# Cross-Region Network ## What Is Cross-Region Network? Cross-Region Network is the global backbone that carries private traffic between regions **inside a single VPC**. The simplest way to see it is through the **Global VPC** — a single VPC named `VPC-Default` that every account gets, and that already spans every region on the platform. Inside the Global VPC, the platform has done the setup for you: * **One subnet per region is auto-provisioned.** Each region has a default subnet in `VPC-Default` with a non-overlapping `/20` range. You don't pick CIDRs, you don't allocate ranges, and you don't worry about collisions. Creating an instance in a region lands it on that region's default subnet. * **All regional subnets in the same VPC are already connected.** Instances in Hong Kong and instances in Frankfurt sit in the same VPC, see each other over private IPs, and can `ping` each other out of the box. * **A free default bandwidth is provisioned at 10 Kbps.** Every VPC ships with a 10 Kbps free baseline between regions — enough for connectivity checks (`ping`, `traceroute`) but not for production traffic. Any production workload needs the bandwidth raised for the region pair it uses. Everything else — encapsulation, path selection across the backbone, failover — is handled by the platform. You don't run a tunnel, you don't manage BGP peerings with the backbone, and you don't install routes for the mesh. ![Cross-Region Network Architecture](/files/Wpy6fYYIh4rrgr9LyOwa) *** ## The VPC Model The platform only offers one kind of VPC: a **Global VPC**. There is no separate "regional VPC" object. * **`VPC-Default`** — the Global VPC you start with. It has one auto-provisioned default subnet in every region. * **Custom Global VPCs** — you can create additional Global VPCs with your own CIDRs. They are still global in scope: you add subnets to whichever regions you want them in. A custom Global VPC with subnets in only one region effectively behaves as a regional VPC. Cross-region reachability exists between subnets **of the same VPC**. Two different VPCs are not connected by Cross-Region Network — regardless of which regions their subnets are in. If you need to bridge two different VPCs, that is a VPC peering / Border Gateway conversation, not a Cross-Region Network one. *** ## When You Don't Need to Configure Anything If your workloads live in the same VPC, and the regions you care about are in the default mesh, the path already exists — `ping` between regions works, and routes are installed in the VPC route table automatically. For anything beyond a connectivity check, **raise the Cross-Region Bandwidth** for the region pair you're using. The 10 Kbps free baseline is enough for connectivity checks (ping, traceroute) but not for production traffic; treat it as the "is this reachable" knob, not the "is this fast" knob. *** ## When You Do Need to Configure Something **Reaching a region that's not in the default mesh.** A small number of regions are regulated or otherwise segmented out of the Global mesh — China is the most common example. Those regions participate in Cross-Region Network differently, and traffic between them and the rest of the world is not automatic. Configure reachability to them explicitly when required. **Raising bandwidth between two regions.** Almost every real workload needs this. The 10 Kbps free baseline is enough for connectivity checks (ping, traceroute) but not for production traffic. Once you have actual traffic flowing — replication, application calls, anything — **purchase additional Cross-Region Bandwidth** for that region pair, anywhere from a few Mbps up to 10 Gbps and beyond. See [Cross-Region Bandwidth](/welcome/elastic-compute/networking/01-overview-1/03-region-links.md). **Exchanging routes dynamically with external networks.** Cross-Region Network carries traffic between subnets inside the same VPC. It does not speak BGP with your on-premises routers or other external networks. For that, use **Border Gateway**, which runs on top of Cross-Region Network and adds dynamic routing. *** ## What It Is Not * **It is not cross-VPC connectivity.** Cross-Region Network connects subnets within the same VPC across regions. It does not connect different VPCs to each other. * **It is not an internet gateway.** To give instances internet access, use a NAT Gateway or assign Elastic IPs. * **It is not a VPN or dedicated line to on-premises.** For connectivity to your own datacenter, use **Border Gateway**. * **It is not a load balancer or service mesh.** It operates at Layer 3 and does not terminate or inspect application traffic. *** ## In This Guide | Page | What You'll Learn | | ---------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------- | | [Default Connectivity](/welcome/elastic-compute/networking/01-overview-1/02-default-connectivity.md) | The Global VPC, its auto-provisioned per-region subnets, and the free default bandwidth | | [Cross-Region Bandwidth](/welcome/elastic-compute/networking/01-overview-1/03-region-links.md) | Raising bandwidth between specific region pairs, how it's priced, and how to purchase it | | [Routing](/welcome/elastic-compute/networking/01-overview-1/04-routing.md) | How cross-region subnets are reached via the existing VPC route table | | [Best Practices](/welcome/elastic-compute/networking/01-overview-1/05-best-practices.md) | Sizing, monitoring, and troubleshooting | *** ## Frequently Asked Questions **Do I have to do anything to turn this on?** No. The Global VPC exists in your account from day one, with per-region subnets already provisioned. Instances created in the Global VPC are immediately reachable from every other region in the default mesh, on the free baseline bandwidth. **Is the default connectivity really free?** Yes, but the free allocation is 10 Kbps — enough for connectivity checks (ping, traceroute) but not for production traffic. Production workloads need to raise the Cross-Region Bandwidth for the pair they're using. **Can instances in different regions ping each other by default?** Yes, as long as they're in the same VPC. An instance in Hong Kong can `ping` an instance in Frankfurt out of the box. Security groups do not gate intra-VPC cross-region traffic — the platform's security groups apply to public-network traffic, not private traffic inside a VPC. **Do I have to pick CIDRs for the default subnets?** No. In `VPC-Default`, the platform assigns each region a non-overlapping `/20`. You can see the assignments in the subnet list but you don't choose them. You only pick CIDRs for VPCs and subnets you create yourself. **Can Cross-Region Network connect two different VPCs?** No. It only connects subnets within the same VPC across regions. Two separate VPCs are independent, even if both have subnets in the same regions. **What about China and other regulated regions?** Regions under regulatory restrictions, China being the typical case, are not part of the default mesh. Reachability into or out of those regions is handled separately. Assume no default mesh for them and plan accordingly. **Does this support IPv6?** Yes. Cross-Region Network carries both IPv4 and IPv6 traffic between subnets in the same VPC. --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://docs.console.zenlayer.com/welcome/elastic-compute/networking/01-overview-1.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.