Comparing Redis and Memcached

Redis and Memcached are both popular open-source in-memory databases which are easy to use and provide higher performance than relational databases.

How can I select between the two key-value databases?

Memcached is suitable for storing simple data structures, whereas Redis is suitable for storing more complex, larger data that requires persistency.

For details, see the following table.

**Table 1** Differences between Redis and Memcached
Item	Redis	Memcached
Latency	In-memory database with sub-millisecond latency	In-memory database with sub millisecond latency
Ease of use	Simple syntax and easy to use	Simple syntax and easy to use
Distributed storage	Horizontal expansion in cluster mode	Supported
Multi-language client	Supports client connections in more than 30 languages including Java, C, and Python.	Supports client connections in more than 10 languages including Java, C, and Python.
Thread/Process	Single-core and single-thread Single-thread communication, avoiding unnecessary context switching and contention Non-blocking I/O (I/O multiplexing) is used to reduce resource consumption when multiple clients are connected.	Multi-thread and scalable The Memcached performance can be improved by increasing the number of CPUs. There is an obvious performance advantage in the scenario where the value of key is great.
Persistent storage	Supported Each write operation (adding, deleting, or modifying data) can be recorded on disk (AOF file).	Supported NOTE: Persistence is not supported by open-source Memcached, but is supported by DCS for Memcached.
Data structure	Supports complex data structures such as hash, list, set, and sorted set, catering to various scenarios.	Supports simple strings.
Lua script support	Supported	Not supported
Snapshot backup	Supported Snapshots are generated periodically. Therefore, there is no guarantee that data will not be lost. Redis forks a subprocess to generate snapshots. When there is a large amount of data, the Redis service may be interrupted for a short time.	Not supported
Key value restriction	The value of a key can be up to 1 GB.	1 MB
Multiple databases	A single-node or master/standby DCS Redis instance supports up to 256 Redis databases. A cluster instance supports only one database, that is, DB0.	Not supported

Based on the preceding comparison, both the Redis and Memcached are easy to use and have high performance. However, Redis and Memcached are different in data structure storage, persistence, backup, migration, and script support. You are advised to select the most appropriate cache engine based on actual application scenarios.

Memcached is suitable for caching scenarios of small amount of static data, where data is only read without further computing and processing, for example, HTML code snippets.

Redis has richer data structures and wider application scenarios.

Parent topic: Service Overview

Previous topic: Comparing Redis Versions

Next topic: Comparing DCS and Open-Source Cache Services