a target service node, the target service node further comprises a second virtual load balancer, and the method further comprises:

sending, by on the target service node, the status of the target backend server on the target service node to the second virtual load balancer.