Amazon MemoryDB for Redis

Amazon MemoryDB for Redis Cheat Sheet 

• An in-memory database service for microservices-based applications.
• MemoryDB stores all of your data in memory, allowing you to achieve microsecond read and single-digit millisecond write latency and high throughput.
• It also uses a Multi-AZ transactional log to store data across multiple AZs in order to enable fast failover, database recovery, and node restarts.

Features

• Primary nodes have strong consistency, and replica nodes have guaranteed eventual consistency. 
• Scale horizontally by adding and removing nodes or vertically by changing node types.
  • • Write throughput – can be increased by adding shards.
    • Read throughput – can be increased by adding replicas.
• With Amazon S3, you can store snapshots with a retention period of up to 35 days.
• Automatic software patching and cluster upgrades.
• Supports both encryption in-transit / at-rest and authentication of users via ACLs.

Concepts

• Clusters

  • • A collection of one or more nodes serving a single dataset. 
    • A dataset is partitioned into shards, and each shard has a primary node and up to 5 optional replica nodes.
      • Primary node – serves both read and write requests.
      • Replica node – only serves read requests.
    • A cluster’s nodes are all designed to be the same node type, with the same parameter and security group settings.
    • When a primary node fails over to a replica node, that replica becomes the new primary node for that shard.
    • Each MemoryDB cluster has its own Redis engine version.
    • Each cluster endpoint has an address and a port that clients can use to discover the specific roles, IP addresses, and slots for each node in the cluster.
    • Supports data tiering using a node type from the r6gd family.
    • When you create or modify a cluster without specifying a preferred maintenance window, MemoryDB assigns a 60-minute maintenance window within your region’s maintenance window on a randomly selected day of the week.
    • Data is automatically backed up to a Multi-AZ transactional log, but you can choose to create point-in-time snapshots of a cluster either periodically or on-demand.

• Nodes

  • • The smallest component of a MemoryDB deployment that runs on an Amazon EC2 instance.
    • A node is part of a shard, which is part of a cluster.
    • Each node runs an instance of the engine at the version you specified when you set up your cluster. 
    • Every node has its own DNS name and port number.
    • Multiple types of nodes are supported, each with different amounts of memory and processing power.
    • When MemoryDB for Redis replaces a node, the new node may have a different IP address. You must update your application configuration to use the new IP address.

• Shards

  • • A hierarchical arrangement of nodes, each wrapped in a cluster.
    • A grouping of one to six nodes, with one serving as the primary write node and the other five as read replicas.
    • A cluster should always have at least one shard.
    • Supports both replication and multiple shards within a cluster.
• Parameter groups are named set of engine-specific settings that can be applied to a cluster so that all nodes in the cluster are configured identically.
• Subnet Groups are a set of subnets that you can assign to your clusters in an Amazon VPC environment.
• Access Control Lists (ACLs) are a collection of one or more users.
• Snapshots
  • • Supports automatic snapshots.
    • The snapshot window has a minimum length of 60 minutes.
    • You can set the retention limit of your snapshots in Amazon S3. The snapshot is automatically deleted when the retention period expires.

Security

• To control who has access to clusters, you can create users with specific permissions and add them to ACLs.
• Configure TLS encryption in transit and encryption at rest using AWS KMS keys. 
• Use a VPC security group to allow inbound traffic on the node’s endpoint and port to control which devices and Amazon EC2 instances can connect to MemoryDB clusters in a VPC.

Networking

• An Amazon EC2 instance in the same Amazon VPC can connect to your MemoryDB node. Alternatively, you can use VPC peering to connect to your MemoryDB node from another Amazon VPC.
• Your application uses the endpoint to connect to your cluster. 
• The endpoint is a unique address that identifies your cluster.

Pricing

• You can choose between two types of nodes:
  • • You are charged per hour for on-demand nodes.
    • By reserving nodes in 1-year or 3-year terms, you get discounts on the ongoing hourly usage rate. You can pay using any of the following methods:
      • No Upfront – low hourly rate.
      • Partial Upfront – lower hourly rate.
      • All Upfront – lowest hourly rate.
• You are charged based on the amount of data (in GB) written to the MemoryDB cluster. 
• You are charged for the additional snapshot storage.

Amazon MemoryDB for Redis Cheat Sheet References:

https://docs.aws.amazon.com/memorydb/latest/devguide/what-is-memorydb-for-redis.html

https://aws.amazon.com/memorydb/