Cross-Cloud Interconnect lets you connect your remote cloud network
to multiple Virtual Private Cloud (VPC) networks. You extend reachability
by adding multiple VLAN attachments to the connection.

As part of this functionality, you can create a VLAN attachment that uses a
Cross-Cloud Interconnect connection that's in a different
project. You can use this approach as long as both resources are in the same
[organization](https://docs.cloud.google.com/resource-manager/docs/cloud-platform-resource-hierarchy#organizations).

If appropriate, you can share connectivity between even more
VPC networks by using Shared VPC or
VPC Network Peering. For more information about these approaches, see
[Options for connecting to multiple VPC networks](https://docs.cloud.google.com/network-connectivity/docs/interconnect/how-to/enabling-multiple-networks-access-same-attachment).

> [!NOTE]
> **Note:** If you have VLAN attachments associated with multiple VPC networks that have overlapping IP address ranges, the conflicting IP ranges might cause traffic to be routed incorrectly by external routers. To avoid this issue, we recommend that you use a [custom mode VPC network](https://docs.cloud.google.com/vpc/docs/vpc#subnet-ranges) to assign unique IP address ranges to all of your VPC subnets.

## Required permissions

*In the project where you create the attachment* , you need the
[Network Admin role](https://docs.cloud.google.com/compute/docs/access/iam#compute.networkAdmin)
(`roles/compute.networkAdmin`) .

*In the project that contains the connection*, you need one of the following:

- The [Network Admin role](https://docs.cloud.google.com/compute/docs/access/iam#compute.networkAdmin) (`roles/compute.networkAdmin`) or the `compute.interconnects.get` permission
- Minimally, the `compute.interconnects.use` permission

### compute.interconnects.get

If the user has the `compute.interconnects.get` permission, then the process
of creating the attachment is the same as it is for a network administrator.

### compute.interconnects.use

If the user has the `compute.interconnects.use` permission and does not have
the `compute.interconnects.get` permission, then the user
does not have visibility into what type of remote cloud the connection uses.
In this case, the Google Cloud console can't detect requirements related to the
remote location.

For example, attachments to Azure require using an IPv4 subnet length of
`/30`.
For an administrative user, the Google Cloud console removes the
option to
request a `/29` IPv4 subnet. However, it does not do so for users who have only
the `compute.interconnects.use` permission.

Some constraints are enforced by the API, and in such cases, the
Google Cloud console returns an error if an invalid parameter is specified.

Other constraints are merely advisory, to guide users toward known valid
configurations. For such constraints, the absence of Google Cloud console
advice can result in incompatible configuration choices that are not discovered
until further along in the configuration sequence.

The following are sample requirements:

- Microsoft Azure attachments require using an IPv4 subnet length of `/30`.
- Azure attachments require using the same VLAN ID on both ports of a pair.
- Amazon Web Services (AWS) attachments require using MD5 authentication for Border Gateway Protocol (BGP) sessions.
- Oracle Cloud Infrastructure (OCI) attachments require using a VLAN ID \>= `100`. Auto-allocated VLAN IDs meet this requirement.

To learn about constraints, consult the following pages:

- [Create VLAN attachments (AWS)](https://docs.cloud.google.com/network-connectivity/docs/interconnect/how-to/cci/aws/configure-google-resources#create-attachments)
- [Create VLAN attachments (Azure)](https://docs.cloud.google.com/network-connectivity/docs/interconnect/how-to/cci/azure/configure-google-resources#create-attachments)
- [Create VLAN attachments (OCI)](https://docs.cloud.google.com/network-connectivity/docs/interconnect/how-to/cci/oci/configure-google-resources#create-attachments)

When you grant a user the `compute.interconnects.use` permission, inform
them of any constraints.

## Create a VLAN attachment

To create a VLAN attachment that uses a Cross-Cloud Interconnect
connection from another project, follow these steps.

### Console

1. In the Google Cloud console, go to the **VLAN attachments** tab.

   [Go to VLAN attachments](https://console.cloud.google.com/hybrid/interconnects/list?tab=attachments)
2. Select the project in which you need to create a VLAN attachment by using
   the project picker.

3. Click **Create VLAN attachments** , and then select **Cross-Cloud Interconnect connection**.

4. In the **Encrypt interconnect** section, select **Set up unencrypted Interconnect** ,
   and then click **Continue**.

5. On the **Choose an Interconnect** page,
   select **In another project**. Enter the following information:

   - **Project ID**: Enter the ID of the project that contains the Cross-Cloud Interconnect connection.
   - **Interconnect name**: Enter the name of the connection.
6. Click **Continue**.

7. For **VLAN attachments** , click **Add VLAN attachment**. Provide
   the following information:

   - **Name**: Specify a name for the VLAN attachment.
   - **Cloud Router**: Choose an existing Cloud Router or create a new one. The Cloud Router that you choose defines the region and VPC network to which your VLAN attachment is available. The Google Cloud ASN is also defined by the Cloud Router that you choose.
8. Click **Done** , and then click **Create**. The attachment takes a few
   minutes to create.

9. After the attachment is created, click **Configure** to create a BGP
   session on the chosen Cloud Router. The Google and Peer BGP IP
   addresses are selected for you automatically.

10. After you add the BGP session, click **Save configuration**. The BGP
    session is not active until you configure resources in your remote
    cloud.

### gcloud

1. If you have not already created a Cloud Router, create one by
   running the following command:

   ```
   gcloud compute routers create ROUTER_NAME \
       --region=REGION \
       --asn=GOOGLE_ASN \
       --network=NETWORK \
       --project=PROJECT_ID
   ```

   Replace the following:
   - `ROUTER_NAME`: a name that you specify for your Cloud Router
   - `REGION`: the Google Cloud region in which your Cloud Router is created; this region must match the region used by the VLAN attachment
   - `GOOGLE_ASN`: the [private ASN](https://tools.ietf.org/html/rfc6996) (`16550`, `64512 - 65534`, `4200000000 - 4294967294`); it is used for all BGP sessions on the same Cloud Router, and it cannot be changed later
   - `NETWORK`: the name of the network to which the Cloud Router manages routes; this is the same network that your VLAN attachment uses
   - `PROJECT_ID`: the ID for the project in which both the Cloud Router and VLAN attachment exist; this project ID is different from the project where the connection is located
2. In the project that contains your connections, list all the
   Cloud Interconnect connections:

   ```
   gcloud compute interconnects list \
       --project=INTERCONNECT_PROJECT_ID
   ```

   Replace `INTERCONNECT_PROJECT_ID` with
   that project's ID. Determine the name of the connection
   that you need to use.
3. Determine the self link for the connection that you need to
   use:

   ```
   gcloud compute interconnects describe INTERCONNECT_NAME \
       --project=INTERCONNECT_PROJECT_ID \
       --format="get(selfLink)"
   ```

   Replace `INTERCONNECT_NAME` with its name, and replace
   `INTERCONNECT_PROJECT_ID` with the ID of the project
   that contains it.
4. Create the VLAN attachment:

   ```
   gcloud compute interconnects attachments dedicated create VLAN_ATTACHMENT_NAME \
       --region=REGION \
       --router=ROUTER_NAME \
       --project=PROJECT_ID \
       --interconnect=INTERCONNECT_SELF_LINK \
       --candidate-subnets=CANDIDATE_SUBNETS \
       --vlan=VLAN_ID
   ```

   Replace the following:
   - `VLAN_ATTACHMENT_NAME`: a name that you specify for your VLAN attachment
   - `REGION`: must match the same region as the associated Cloud Router
   - `ROUTER_NAME`: the name of the Cloud Router from the first step
   - `PROJECT_ID`: the ID for the project in which both the Cloud Router and the VLAN attachment exist; this project ID is different from the project where the connection is located
   - `INTERCONNECT_SELF_LINK`: the self link for the connection that the VLAN attachment uses; the self link includes the ID of the project that contains the connection itself
   - `CANDIDATE_SUBNETS`: an optional flag that lets you specify comma-delimited ranges of link-local IP addresses (as `CANDIDATE_SUBNETS`) to be used for the BGP session that manages routes for the VLAN attachment; for more information, see the [Google Cloud CLI documentation](https://docs.cloud.google.com/sdk/gcloud/reference/compute/interconnects/attachments/dedicated/create#--candidate-subnets)
   - `VLAN_ID`: an optional flag that lets you specify a VLAN ID; for more information, see the [Google Cloud CLI
     documentation](https://docs.cloud.google.com/sdk/gcloud/reference/compute/interconnects/attachments/dedicated/create#--vlan)
5. Describe the VLAN attachment that you just created:

   ```
   gcloud compute interconnects attachments dedicated describe VLAN_ATTACHMENT_NAME \
       --region=REGION \
       --project=PROJECT_ID \
       --format="get(cloudRouterIpAddress,customerRouterIpAddress,tag8021q)"
   ```

   Replace the following:
   - `VLAN_ATTACHMENT_NAME`: its name
   - `REGION`: its region
   - `PROJECT_ID`: its project ID

   Note the following:
   - `cloudRouterIpAddress` is the BGP IP address that is used by the BGP session on the associated Cloud Router.
   - `customerRouterIpAddress` is the BGP IP address that you use to configure the BGP session in your remote cloud.
   - `tag8021q` is the VLAN ID, which you might have specified manually in the previous step.
6. Create an interface on the Cloud Router:

   ```
   gcloud compute routers add-interface ROUTER_NAME \
       --interconnect-attachment=VLAN_ATTACHMENT_NAME \
       --region=REGION \
       --interface-name=INTERFACE_NAME \
       --project=PROJECT_ID \
   ```

   Replace the following:
   - `ROUTER_NAME`: the name of the Cloud Router from the first step
   - `VLAN_ATTACHMENT_NAME`: the name of the VLAN attachment that you created and described in these steps
   - `REGION`: the region used by the Cloud Router and the VLAN attachment
   - `INTERFACE_NAME`: a name that you specify for the new interface on the Cloud Router
   - `PROJECT_ID`: the ID for the project in which both the Cloud Router and VLAN attachment exist; this project ID is different from the project where the connection is located
   - `CLOUD_ROUTER_IP`: the `cloudRouterIpAddress` that you determined in the previous step
7. Add a BGP peer to the new interface on the Cloud Router:

   ```
   gcloud compute routers add-bgp-peer ROUTER_NAME \
       --region=REGION \
       --interface-name=INTERFACE_NAME \
       --peer-name=BGP_PEER_NAME \
       --project=PROJECT_ID \
       --peer-asn=PEER_ASN \
       --advertised-route-priority=PRIORITY \
       --advertisement-mode=ADVERTISEMENT_MODE
   ```

   Replace the following:
   - `ROUTER_NAME`: the name of the Cloud Router from the first step
   - `REGION`: the region used by the Cloud Router and the VLAN attachment
   - `INTERFACE_NAME`: the name of the interface that you created in the previous step
   - `BGP_PEER_NAME`: a name that you specify for the BGP peer
   - `PROJECT_ID`: the ID for the project in which both the Cloud Router and the VLAN attachment exist; this project ID is different from the project where the connection is located
   - `CUSTOMER_ROUTER_IP`: the BGP IP address for your remote cloud; this is the `customerRouterIpAddress` that you noted when you described your VLAN attachment
   - `PEER_ASN`: the ASN for your remote cloud
   - `PRIORITY`: an optional flag that you can use to set the base priority for the routes that the Cloud Router advertises to the remote cloud; for more details about this option and the base metric, see [Advertised prefixes and priorities](https://docs.cloud.google.com/network-connectivity/docs/router/concepts/advertised-routes#advertised-priority)
   - `ADVERTISEMENT_MODE`: an optional flag that you can use if you want to customize the routes that the Cloud Router advertises to the remote cloud; for information about default and custom advertisements, see [route advertisements](https://docs.cloud.google.com/network-connectivity/docs/router/concepts/advertised-routes#overview-am)

## What's next

- To learn more about Cross-Cloud Interconnect, see the
  [Cross-Cloud Interconnect overview](https://docs.cloud.google.com/network-connectivity/docs/interconnect/concepts/cci-overview).

- To learn about common issues that you might encounter when using
  Cross-Cloud Interconnect, see
  [Troubleshooting](https://docs.cloud.google.com/network-connectivity/docs/interconnect/support/troubleshooting#cci).