What is Cloud Load Balancer?
Tencent Cloud Load Balancer (CLB) is a service that distributes traffic to multiple real servers so as to elevate service capabilities of applications and eliminate single points of failure for a higher availability.
CLB virtualizes multiple real servers in the same region into a high-performance and high-availability application service pool by setting a virtual IP address (VIP) and then distributes the network requests from clients to the pool in the manner specified by the applications.
CLB checks the health of real servers in the pool and automatically isolates unhealthy ones, thus resolving single points of failure issues and improving the overall service capabilities of the applications.
Tencent CLB, featuring self-service management, self-service fault repair, and defense against network attacks, is applicable for application scenarios such as enterprises, communities, e-commerce, and games.
A working CLB group usually consists of the following components:
Cloud Load Balancer: A CLB instance used for traffic distribution.
**VIP (virtual IP)**: The IP address used for the CLB to provide service to clients.
Backend/Real Server: A set of server instances in the backend used for processing requests.
VPC/Classic Network: Overall network environment.
The access requests from outside the CLB are distributed by CLB instances to the real servers for processing based on policies and forwarding rules.
Cloud Load Balancer
A network load balancing service provided by Tencent Cloud, which works with real servers to provide users with TCP/UDP- and HTTP-based load balancing services.
Cloud Load Balance Listener
CLB listeners allow you to configure listening ports, load balancing policies, and health check settings. Each listening item corresponds to an application service in the backend.
CLB forwards requests to a group of real server instances for processing based on the rules specified by users.
Virtual IP Address
The service address (currently an IP address) assigned by the system. Users can choose whether to disclose the service address to create CLB instances of the public or private network type.
Public network VIP
Regular IP: A general Border Gateway Protocol (BGP) IP address, which balances network quality and costs.
Accelerated IP: An IP address accelerated by using Anycast Internet Acceleration (AIA), which makes public network access more stable, reliable, and low-latency.
Static single-line IP: An IP address used to access the public network through a single ISP, which is low-cost and convenient for autonomous scheduling.
Private network VIP
VPC: An IP address within a VPC.
Classic network: A private IP address on the classic network.
How it Works
Basic working principle
A CLB instance receives traffic from clients and routes requests to real servers in one or more availability zones for processing.
The CLB service works primarily by the load balancing listener. The listener monitors the requests on the CLB instance and distributes them to real servers based on policies. The CLB can forward the requests to real servers directly by configuring the Client - CLB and CLB - Real server forwarding protocols and protocol ports.
We recommend that you configure real server instances of CLB across multiple availability zones. This enables CLB to route traffic to healthy instances in other availability zones when one availability zone becomes unavailable, thereby avoiding service interruptions caused by availability zone failures.
The client requests to access the service through the domain name. Before the request is sent to the load balancer, the DNS server resolves the CLB domain name and returns the requested CLB IPs to the client. When the CLB listener receives a request, it uses different load balancing algorithms to distribute the request to the real servers. Tencent Cloud currently supports multiple balancing algorithms including weighted round robin and
ip_hash weighted minimum connections.
Real server monitoring
A CLB instance can also monitor the running status of real server instances to ensure that traffic is only routed to the normally running instances. The CLB instance will stop routing traffic to an abnormal instance and reroute traffic to it when it is running normally again.
CLB works with the following services to improve application availability and scalability:
Auto Scaling: Controls the instance quantity elastically. If CLB is enabled in auto scaling, the scaled instance is automatically added to the CLB group, and the terminated instance is automatically removed from the CLB group. Domains and DNSPod: Convert your custom domain name (such as
www.example.com) to the IP address used for network communication (such as
192.0.2.1), thus facilitating the routing of requests to CLB instances.