BLB Selection Guide
Updated at：2025-10-16
Overview
This document guides you in choosing suitable BLB types, deployment models, and charge types.
Product overview
Product type
BLB is categorized into the following types:
| Instance type | Description | Differences | Applicable scenarios |
|---|---|---|---|
| General-purpose BLB instance | The standard load balancer includes functional components to fulfill general load balancing needs. It supports distributing both IPv6 and IPv4 traffic from the internet or intranet to multiple IPv4 real servers, offering a lightweight design, easy configuration, and cost efficiency. | Real servers are grouped by instance dimension, with all servers under the same listener sharing a unified listening port. Traffic forwarding options include weighted round-robin, weighted least connections, and source IP-based scheduling algorithms through unified listener configuration. | IPv4 or IPv6 clients can access IPv4 real servers, which supports NAT44 and NAT64 conversions. HTTP/HTTPS listeners do not support complex forwarding policies |
| Application BLB instance | It is tailored for complex application deployment architectures, especially large-scale website setups. It employs a policy-driven network management framework to balance IPv6 and IPv4 traffic from the internet or intranet across multiple IPv4 real servers, ensuring traffic distribution is service-driven. | The backend service supports server groups or IP groups, organizing real servers or backend service IPs by the listener dimension. The listening ports can vary among different real servers or backend service IPs under each listener. HTTP/HTTPS listeners enable policy-based traffic forwarding by domain name, URL, and more. | IPv4 or IPv6 clients can access IPv4 real servers, which supports NAT44 and NAT64 conversions. HTTP/HTTPS listeners support complex forwarding policies. |
| General-purpose IPv6 BLB instance | The standard load balancer includes functional components tailored to general load balancing requirements. It distributes IPv6 traffic from the internet or intranet to multiple IPv6 real servers, providing a lightweight design, simplicity, and cost-effectiveness. | Real servers are grouped by instance dimension, with all servers under the same listener sharing a unified listening port. Traffic forwarding options include weighted round-robin, weighted least connections, and source IP-based scheduling algorithms through unified listener configuration. | IPv6 clients can access IPv6 real servers, which supports NAT66 conversions. HTTP/HTTPS listeners do not support complex forwarding policies. |
| Application IPv6 BLB instance | It is designed for complex application deployment architectures, especially large-scale website structures. Built on a policy-driven network management framework, it balances IPv6 traffic from the internet or intranet across multiple IPv6 real servers, achieving service-oriented load balancing. | The backend service supports server groups or IP groups, organizing real servers or backend service IPs by the listener dimension. The listening ports can vary among different real servers or backend service IPs under each listener. HTTP/HTTPS listeners enable policy-based traffic forwarding by domain name, URL, and more. | IPv6 clients can access IPv6 real servers, which supports NAT66 conversions. HTTP/HTTPS listeners support complex forwarding policies. |
Product deployment mode
| Cluster | Performance type | Performance limitations |
|---|---|---|
| Dedicated cluster | Shared performance type | No speed limits: During traffic surges, its performance might be impacted by other instances within the same dedicated cluster. However, the dedicated cluster ensures consistent overall performance for all BLB instances within the cluster. |
| Dedicated cluster | Specified performance type | Speed limit: During traffic peaks, the instance's performance remains unaffected by other instances within the dedicated cluster. |
| Shared cluster | No speed limit | No speed limit: During traffic peaks, its performance can be impacted by other instances within the shared cluster, and overall user performance is not guaranteed. |
| Shared cluster | Specified performance type | Speed limit: During traffic peaks, the instance's performance remains unaffected by other instances within the shared cluster. |
| Availability zone mode | Applicable scenarios |
|---|---|
| Primary and standby availability zones | Applicable for scenarios requiring high disaster recovery capability. |
| Single availability zone | Only supports network structure scenarios within single availability zones, offering lower disaster recovery capability. |
Product billing mode
The BLB product has the following billing models
| Billing mode | Performance limitations | Applicable scenarios |
|---|---|---|
| Subscription billing | Performance is limited based on the selected specifications, and different specifications have different charges | Long-term. Scenarios with relatively minor service fluctuations |
| Postpay - Billed based on usage duration | Performance is limited based on the selected specifications, and different specifications have different charges | Short-term. Scenarios with relatively minor service fluctuations |
| Postpay - Billed based on usage | Performance is limited based on the selected specifications or unlimited, with unified billing according to actual usage | Scenarios with relatively significant service fluctuations |