This page explains networking for Memorystore for Redis Cluster.
On this page:
Networking setup guidance
As a reader of this page, you likely fit into one of two roles, each having different tasks to accomplish (you may have overlap between these roles). Knowing which role you fit in and your role's goals helps you accomplish your networking and instance creation tasks quickly and efficiently.
Two possible roles you may fit are as follows:
Role 1 – Redis Admin
Your goal is to create a Memorystore for Redis Cluster instance. You are reading this page to learn if you have the required prerequisites necessary to create the instance.
Once you know the proper service connection policy has been established for your network, your goal is to get the full network name (that has the format
projects/NETWORK_PROJECT_ID/global/networks/NETWORK_ID
) from your Network Admin so you can use it to Create an instance.
Role 2 – Network Admin
Your goal is to find out if the proper service connection policy has been created for the network on which the Redis Admin wants to deploy a Memorystore for Redis Cluster instance. If the service connection policy hasn't been created, create it. The purpose of the policy is to allow Memorystore to automate private connectivity to the Memorystore service. For instructions on how to configure and manage a service connection policy, see Configure service connection policies.
- Use the service class
gcp-memorystore-redis
when creating the service network policy.
- Use the service class
Your next goal is to provide the network name to the Redis Admin so they can use it to create an instance.
If you are the Network Admin, you must have the required roles to create a service connection policy.
Memorystore for Redis Cluster has the following networking characteristics:
The only networking connectivity method that can be used for Memorystore for Redis Cluster is Private Service Connect service connectivity automation (enabled by service connection policies). For more details, see Service connection policies.
If the correct service connection policy exists, service connectivity automation automatically deploys connectivity for a Memorystore for Redis Cluster instance during instance creation.
Prerequisites required before creating a cluster
As described in Service connection policies,
a service connection policy is unique to your network, region, and service class.
A service connection policy must exist for your region, network, and
gcp-memorystore-redis
service class before you can create a
Memorystore for Redis Cluster instance. You must also make sure the necessary APIs
are enabled before you can create an instance. For additional details about
how the Memorystore for Redis Cluster service creates a
Private Service Connect endpoint and the lifecycle of that endpoint,
see Deploy a managed service and configure connectivity.
Communicate networking requirements
If you are reading this as a Redis Admin, ask your Network Admin if a service
connection policy exists for the region, network, and gcp-memorystore-redis
service class where you want to create your instance. Once your Network Admin
has created the policy, ask them for the full network name (that has the format projects/NETWORK_PROJECT_ID/global/networks/NETWORK_ID
) so you can use it
to create a Memorystore for Redis Cluster instance.
Send your Network Admin a link to this page so they can understand the service connection policy prerequisites needed for you to create an instance.
Enable APIs
As a Redis Admin, before you can create a Memorystore for Redis Cluster instance, you must enable all of the APIs listed in Before you begin.
Shared VPC
In addition to standard VPC networks, Memorystore for Redis Cluster also supports Shared VPC networks.
Shared VPC setups have a host project and one or more service projects. The service connection policy for Memorystore for Redis Cluster is defined in the host project by the Network Admin. Redis Admins typically create Memorystore for Redis Cluster instances in service projects.
For a quickstart on creating an instance with Shared VPC, see Instance provisioning on a Shared VPC network.
Reserved network addresses
After you successfully create a Memorystore for Redis Cluster instance, Memorystore reserves two network addresses for the instance. They are used to serve the traffic for your instance. One of these is the discovery endpoint that you use to connect to your instance.
Supported networking architecture
Memorystore for Redis Cluster supports the network architectures described in this section.
Same network, project, and region client access example
In this example the client and Memorystore endpoint IPs are located in the same network, project, and region.
Same network and project, but multi region client access example
In this example the client and Memorystore endpoint IPs are located in the same network and project, but in multiple regions.
Shared VPC client access example
In this example the clients are located in different Shared VPC projects. Although clients in this example are in the same region, clients from different regions are also supported.
On-premises access example
This diagram shows an example of a client connecting to Memorystore from an on-premises network using Cloud Interconnect and Cloud Router. Although Cloud Interconnect and Cloud Router intfrastructure is used, the client machines in the on-premises network connect to Redis using the Memorystore endpoint IP addresses. For example, in the diagram below, clients connect directly to 10.142.0.10 and 10.142.0.11.
For instruction on finding your instance's discovery endpoint IP address, see View your cluster's discovery endpoint.
Multi VPC network access
If you want to connect to Memorystore for Redis Cluster across multiple VPCs, you can do so by creating an internal VPN tunnel between the VPC that has the PSC endpoint and any other VPCs that require access. Since multiple VPCs cannot directly connect to Private Service Connect, VPN tunnel infrastructure must be used.
Frequently asked questions
This section covers networking FAQs for Memorystore for Redis Cluster.
Why do I need a service connection policy?
You must define a service connection policy before you can create a Memorystore for Redis Cluster instance because Memorystore for Redis Cluster uses Private Service Connect service connectivity automation to automate deployment and connectivity in the consumer network. This requires authorization from the consumer network admin. A service connection policy defines this authorization policy.
Why do the Network Connectivity and Service Consumer Management APIs need to be enabled?
Memorystore for Redis Cluster uses Private Service Connect service connectivity automation to automate deployment and connectivity in the consumer network. The automation needs these APIs to be enabled. If they are not enabled, cluster creation operations fail.
What permissions are needed for setting up Memorystore for Redis Cluster networking?
If you are a Redis Admin trying to accomplish the Redis Admin tasks described on this page, you need the
redis.admin
role. The Permissions and their roles section explains what roles are needed for different Redis cluster permissions.If you are a Network Admin trying to accomplish the Network Admin tasks described on this page, you need the
compute.networkAdmin
role.
How can set up connectivity for my on-premises network?
In addition to the guidance explained on this page, you can learn about setting up on-premises connectivity at: