1. Non-transitive Connectivity of Peering Connection
A peering connection enables connectivity between two VPCs, but the connectivity cannot be transited.For example, as shown in the following figure, peering connections are created between VPC 1 and VPC 2, and between VPC 1 and VPC 3. However, due to the non-transitivity of peering connections, traffic cannot flow between VPC 2 and VPC 3.
Note:
Even if a peering connection is created, both ends of the connection cannot communicate with each other if routes for sending and returning packets are not configured at both ends.
2. Intra-Region and Cross-Region Peering Connection
VPC supports both intra-region and cross-region peering connections. Due to the differences in physical distance and underlying implementation architecture, there are some differences between the two types of connections, as shown in the table below.
| Comparison Items | Intra-Region Peering Connection | Cross-Region Peering Connection |
|---|---|---|
| Underlying architecture | Local private network within a single region | Cross-region internal MPLS network |
| Cross-account connection | Supported | Supported |
| Access permission | The CVMs at either end of a peering connection can access all the resources at the peer end, including CVMs, databases, and CLBs. | The CVMs at either end of a peering connection can access all the resources at the peer end, including CVMs, databases, and CLBs. |
| Feature Limits | The VPC IP ranges at both ends of the peering connection cannot overlap. Peering connections are independent of each other. | The VPC IP ranges at both ends of the peering connection cannot overlap. If multiple peer VPCs are connected to the same VPC, the IP ranges to which these peer VPCs belong cannot overlap. |
| Use Cases | Mainly used to connect applications located in different Virtual Private Clouds within the same region. | Typical use cases: cross-region disaster recovery. By utilizing cross-region peering connections, VPCs in different regions can communicate with each other, which helps rapidly deploy a 2-region-3-center solution for disaster recovery and meets the requirement for financial-level network disaster recovery with high bandwidth and reliability. |