Support providing Load Balancer Service for Layer 7 business and Layer 4 business with higher security requirements
Support providing stateful Layer 4 Load Balancer Service for business.
Support providing stateless Layer 4 Load Balancer Service for business
Load balancer provides Load Balancer Service that is based on Layer 4 (TCP/TLS protocol) and Layer 7 (HTTP/HTTPS), and can independently set listening protocols and ports to meet different business needs.
The load balancer will periodically detect running status of backend server, customize detection frequency; once any backend server operation exception is detected, the traffic will not be assigned to the servers with exception. In this way, the service availability is guaranteed.
Load Balancer supports round robin, least connection number, source IP and other calling algorithms. Appropriate algorithm can be selected based on business demand of one’s own. For example, you can select round robin when there isn’t big difference among the backend servers, so as to evenly assign traffic of backend servers and improve external service capacity. At the same time, weight can be configured based on backend server performance. The server of high configuration can be set with high weight, to host more access traffic.
The user may independently select whether to associate EIP, flexibly set up intranet and public network Load Balancer, free switch, to meet business demands at different scenarios. By associating public IP, the Intranet structure can be hid and system security can be improved.
The Application Load Balancer supports forwarding the requests from the same resource to the same backend server for processing through Source IP forward and Layer 7 cookie session persistence to meet the business needs in different scenarios.
Load Balancer instance resource automatic redundancy configuration, cluster method is adopted to provide services; and resources can be deployed in the multiple availability zones so as to ensure high availability of services.
The distributed network Load Balancer adopts the All AZs distributed deployment architecture. There isn’t any Load Balancer entity. There is no case where the Load Balancer resource performance reaches the cap and the forwarding is restricted. The forwarding performance has no bottleneck.
Load Balancer is capable of automatically detecting the availability status of backend servers through the strategy of physical examination policy. Once finding abnormal running, it will not forward the traffic to this server. It will restore automatically after normal operation, or the backend server can be manually added or deleted according to the practical business needs. When it is necessary for higher availability, the resources can be deployed in different availability zones.
LB can be configured in an intranet environment, and provide external services by associating EIP. Therefore, the private IP structure can be concealed, so as to enhance the system security. Besides, a more secure protection system can be built by setting a firewall or taking other measures.
Support unvarnished transmission of the real IP address of the client through LB under HTTP listening protocol, making it convenient to analyze the source of customer and make statistics on the operation and maintenance data.
Load Balancer can be used to evenly assign the traffic to multiple backend servers, and set up different traffic distribution schemes through the scheduling algorithm, giving a solution to the system impact caused by mass-traffic and burst traffic. It can be applied in conventional websites, large-scale business websites, seckilling, panic buying or other business systems that may lead to the increase of burst traffic.