> [!CAUTION]
> **Caution:** Support for backing up and restoring workloads managed using the legacy appliance management console is deprecated. For more information, see [Feature deprecations](https://docs.cloud.google.com/backup-disaster-recovery/docs/deprecations.md.txt) in the Backup and DR documentation.

Following are some important considerations that affect how you deploy
Backup and DR backup/recovery appliances:

- What are your organization's Recovery Time Objective (RTO) requirements?
  The RTO is the maximum amount of time that you can afford to be without your
  data. For example, if your RTO is 4 hours,then you need to be able to access
  your data within 4 hours of a failure.

- Do you need to centralize your backup management? You need to decide whether
  you want to manage your backups centrally or not.

  - Centralized backup management means that you have a single appliance management console to manage backups for all your workloads across all lines of business. This can be a more efficient way to manage backups, as you only need to manage a single appliance management console.
  - Decentralized backup management means that you have a separate appliance management console for each line of business. The mode of operation varies from organization to organization.
- What is your backup use case? Do you need remote backups for disaster recovery
  in the event of a disaster in a production region, or is it enough
  to protect your data locally? If you need disaster recovery capability, you must
  consider cross-region backups. This means storing your backups in multiple
  locations, so that if one location is affected by a disaster, your data will
  still be accessible.

## Workloads are in a single region

Your best backup strategy within a region depends upon your needs.

### If you don't need disaster recovery (DR)

For fastest performance and lower cost, deploy the
appliance management console and the backup/recovery appliances in the same
region where your workloads are running. Store your backup images in the same
region as your workloads.

If you also want to have backup copies off-site, then you can either store
backups in a different region or use dual-region or multi-region storage.
Storing backups in a different region incurs network and storage charges.

### If you need both backup and DR

For fastest performance and lower cost, deploy an appliance management console
in the same region as the production workload region and a second
appliance management console in a region that you can use for disaster
recovery.

Deploy backup/recovery appliances in both the production workload region and the
DR region to minimize recovery time objective (RTO). This makes sure that a
DR environment is fully pre-provisioned and available in the event of a disaster.

> [!NOTE]
> **Note:** To optimize DR cost, you can wait to deploy the DR backup/recovery appliances until you actually perform DR.

Store backup images in the production region and a copy in the disaster recovery
DR region, or use dual-region or multi-region storage. The production region
backup copy can meet routine backup needs with faster performance. Data copied
to your DR region can be used to recover your workloads in case the production
region is down.

## Workloads are in multiple regions

Your best backup strategy across regions depends upon your needs.

### If you don't need disaster recovery (DR)

> [!NOTE]
> **Note:** A regional outage where the appliance management console is deployed affects backup operations in other regions. This is an extremely rare event. If this is undesirable, then use the strategy in [If you need both backup and DR](https://docs.cloud.google.com/backup-disaster-recovery/management-console/deployment/deployment-considerations.md.txt#backup-and-dr).

For fastest performance and lower cost, deploy the appliance management console in one
of the regions where your workloads are running. This allows centralized
management across all the workloads and regions.

Deploy one or more backup/recovery appliances in each region where the workloads
are running. Store your backups in the same region as your workloads.

If you also want to have backup copies off-site, then you can either store
backups in a different region or use dual-region or multi-region storage.
Storing backups in a different region or multi-region incurs network and
storage charges.

### If you need both backup and DR

Deploy an appliance management console in each of the production workload
regions and another appliance management console in the DR region.

> [!NOTE]
> **Note:** If the production and DR regions overlap (for example, region A is production for some workloads and DR for others), you can use the same appliance management console for both production backups and DR recoveries in that region.

Deploy backup/recovery appliances in both the production workload regions and the
DR region to minimize recovery time objective (RTO). This ensures that a DR
environment is fully pre-provisioned and available in the event of a disaster.

> [!NOTE]
> **Note:** To reduce DR cost, you can choose to deploy the backup/recovery appliances later, when you actually perform DR.

Store backups in the production workload region and a copy in the DR region,
or use dual-region or multi-region storage. The production region backup copy
can be used to meet backup needs.

A backup image in the DR can be used to recover workloads if the production
region is down.

## Recommended network topology for Backup and DR

Google Cloud recommends the use of Shared VPC while deploying Backup and DR.
Shared VPC allows an
[organization](https://docs.cloud.google.com/resource-manager/docs/cloud-platform-resource-hierarchy.md.txt)
to connect resources from multiple projects to a common
[VPC](https://docs.cloud.google.com/vpc/docs/vpc.md.txt) (VPC) network, so that they can
communicate with each other securely and efficiently using internal IPs from
that network. When you use Shared VPC, you designate a project as a host
project and attach one or more other service projects to it. The VPC networks in
the host project are called Shared VPC networks.
[Eligible resources](https://docs.cloud.google.com/vpc/docs/shared-vpc.md.txt#resources_that_can_be_attached_to_shared_vpc_networks_from_a_service_project)
from service projects can use subnets in the Shared VPC network.

Shared VPC lets
[organization administrators](https://docs.cloud.google.com/resource-manager/docs/cloud-platform-resource-hierarchy.md.txt#organizations)
delegate administrative responsibilities, such as creating and managing
instances, to
[service project administrators](https://docs.cloud.google.com/vpc/docs/shared-vpc.md.txt#iam_in_shared_vpc)
while maintaining centralized control over network resources like subnets,
routes, and firewalls.

The appliance management console is activated into a [service producer VPC network](https://docs.cloud.google.com/vpc/docs/private-services-access.md.txt)
VPC. This service producer VPC communicates with your project using [Private Google Access](https://docs.cloud.google.com/vpc/docs/private-google-access.md.txt).
The primary purpose of this connection is for the appliance management console and the
backup/recovery appliances to exchange metadata. Backup traffic doesn't traverse
this link. However, an appliance management console needs to
communicate with all of the backup/recovery appliances deployed in any network.

### Shared VPC best practices

The following best practices are recommended:

- **Connecting to the appliance management console**: It's best to connect
  the service provider network to a Shared VPC within your network. All
  traffic from the appliance management console flows through this VPC,
  and therefore, through the host project. Provisioning the connectivity to the
  Backup and DR through a Shared VPC also enables a seamless
  connection between projects where the workloads are running (service projects)
  and the Backup and DR.

- **Backup/recovery appliance appliance location**: The backup/recovery appliances
  must be deployed in a subnet that has Private Google Access enabled for
  connectivity to the appliance management console. There are two recommended strategies
  for selecting the projects for the backup/recovery appliances:

  - **In the central host project**: In this strategy, Backup and DR is treated
    as a central service from IT. The central backup team governs provisioning
    of the service. As such, all backup/recovery appliances are provisioned in
    the host project enabling central admins to consolidate all backup resources
    into a central project. This approach has the benefit of consolidating all
    backup-related resources and their billing into a single project.

  - **In service projects**: This strategy is suitable for more decentralized teams
    where service projects are created and their management is delegated to
    distributed teams. In this scenario, the recommended best practice is to
    provision VPC for downstream service projects. Backup/recovery appliance
    are installed in the service projects within these VPCs. This enables colocation
    of the workload and the backup/recovery appliance within a single project.

  - **Private Google Access** : It is a best practice to enable Private Google Access
    for each subnet where you install a backup/recovery appliance. This ensures
    the backup/recovery appliance can communicate with APIs, such as Compute Engine,
    Cloud Storage, and Cloud Logging, which is important for monitoring and alerting
    to work. To simplify and enhance connections to Google Cloud APIs, consider
    configuring DNS resolution for `private.googleapis.com` as documented in the
    [Summary of configuration options](https://docs.cloud.google.com/vpc/docs/configure-private-google-access.md.txt#config-options)
    section. For Private Google Access, configure the firewall rules from the
    backup/recovery appliances to allow connections to the CIDR
    range `199.36.153.8/30` on TCP port 443.

## What's next

- [Set up and plan a backup/recovery appliance deployment](https://docs.cloud.google.com/backup-disaster-recovery/management-console/deployment/deployment-plan.md.txt)