This topic describes product and documentation updates for Global Accelerator (GA).
March 2026
|
Feature |
Change type |
Description |
References |
|
Service-based WAF 3.0 integration for pay-as-you-go standard GA instances |
Added |
Pay-as-you-go standard GA instances support one-click integration with WAF 3.0. Once enabled, the service-based web application firewall (WAF) is integrated into the acceleration region in an out-of-path mode. It inspects and protects traffic without being part of the traffic forwarding path. You can enable or disable WAF protection without modifying your network architecture. |
January 2026
|
Feature name |
Change type |
Description |
Related documents |
|
Custom TLS security policies for standard pay-as-you-go Global Accelerator (GA) |
Added |
You can now configure custom TLS security policies for standard pay-as-you-go Global Accelerator (GA) instances, which lets you specify TLS protocol versions and disable specific cipher suites in an HTTPS listener to meet your security and compliance requirements. |
October 2025
|
Feature name |
Change type |
Description |
Documentation |
|
Traffic mirroring for layer-7 HTTP(S) listeners on pay-as-you-go standard Global Accelerator instances |
New |
Forwarding rules for layer-7 HTTP(S) listeners on pay-as-you-go standard Global Accelerator instances now support traffic mirroring. This feature lets you copy business traffic to different backend services. GA automatically ignores responses from the mirrored backend, so it does not affect the original request. |
|
|
Custom HTTP headers in access logs for pay-as-you-go standard Global Accelerator instances |
New |
Access logs for layer-7 HTTP(S) listeners on pay-as-you-go standard Global Accelerator instances now support custom HTTP headers. You can specify the HTTP headers to record, and GA writes the matched content to the |
July 2025
|
Feature |
Update type |
Description |
References |
|
Pay-as-you-go standard GA instances support multiple endpoint groups for Layer 7 (HTTP/HTTPS) listeners |
Update |
Pay-as-you-go standard GA instances now allow you to configure multiple default endpoint groups in different regions for a single HTTP or HTTPS listener. You can set traffic distribution ratios for the endpoint groups to control the traffic forwarded to each group. Additionally, pay-as-you-go standard GA instances now offer these capabilities:
|
|
|
Pay-as-you-go standard GA now integrates with Anti-DDoS |
New |
Pay-as-you-go standard GA instances now integrate with Anti-DDoS. Using Alibaba Cloud's global scrubbing centers, this feature provides terabit-level DDoS protection for the secure CNAME (secure accelerated IP address) of a GA instance. |
June 2025
|
Feature |
Update type |
Description |
References |
|
New region available |
Update |
Global Accelerator now supports Mexico as an acceleration region for pay-as-you-go standard Global Accelerator instances. |
|
|
Pay-as-you-go standard Global Accelerator instances now support Anycast Elastic IP Address |
New |
Pay-as-you-go standard Global Accelerator instances now support Anycast Elastic IP Address as a type of accelerated IP address. If you use an Anycast EIP, Global Accelerator (GA) provides a unified IP entry point across multiple regions worldwide. You do not need to select an acceleration area or region. Users in different regions can automatically access the Alibaba Cloud acceleration network from a nearby location via the Anycast EIP or a CNAME record. |
May 2025
|
Feature |
Update type |
Description |
References |
|
Backend IP address protocol configuration for pay-as-you-go standard GA instances |
New |
Pay-as-you-go standard GA instances now support setting the backend IP address protocol to IPv4-only, IPv6-only, or IPv4/IPv6 protocol affinity. This lets you select the IP protocol for communication between GA and backend services based on your deployment, ensuring end-to-end accelerated access across various network environments. |
February 2025
|
Feature |
Change type |
Description |
References |
|
Port mapping for TCP and UDP listeners |
New |
Standard GA instances now support port mapping for TCP and UDP listeners. Port mapping lets you customize the mapping between listener ports and backend application ports, improving the flexibility and security of your application deployments. |
|
|
Preserving client source IP for UDP listeners |
New |
For standard GA instances, UDP listeners now support configuring Preserve Client IP to get the real IP addresses of clients. |
|
|
Multiple port ranges for TCP and UDP listeners |
Update |
For pay-as-you-go standard GA instances, you can now configure multiple consecutive port ranges for a single TCP or UDP listener. The supported port range is 1 to 65499. |
|
|
New backend service type for standard GA instances |
New |
Endpoints on standard GA instances now support the Custom Private IP backend service type. Use endpoints of the Custom Private IP type to accelerate unsupported Alibaba Cloud backend services or on-premises services over a private connection. |
December 2024
|
Feature |
Update type |
Description |
References |
|
Listener-granularity monitoring |
New |
Pay-as-you-go standard Global Accelerator instances now support listener-granularity monitoring. |
July 2024
|
Feature |
Update type |
Description |
References |
|
Connect to backend services over HTTP/2 |
New |
Standard Global Accelerator now supports connecting to backend services over HTTP/2. This allows backend services to take full advantage of the protocol to significantly improve performance and reduce latency and network overhead. |
May 2024
|
Feature |
Update type |
Description |
References |
|
instance monitoring |
Update |
The Used CUs and GaIpProcessedBytes metrics are now available for pay-as-you-go standard GA instances. |
March 2024
|
Feature |
Update type |
Description |
References |
|
HTTP/3 support |
New |
Standard GA supports HTTP/3 to enhance the stability, security, and efficiency of data transmission. |
February 2024
|
Feature |
Update type |
Description |
References |
|
CU calculator |
New |
Use the CU calculator to estimate the instance fee and CU fee for a pay-as-you-go standard GA instance. |
|
|
CNAME |
Update |
Check the DNS record of a standard GA instance to view the effective regions for its CNAME. |
December 2023
|
Feature |
Update type |
Description |
References |
|
Connection timeout |
Update |
Smart routing listeners now support the Idle Connection Timeout Period and Connection Request Timeout. |
October 2023
|
Feature |
Update type |
Description |
References |
|
Forwarding rules |
Update |
Forwarding rules that use domain name and path conditions now support regular expressions. |
|
|
Accelerated IP protocol version |
Update |
Pay-as-you-go standard GA instances support dual-stack accelerated IP addresses. You can allocate both IPv4 and IPv6 accelerated IP addresses to an acceleration area. This enables IPv4 and IPv6 clients to access IPv4 services. |
|
|
New backend service types for standard GA instances |
Update |
Endpoints for standard GA instances now support Network Load Balancer (NLB) and elastic network interfaces (ENIs) as backend service types. |
August 2023
|
Feature |
Change type |
Description |
Related documentation |
|
GA resource plan |
New |
A GA resource plan is a prepaid package for pay-as-you-go GA instances. After a resource plan is purchased and becomes active, it automatically offsets CU fees and instance fees at a specific ratio. Resource plans are more cost-effective than pay-as-you-go. |
|
|
Basic GA instances support multiple accelerated IP addresses and endpoints |
Update |
You can now add multiple accelerated IP addresses and endpoints in the same region to Basic GA instances. You can bind accelerated IP addresses to endpoint groups to map them to endpoints and distribute traffic to backend services. |
|
|
New backend service type for Basic GA instances |
Update |
Network Load Balancer (NLB) is now a supported backend service for endpoints in Basic GA instances. |
July 2023
|
Feature |
Update type |
Description |
References |
|
Basic DDoS protection |
New |
Basic DDoS protection is enabled by default for accelerated IP addresses of Global Accelerator instances and public IP addresses of endpoints. It provides up to 5 Gbps of DDoS mitigation. The maximum amount of free mitigation traffic varies by region. |
|
|
Bandwidth billing method |
New |
Global Accelerator supports a pay-by-traffic bandwidth billing method. Traffic fees for accessing the Global Accelerator network are settled and billed by CDT. You can select the public network quality type for your Global Accelerator instances, including BGP (Multi-ISP) and BGP (Multi-ISP) Pro. Standard Global Accelerator instances support two transmission network quality types: Premium Bandwidth for Cross-domain Acceleration and Leased Line for Cross-domain Acceleration. |
|
|
Instance billing method |
New |
Global Accelerator supports the pay-as-you-go billing method for instances. |
March 2023
|
Feature |
Update type |
Description |
References |
|
domain name management |
New |
GA provides an ICP filing exemption for accelerated domain names. If your accelerated domain name already has a valid ICP filing, you can associate it with a GA instance. GA then exempts the domain name and its subdomains from the Alibaba Cloud ICP filing requirement. |
|
|
Self-service unsubscription for basic bandwidth plans |
Update |
You can unsubscribe from subscription-based basic bandwidth plans in the GA console or the Expenses and Costs console. |
|
|
Instance purchase page optimization |
Update |
GA has optimized the instance purchase page. You can follow the on-screen instructions to purchase a suitable Global Accelerator configuration. |
|
|
Business configuration wizard |
New |
GA provides configuration prompts and a business configuration wizard page to simplify deployment. |
February 2023
|
Feature |
Update type |
Description |
References |
|
basic bandwidth plan |
Update |
You can now downgrade basic bandwidth plans. |
|
|
Downgrade support for standard GA instances |
Update |
You can downgrade the instance type for standard GA instances. |
|
|
Self-service unsubscription for subscription GA instances |
Update |
You can request to unsubscribe from subscription GA instances in the GA console or the Expenses and Costs console. |
January 2023
|
Feature |
Update type |
Description |
References |
|
Configuration-free 7-day free trial |
Discontinued |
The configuration-free 7-day free trial for Alibaba Cloud GA has been discontinued. To try the GA service, consider the limited-time first-month free trial. For more information, see [Trial] Limited-time first-month free trial. |
None |
December 2022
|
Feature |
Update type |
Description |
References |
|
Tags |
New |
GA now supports tags. You can add tags to standard GA instances, basic GA instances, bandwidth plans, and access control lists to categorize resources for simplified searching and aggregation. |
|
|
Resource groups |
New |
Resource groups help you manage resources and user permissions for multiple projects or applications within a single Alibaba Cloud account. GA supports resource groups for the following resources: standard GA instances, basic GA instances, bandwidth plans, and access control lists. |
|
|
New backend service type for basic GA instances |
Update |
You can now add Elastic Compute Service (ECS) instances as endpoints for basic GA instances. |
November 2022
|
Feature |
Update type |
Description |
References |
|
Instance diagnosis |
New |
The instance diagnosis feature checks the configuration and running status of a standard GA instance and provides quick fixes for detected issues. |
|
|
Certificate management |
Update |
You can replace additional certificates. This is useful for replacing an expired certificate without changing the associated domain name. |
October 2022
|
Feature |
Update type |
Description |
References |
|
Origin probing |
New |
Use origin probing to monitor end-to-end network quality from detection points to the origin server in real time, allowing you to quickly locate network faults and apply targeted network optimizations. |
September 2022
|
Feature |
Update type |
Description |
References |
|
custom routing listener |
New |
After you configure a custom routing listener for a GA instance, GA generates a port mapping table from the listener port range, endpoint group port ranges, and endpoint IP addresses (vSwitches). The listener forwards requests to specific vSwitch IP addresses and ports based on this mapping table. |
June 2022
|
Feature |
Update type |
Description |
References |
|
Configuration-free 7-day free trial |
New |
Alibaba Cloud Global Accelerator offers a configuration-free 7-day free trial. Just enter your service details, and the system automatically configures a Global Accelerator instance. |
None |
|
Accelerated IP address type |
New |
Standard Global Accelerator instances support the following two types of accelerated IP addresses:
|
|
|
Basic GA instance |
New |
Global Accelerator now offers basic instances for Layer 3 (IP) network acceleration. |
March 2022
|
Feature |
Update type |
Description |
References |
|
Endpoint groups |
Update |
GA now lets you configure multiple endpoint groups in different regions for a single TCP or UDP listener. You can set traffic dialing values for these groups to control the proportion of traffic routed to each group. |
February 2022
|
Feature |
Update type |
Description |
References |
|
Additional HTTP header fields |
New |
You can add additional HTTP header fields when you configure an HTTP or HTTPS listener for a Global Accelerator (GA) instance. |
|
|
TLS security policy |
New |
You can select a TLS security policy when you configure an HTTPS listener for a Global Accelerator (GA) instance. |
|
|
New backend service types for standard GA instances |
New |
Global Accelerator (GA) now supports Application Load Balancer (ALB) instances and Object Storage Service (OSS) buckets as backend service types. |
January 2022
|
Feature |
Change type |
Description |
Related documentation |
|
Health check |
New |
You can enable health checks for the endpoint groups of your GA instance. When an endpoint fails a health check, GA automatically redirects new requests to other healthy endpoints. After an unhealthy endpoint recovers and passes health checks, GA automatically resumes routing traffic to it. |
|
|
Large-scale port listener |
Update |
You can now configure a single TCP or UDP listener to use a range of more than 300 consecutive ports. Such a listener is called a large-scale port listener. |
September 2021
|
Feature |
Update type |
Description |
References |
|
Preserve client IP |
Update |
Added support for Proxy Protocol. This method is ideal when IPv6 clients access backend services. Your backend server must support parsing Proxy Protocol to obtain the source IP address. |
August 2021
|
Feature |
Update type |
Description |
References |
|
Access control |
New |
You can configure different access control policies for different listeners. |
|
|
Access log |
New |
The access log records traffic for all requests to endpoints, which helps you check access control rules and troubleshoot network issues. |