What Is an In-Memory Database?

Latency is the lag between the request to access data and the application's response. In-memory databases offer predictable low latencies irrespective of scale. They deliver microsecond read latency, single-digit millisecond write latency, and high throughput.

As a result, in-memory storage allows enterprises to make data-based decisions in real-time. You can design applications that process data and respond to changes before it's too late. For example, in-memory computing of sensor data from self-driving vehicles can give the desired split-second response time for emergency braking.

In-memory databases are known for their high throughput. Throughput refers to the number of read (read throughput) or write (write throughput) operations over a given period of time. Examples include bytes/minute or transactions per second.

You can scale your in-memory database to meet fluctuating application demands. Both write and read scaling is possible without adversely impacting performance. The database stays online and supports read-and-write operations during resizing.

A cache is a high-speed data storage layer that stores a subset of typically transient data. A cache's primary purpose is to increase data retrieval performance by reducing the need to access the underlying slower storage layer. This means future requests for that data are served up faster than is possible by accessing the data’s primary storage location.

Caching allows you to reuse previously retrieved or computed data efficiently. In-memory data storage works well for faster access to cached data. Caching trades off durability for response time. Response time is faster because data is retrieved from memory, but caching doesn't protect against the loss of data in memory. This is why caching is often used in combination with a disk-based durable database.

Real-time bidding refers to the buying and selling of online ad impressions. Usually, the bid must be made while the user loads a webpage in 100-120 milliseconds and sometimes as little as 50 milliseconds. During this period, real-time bidding applications request bids from all buyers for the ad spot, select a winning bid based on multiple criteria, display the offer, and collect post-ad-display information. In-memory databases are ideal for ingesting, processing, and analyzing real-time data with sub-millisecond latency.

A relative gaming leaderboard shows a gamer's position relative to other players of a similar rank. These leaderboards can help build engagement among players and keep gamers from becoming demotivated compared to top players. In-memory databases can quickly deliver sorting results and update the leaderboard in real-time for a game with millions of players.

An in-memory cache relies on random access memory (RAM) to store data. The technology stores data tables directly in the RAM instead of on external drives. Specialized data structures allow the indexing of your data records. The indexes act as direct pointers to specific rows and columns. However, the actual physical data is compressed and non-relational in format. When you make an access request, the database navigates to the exact data value using the index. Stored data is always available in a directly usable format.

Things like 64-bit computing, lower RAM prices, and multi-core servers have made in-memory storage more common. In addition, cloud-based data stores allow you to scale your RAM resources up or down as required, making in-memory technology more flexible and accessible.

A traditional database persists all data to external disk drives or solid-state drives. Every single read-and-write operation requires disk access. Conversely, in-memory cache does not prioritize data persistence. For example, caches may only save data periodically to external storage media. We have summarized the differences between in-memory caches and traditional databases below.

Snapshot files record the database state at a given moment in time. The in-memory cache generates snapshots periodically or during a controlled shutdown. While snapshotting improves durability to some extent, data loss may still occur between snapshots.

Transaction logging records changes to the database in an external journal file. Logging is independent of data read/write and does not impact performance. The journal file facilitates the automatic recovery of an in-memory cache.

Some in-memory caches rely on redundancy to provide high availability. They maintain multiple copies of the same data in different memory modules. Module failure results in automatic failover to the duplicate backup copy. This mitigates the risk of data loss with a cache.

Amazon MemoryDB

Amazon MemoryDB is a durable, in-memory database service that delivers ultra-fast performance. It is Redis OSS-compatible, enabling customers to quickly build applications using the same flexible and friendly Redis OSS data structures, APIs, and commands they already use today. MemoryDB also stores your data durably across multiple Availability Zones (AZs) using a Multi-AZ transactional log to enable fast failover, database recovery, and node restarts.

Amazon ElastiCache

Amazon ElastiCache is a blazing-fast, in-memory caching service that enables microsecond latency to power real-time internet-scale applications. It is Redis OSS and Memcached compatible. Developers can use ElastiCache as an in-memory cache, or use cases that don't require high data durability. The ElastiCache cluster configuration supports customers to run workloads with up to 6.1 TB of in-memory capacity in a single cluster. ElastiCache also provides the ability to add and remove shards from a running cluster. You can dynamically scale your ElastiCache cluster workloads in and out to adapt to changes in demand.

Get started with in-memory caches or databases on AWS by creating a free account today!