Understanding Amazon NoSQL data stores - Choosing an AWS NoSQL Database

This whitepaper is for historical reference only. Some content might be outdated and some links might not be available.

Understanding Amazon NoSQL data stores

AWS provides the broadest selection of managed NoSQL databases, allowing you to save, grow, and innovate faster. With Amazon NoSQL databases, you get high performance, enterprise-grade security, automatic, and instant scalability. The following table lists some of the AWS managed NoSQL database services offered, and their key characteristics:

Table 1 — AWS database service comparison

AWS database service Use cases Strengths Security Performance Cross-Region resiliency

Amazon DocumentDB (with MongoDB compatibility)*

  • Mid TB range

  • Data format: JSCO, BSON, XML

  • NoSQL type: document

  • Consistency: strong/eventual

User profile/personalization, catalogs, mobile, and content management, retail and marketing (for example, tracking customers who purchase similar items) Flexible schema and indexing, ad hoc queries on any attributes, including nested attributes Capability to enable data encryption at rest and in transit Millions of requests per second with millisecond latency

DocumentDB Global clusters

Amazon DynamoDB*

  • High TB range

  • Data format: JSON, BSON, or XML

  • NoSQL type: key-value, document

  • Consistency: strong/eventual

  • User preferences

  • Session management

  • Shopping cart

  • Product catalog

  • High-traffic web apps

  • Near real-time bidding

  • Performance at scale

  • Serverless

  • Simple data model

Encrypts all data at rest by default, row/column level security Single digit millisecond at any scale, In-memory acceleration with DynamoDB (DAX) for low latency access to eventually consistent data Global tables

Amazon Keyspaces (for Apache Cassandra)

  • Data format: JSON

  • NoSQL type: wide column

  • Consistency: one, local_one, local-quorum

High scalable apps for:

  • Equipment maintenance

  • Fleet management

  • Route optimization

  • Extreme write speeds with relatively less velocity reads

  • Being serverless, allocates storage and read/write throughput directly to tables

  • Tables are encrypted by default

  • Capability to enable data encryption at rest and in transit

Single digit millisecond response at any scale Multi-Region Replication

Amazon Neptune*

  • Mid TB range

  • Data format: Germalin, RDF, open Cypher

  • NoSQL type: graph

  • Consistency: immediate consistency

  • Recommendations

  • Social patterns

  • relationship traversal

  • Fraud detection

  • Risk assessment

  • Highly connected data is locally indexed and purpose-built to answer questions about relationships

  • Optimized for efficient storage and retrieval

Capability to enable data encryption at rest and in transit High throughput, low latency
  • Cross-Region snapshot

  • Neptune streams to replicate data between cross-Region clusters

Amazon Timestream*

  • NoSQL type: TimeSeries

  • Consistency: eventual

  • Server metrics

  • Application performance monitoring

  • Network data

  • IoT apps

  • Sensor data

  • Events

  • Clicks

  • Financial forecasting

  • Many other types of analytics data

Analytics over time series data

Encrypts all data by default Capable of ingesting trillions of events daily. The adaptive SQL query engine provides rapid point-in-time queries with its in-memory store, and fast analytical queries through its magnetic store Cross-Region backup

Amazon ElastiCache (Memcached)

  • Low TB range

  • NoSQL type: in-memory, key-value

  • Caching repeat requests

  • Sticky sessions (to store session state)

  • Simple caching model

  • Multi-threaded performance

Capability to enable data encryption at rest and in transit Sub-millisecond response time N/A
Amazon ElastiCache (Redis OSS)
  • Low TB range

  • NoSQL type: in-memory, key-value

  • Gaming leaderboards

  • Geospatial applications

  • Complex data structures

  • Sorting and ranking

  • Pub/sub messaging

  • Geospatial capabilities

Capability to enable data encryption at rest and in transit Sub-millisecond response time Global datastores

Amazon MemoryDB

  • NoSQL type: in-memory, database

  • Consistency: strong/eventual

  • High concurrency

  • Streaming media

  • Data feeds

  • Durable database

  • Complex data structures

Capability to enable data encryption at rest and in transit Microsecond read and single-digit millisecond write latency Cross-Region snapshot

* ACID compliant