Accessing your ElastiCache cluster or replication group - Amazon ElastiCache (Redis OSS)
Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Accessing your ElastiCache cluster or replication group

Your Amazon ElastiCache instances are designed to be accessed through an Amazon EC2 instance.

If you launched your ElastiCache instance in an Amazon Virtual Private Cloud (Amazon VPC), you can access your ElastiCache instance from an Amazon EC2 instance in the same Amazon VPC. Or, by using VPC peering, you can access your ElastiCache instance from an Amazon EC2 in a different Amazon VPC.

If you launched your ElastiCache instance in EC2 Classic, you allow the EC2 instance to access your cluster by granting the Amazon EC2 security group associated with the instance access to your cache security group. By default, access to a cluster is restricted to the account that launched the cluster.

Grant access to your cluster or replication group

You launched your cluster into EC2-VPC

If you launched your cluster into an Amazon Virtual Private Cloud (Amazon VPC), you can connect to your ElastiCache cluster only from an Amazon EC2 instance that is running in the same Amazon VPC. In this case, you will need to grant network ingress to the cluster.

Note

If your are using Local Zones, make sure you have enabled it. For more information, see Enable Local Zones. By doing so, your VPC is extended to that Local Zone and your VPC will treat the subnet as any subnet in any other Availability Zone and relevant gateways, route tables and other security group considerations. will be automatically adjusted.

To grant network ingress from an Amazon VPC security group to a cluster
  1. Sign in to the Amazon Web Services Management Console and open the Amazon EC2 console at https://console.amazonaws.cn/ec2/.

  2. In the navigation pane, under Network & Security, choose Security Groups.

  3. From the list of security groups, choose the security group for your Amazon VPC. Unless you created a security group for ElastiCache use, this security group will be named default.

  4. Choose the Inbound tab, and then do the following:

    1. Choose Edit.

    2. Choose Add rule.

    3. In the Type column, choose Custom TCP rule.

    4. In the Port range box, type the port number for your cluster node. This number must be the same one that you specified when you launched the cluster. The default port for Redis OSS is 6379.

    5. In the Source box, choose Anywhere which has the port range (0.0.0.0/0) so that any Amazon EC2 instance that you launch within your Amazon VPC can connect to your ElastiCache nodes.

      Important

      Opening up the ElastiCache cluster to 0.0.0.0/0 does not expose the cluster to the Internet because it has no public IP address and therefore cannot be accessed from outside the VPC. However, the default security group may be applied to other Amazon EC2 instances in the customer’s account, and those instances may have a public IP address. If they happen to be running something on the default port, then that service could be exposed unintentionally. Therefore, we recommend creating a VPC Security Group that will be used exclusively by ElastiCache. For more information, see Custom Security Groups.

    6. Choose Save.

When you launch an Amazon EC2 instance into your Amazon VPC, that instance will be able to connect to your ElastiCache cluster.

Accessing ElastiCache resources from outside Amazon

Amazon ElastiCache is an Amazon service that provides cloud-based in-memory key-value store. The service is designed to be accessed exclusively from within Amazon. However, if the ElastiCache cluster is hosted inside a VPC, you can use a Network Address Translation (NAT) instance to provide outside access.

Requirements

The following requirements must be met for you to access your ElastiCache resources from outside Amazon:

  • The cluster must reside within a VPC and be accessed through a Network Address Translation (NAT) instance. There are no exceptions to this requirement.

  • The NAT instance must be launched in the same VPC as the cluster.

  • The NAT instance must be launched in a public subnet separate from the cluster.

  • An Elastic IP Address (EIP) must be associated with the NAT instance. The port forwarding feature of iptables is used to forward a port on the NAT instance to the cache node port within the VPC.

Considerations

The following considerations should be kept in mind when accessing your ElastiCache resources from outside ElastiCache.

  • Clients connect to the EIP and cache port of the NAT instance. Port forwarding on the NAT instance forwards traffic to the appropriate cache cluster node.

  • If a cluster node is added or replaced, the iptables rules need to be updated to reflect this change.

Limitations

This approach should be used for testing and development purposes only. It is not recommended for production use due to the following limitations:

  • The NAT instance is acting as a proxy between clients and multiple clusters. The addition of a proxy impacts the performance of the cache cluster. The impact increases with number of cache clusters you are accessing through the NAT instance.

  • The traffic from clients to the NAT instance is unencrypted. Therefore, you should avoid sending sensitive data via the NAT instance.

  • The NAT instance adds the overhead of maintaining another instance.

  • The NAT instance serves as a single point of failure. For information about how to set up high availability NAT on VPC, see High Availability for Amazon VPC NAT Instances: An Example.

How to access ElastiCache resources from outside Amazon

The following procedure demonstrates how to connect to your ElastiCache resources using a NAT instance.

These steps assume the following:

  • iptables -t nat -A PREROUTING -i eth0 -p tcp --dport 6380 -j DNAT --to 10.0.1.231:6379

  • iptables -t nat -A PREROUTING -i eth0 -p tcp --dport 6381 -j DNAT --to 10.0.1.232:6379

Next you need NAT in the opposite direction:

iptables -t nat -A POSTROUTING -o eth0 -j SNAT --to-source 10.0.0.55

You also need to enable IP forwarding, which is disabled by default:

sudo sed -i 's/net.ipv4.ip_forward=0/net.ipv4.ip_forward=1/g' /etc/sysctl.conf sudo sysctl --system

  • You are accessing a Redis OSS cluster with:

    • IP address – 10.0.1.230

    • Default Redis OSS port – 6379

    • Security group – sg-bd56b7da

    • Amazon instance IP address – sg-bd56b7da

  • Your trusted client has the IP address 198.51.100.27.

  • Your NAT instance has the Elastic IP Address 203.0.113.73.

  • Your NAT instance has security group sg-ce56b7a9.

To connect to your ElastiCache resources using a NAT instance
  1. Create a NAT instance in the same VPC as your cache cluster but in a public subnet.

    By default, the VPC wizard will launch a cache.m1.small node type. You should select a node size based on your needs. You must use EC2 NAT AMI to be able to access ElastiCache from outside Amazon.

    For information about creating a NAT instance, see NAT Instances in the Amazon VPC User Guide.

  2. Create security group rules for the cache cluster and NAT instance.

    The NAT instance security group and the cluster instance should have the following rules:

    • Two inbound rules

      • One to allow TCP connections from trusted clients to each cache port forwarded from the NAT instance (6379 - 6381).

      • A second to allow SSH access to trusted clients.

      NAT instance security group - inbound rules
      Type Protocol Port range Source
      Custom TCP Rule TCP 6379-6380 198.51.100.27/32
      SSH TCP 22 203.0.113.73/32
    • An outbound rule to allow TCP connections to cache port (6379).

      NAT instance security group - outbound rule
      Type Protocol Port range Destination
      Custom TCP Rule TCP 6379 sg-ce56b7a9 (Cluster instance Security Group)
    • An inbound rule for the cluster's security group that allows TCP connections from the NAT instance to the cache port (6379).

      Cluster instance security group - inbound rule
      Type Protocol Port range Source
      Custom TCP Rule TCP 6379 sg-bd56b7da (Cluster Security Group)
  3. Validate the rules.

    • Confirm that the trusted client is able to SSH to the NAT instance.

    • Confirm that the trusted client is able to connect to the cluster from the NAT instance.

  4. Add an iptables rule to the NAT instance.

    An iptables rule must be added to the NAT table for each node in the cluster to forward the cache port from the NAT instance to the cluster node. An example might look like the following:

    iptables -t nat -A PREROUTING -i eth0 -p tcp --dport 6379 -j DNAT --to 10.0.1.230:6379

    The port number must be unique for each node in the cluster. For example, if working with a three node Redis OSS cluster using ports 6379 - 6381, the rules would look like the following:

    iptables -t nat -A PREROUTING -i eth0 -p tcp --dport 6379 -j DNAT --to 10.0.1.230:6379 iptables -t nat -A PREROUTING -i eth0 -p tcp --dport 6380 -j DNAT --to 10.0.1.231:6379 iptables -t nat -A PREROUTING -i eth0 -p tcp --dport 6381 -j DNAT --to 10.0.1.232:6379
  5. Confirm that the trusted client is able to connect to the cluster.

    The trusted client should connect to the EIP associated with the NAT instance and the cluster port corresponding to the appropriate cluster node. For example, the connection string for PHP might look like the following:

    redis->connect( '203.0.113.73', 6379 ); redis->connect( '203.0.113.73', 6380 ); redis->connect( '203.0.113.73', 6381 );

    A telnet client can also be used to verify the connection. For example:

    telnet 203.0.113.73 6379 telnet 203.0.113.73 6380 telnet 203.0.113.73 6381
  6. Save the iptables configuration.

    Save the rules after you test and verify them. If you are using a Redhat-based Linux distribution (like Amazon Linux), run the following command:

    service iptables save

Related topics

The following topics may be of additional interest.