Cloud Load Balancer
Cloud Load Balancer (CLB) provides a secure and fast layer-4 and layer-7 traffic distribution service to expand service capabilities and eliminate single points of failure, making it easy to handle high-traffic access. Gateway Load Balancer (GWLB) is a load balancer running at the network layer, helping customers to easily deploy, scale, and manage third-party virtual devices with high security.
One single CLB cluster (not one instance) supports up to 120 million concurrent connections and can handle up to 40 Gbps of traffic bandwidth or process 6 million data packets per second. This enables you to easily sustain ecommerce websites, social networking platforms and gaming businesses with over 10 million daily page views.
One CLB cluster consists of 4 physical servers, offering an availability of up to 99.95%. In the extreme case where only one CLB instance is available, it can still support over 30 million concurrent connections. The cluster system will quickly remove faulty instances and keep healthy instances to ensure that the backend server continues to operate properly.
The CLB cluster scales the service capabilities of the application system elastically according to the business load, and automatically creates and releases CVM instances through the dynamic scaling group of Auto Scaling. These features, in conjunction with a dynamic monitoring system and a billing system that is accurate to the second, eliminate the need to manually intervene or estimate resource requirements, helping you efficiently allocate computing resources and prevent resource waste.
With the aid of the BGP Anti-DDoS system, CLB can defend against most network attacks (such as DDoS, CC and web intrusion attacks) and cleanse attack traffic in a matter of seconds, preventing blocked IPs and fully occupied bandwidth. CLB's built-in synproxy anti-attack mechanism prevents backend CVMs from being attacked and overloaded before the BGP system takes effect, ensuring data security and stability.
CLB eliminates the need to invest in additional load balancing hardware and devote time to tedious OPS work, saving you up to 99% of hardware and labor costs. CLB is billed on a pay-as-you-go basis, meaning you only pay for what you use. You can create and terminate CLB instances at any time according to your business needs.
CLB supports requests from layer-4 protocols (including TCP and UDP) and layer-7 protocols (including HTTP and HTTPS). Its centralized certificate management system supports the HTTPS protocol, making data transfer faster, more secure and more reliable.
The weight configuration feature accelerates the request processing speed of backend CVMs. The weight of a backend CVM can be set to an integer between 1 and 100, and is set to 10 by default. CLB distributes the traffic to different backend CVMs based on the set weights or the hash and polling results of the access source IP addresses. Higher weights can be set for CVMs with higher specs (such as CPU and memory performance) so that they can take on more requests.
The application-based CLB layer-7 protocol provides content-based routing, which supports user-defined domain names/URLs and custom forwarding rules and groups and enables the distribution of requests to backend CVMs under different forwarding rules. You can use CLB to divide the target groups for business separation, greatly improving the service efficiency.
The application-based CLB layer-7 protocol supports custom redirection (rewrite), which solves two major problems:
- Forced HTTPS: When a PC or mobile browser accesses a web service with an HTTP request, an HTTPS response is returned to the browser after the request passes through the CLB proxy, forcing the browser to access the webpage using HTTPS.
- Custom redirection: Redirection is necessary when the web service needs to go offline temporarily in situations such as sold-out events for ecommerce websites, or when the site needs to be maintained, updated or upgraded. If no redirection is performed, the old address saved in the user’s bookmarks and search engine database will return a "404" or "503" error page, degrading the user experience and resulting in additional traffic loss.
CLB supports three scheduling algorithms (weighted round robin, IP hash and weighted least connections) and can set the weights of backend CVMs to ensure even traffic scheduling and improve the CLB performance. In addition, CLB supports the scheduling algorithm of least connections in the layer-7 HTTP protocol, which ensures that the load of a backend CVM can be quickly increased when it joins the cluster, reducing the load pressure of other backend CVMs.
For users that use layer-4, CLB provides a simple session persistence function (aka source address-based or IP-based session persistence). When CLB performs load balancing, it uses the source address of the access request as the basis for judging the associated session and forwards all requests from the same IP address (IP address range) to the same CVM. For users that use layer-7, CLB provides a cookie injection-based session persistence method in which CLB is responsible for injecting cookies without requiring backend CVM configurations.
You can customize the frequency at which CLB checks whether the backend CVMs are running normally. When a CVM with an exception is detected, CLB will stop distributing traffic to it and filter out healthy instances to ensure business continuity.
CLB is deployed in multiple availability zones in the same region. For example, two high-performance primary/secondary clusters can be deployed in Guangzhou Zone 2 and Guangzhou Zone 3 respectively, and when the services in Guangzhou Zone 2 are unreachable, the traffic can be automatically switched to Guangzhou Zone 3, making it easy to cope with faults at the data center level. Combined with the DNS resolution capabilities of DNSPod, CLB can support global load balancing to achieve cross-regional disaster recovery.
Business Type
- The business has obvious peaks and valleys, and business stability and low costs are desired.
Scenario Description
- The business goes through significant peaks and troughs and requires the flexible control of backend resources.
The business has requirements that are highly time-sensitive. For example, during ecommerce promotion campaigns, web page views may increase by over 10 times within a few hours, and a large number of backend CVMs need to be added when the number of visitors surges. When the traffic drops, the unnecessary CVMs needs to be terminated. You can use CLB to distribute traffic to the business systems, which helps flexibly enhance the service capabilities of the application systems and elastically adjust the backend resources in real time.
- The business has high peak traffic and needs to build a low-cost architecture.
When a business has high peak traffic, using traditional hardware to build a cluster is very costly and requires tedious OPS work. CLB can work together with Cloud Monitor to effectively manage high peak traffic and reduce deployment costs, relieving you from hardware OPS duties and allowing you to focus instead on developing product features.
Business Type
- The business has complex modules and requires the separation of dynamic and static services.
Scenario Description
- The business requires the separation of dynamic and static services and flexible forwarding.
When the number of requests is high, the targeted distribution of requests to the website can be achieved by differentiating the static and dynamic requests, effectively reducing the backend load pressure. The dynamic requests can be processed by the independently deployed CLB and associated backend CVM clusters, while the static content can be connected to CDN and optimized by COS to significantly improve the loading speed.
Business Type
- The website or business requires data interconnection.
Scenario Description
- The business requires content-based routing and forwarding for IP convergence.
A large website generally has hundreds of business modules. CLB can analyze HTTP headers by setting different forwarding rules or groups to separate the business modules and achieve content-based routing and forwarding. In addition, you can use custom forwarding paths instead of second-level domain names to reduce the number of DNS polls, converge IPs and improve the service accessing speed.
- The Direct Connect-hybrid cloud solution achieves stable data transfer and flexible migration.
With Tencent Cloud's hybrid cloud solution, the core internal systems and data of the business can be stored in the user-built IDC, and the services can be deployed in the cloud to cope with the surge in user traffic. Direct Connect ensures the stability and speed of in/off-cloud data transfer to guarantee data consistency. In addition, the business can be gradually migrated to the cloud through flexible weight configuration, and combined with the data transfer feature of Direct Connect, an elastic, fast, reliable and low-cost business deployment can be easily implemented.
Gateway Load Balancer (GWLB) is a load balancer running at the network layer. The GWLB instance helps customers deploy, scale, and manage third-party virtual devices such as firewalls, intrusion detection and prevention systems, analysis, and visibility, making operations simpler and security stronger.
GWLB, in conjunction with third-party virtual devices such as firewalls, can detect inbound and outbound traffic while ensuring high availability zones. GWLB uses the GENEVE protocol to communicate with firewalls, forwards the received traffic to a group of firewalls at the backend, and performs management features such as health checks and load balancing.
Gateway Load Balancer (GWLB) is a load balancer running at the network layer. The GWLB instance helps customers deploy, scale, and manage third-party virtual devices such as firewalls, intrusion detection and prevention systems, analysis, and visibility, making operations simpler and security stronger.
GWLB associates the Elastic IP with the terminal node. Access from the public network to the VPC network where the GWLB instance resides is available through Private Link. Third-party virtual devices, such as firewalls, are deployed directly in the network data path so that all passing data packets will be checked and processed by the firewalls. This stops malicious data from entering the network and protects internal resources from attacks.
Billing items for Tencent Cloud CLB include instance fees, LCU fees, network fees, and cross-domain binding fees. Different scenarios correspond to different fees. For billing and pricing details, please refer to the Billing Overview .
Billing items for Tencent Cloud GWLB include instance fees and GLCU fees. For billing and pricing details, please refer to the Billing Overview.