This topic describes the configuration limits and default resource quotas for Global Accelerator (GA) and explains how to request a quota increase.
Instance configuration limits
The following diagram shows the configuration options and dependencies for standard and basic instances.
Pay-as-you-go standard GA instances support multiple acceleration IP types, including EIP and Anycast EIP. Subscription standard GA instances support only the EIP acceleration IP type.
For a Subscription standard instance, you must select an appropriate instance specification and associate a basic bandwidth plan.
You can associate Subscription basic instances only with basic bandwidth plans that provide basic or premium bandwidth.
For basic GA instances, the acceleration IP type defaults to EIP and does not support Anycast EIP.
Instance specifications of Large III and higher are not available by default. To use these specifications, contact your account manager.
GA instance limits
Resource | Standard instance default quota | Basic instance default quota | Increase request |
Number of standard instances per Alibaba Cloud account | 10 | No limit | You can increase quotas in Quota Center or on the Quota Management page. For more information, see Increase a quota. |
Supported endpoint types |
Note For more information about endpoint types, see Endpoints. | elastic network interfaces (ENIs), Classic Load Balancer (CLB), Network Load Balancer (NLB), and Elastic Compute Service instances. Note For more information about endpoint types, see Add and manage endpoint groups and endpoints. | Not adjustable |
Number of origin probing tasks per GA instance | 5 | Not applicable | You can increase quotas in Quota Center or on the Quota Management page. For more information, see Increase a quota. |
Support for TCP and UDP fragmented packets | Not supported | Not supported | Not adjustable |
Acceleration area limits
Resource | Standard instance default quota | Basic instance default quota | Increase request |
Supported acceleration areas | China North, China South, China East, China Southwest, Asia Pacific, Europe, North America, South America, and Middle East Note Each acceleration area contains specific acceleration regions. For more information about the mappings between acceleration areas and acceleration regions, see Acceleration areas and regions for standard instances and Acceleration areas and regions for basic instances. | Not adjustable | |
Supported acceleration IP types |
Note For more information about acceleration IP types, see Acceleration IPs. | Elastic IP Address (default) | Not adjustable |
Number of acceleration IPs per GA instance | The quota varies based on the instance billing method and acceleration IP type:
| 20 | You can increase quotas in Quota Center or on the Quota Management page. For more information, see Increase a quota. |
Number of idle acceleration IPs per GA instance | Not applicable | 10 | You can increase quotas in Quota Center or on the Quota Management page. For more information, see Increase a quota. |
Minimum bandwidth per acceleration region |
| Acceleration IP type: EIP: 2 Mbps | Not adjustable |
GA uses a distributed high-availability cluster architecture. The forwarding capacity of a single connection is limited by the single-core performance of an individual data plane node, and your configured peak bandwidth is distributed across multiple servers in the HA cluster. If your workload is dominated by high-throughput, long-lived connections, a single connection may not reach the instance's peak bandwidth. We recommend splitting your traffic into multiple concurrent connections to fully utilize the total bandwidth capacity of your GA instance.
Listener limits
Resource | Standard instance default quota | Basic instance default quota | Increase request |
Number of listeners per GA instance | 50 | Not applicable | Contact your account manager. |
Supported routing types for listeners | intelligent routing or custom routing. Note For more information about routing types, see Listener routing types. | Not adjustable | |
Number of listener ports per listener |
| Not adjustable | |
Supported port range for listeners | 1 to 65,499 | Not adjustable | |
Number of forwarding rules per listener | 20 | You can increase quotas in Quota Center or on the Quota Management page. For more information, see Increase a quota. | |
Number of forwarding conditions per listener | 40 | You can increase quotas in Quota Center or on the Quota Management page. For more information, see Increase a quota. | |
Number of additional certificates per HTTPS listener | 25 | Not adjustable |
Endpoint group limits
Resource | Standard instance default quota | Basic instance default quota | Increase request |
Number of endpoint groups per GA instance | The quota depends on the instance's billing method, as well as the number and protocol of its listeners:
| 1 | You can request a quota increase for endpoint groups that are associated with TCP or UDP listeners only. You can increase quotas in Quota Center or on the Quota Management page. For more information, see Increase a quota. |
Number of virtual endpoint groups per listener | 10 | Not applicable | You can increase quotas in Quota Center or on the Quota Management page. For more information, see Increase a quota. |
Number of endpoints per endpoint group | 4 | 200 | You can increase quotas in Quota Center or on the Quota Management page. For more information, see Increase a quota. |