Skip to content
DoCodeHereDeveloper tools and services

Cloud, hosting and databases · Researched comparison

How to Pick In-Memory Databases: 7 Tools Compared

Aerospike leads the list, best for very large live data volumes. If your case is different, KeyDB suits Multithreaded Redis-compatible caching and Apache Ignite suits in-memory SQL and compute at scale. Prices, free plans and limits for all 7 tools follow.

By · Published 21 Nov 2026

We run no ads and no affiliate links at launch. The order of a list follows our published method.

Prices are the starting prices vendors list publicly. Plans change, so confirm the current price on the vendor page before you buy.

Key takeaways

ToolBest forStandout featureListed starting priceFree plan or trial
1 AerospikeVery large live data volumesNo charge per operationQuote onlyFree plan
2 KeyDBMultithreaded Redis-compatible cachingActive Replication and FLASH storage included$0 for self-hosted open sourceFree plan
3 Apache IgniteIn-memory SQL and compute at scaleKey-value and compute APIs in one system$0 for self-hosted open sourceFree plan
4 Dragonfly CloudHigh throughput Redis workloadsMultithreaded engine uses numerous cores on one node$36/month (Flex)Free trial
5 Redis CloudMost teams starting a managed cachePro layers on dedicated deployment and Active-Active replication$5/month for 250 MB (Essentials)Free plan
6 MemcachedStraightforward distributed caching at no license costVery light and straightforward to run$0 for self-hosted open sourceFree plan
7 Upstash RedisServerless apps with spiky trafficIdle usage-plan database costs $0$0.20 per 100K commands (pay-as-you-go)Free plan

How we picked

Selection started from tools commonly considered for in memory databases, then narrowed to 7 that have clear pricing information or a clear quote process. Seat minimums, billing terms and add-ons can move the real cost, so the starting prices here are a guide and not a quote. The ranking is editorial: typical fit first, then clear pricing, then free access. If your needs are unusual, read the profiles rather than relying on position.

01

Aerospike

Best for: Very large live data volumes

Aerospike is built with very large live data volumes in mind. Aerospike is a live NoSQL data product made for very large datasets and high transaction rates, with a hybrid memory design that keeps hot indexes in RAM and data on SSD. It charges by unique production data volume instead of by operations or by server count.

On top of that, choose Aerospike if you run very large live workloads and want predictable latency at scale. Enterprise brings features and support under a license based on data volume. Beyond that, one limit: since pricing follows unique data volume, a dataset with heavy duplication or quick growth can move the quote more than a request-count mo There is no fixed entry price for Aerospike; pricing comes through a quote. Free access exists with Aerospike.

Downsides: Production pricing is quote only.

Pros

  • No charge per operation
  • Hybrid memory design mixes RAM and SSD

Cons

  • Enterprise setup needs careful capacity planning

Price: Quote only, as listed by the vendor (Aerospike website). Confirm the current plan price before you buy.

Skip it if this limit matters to you: enterprise setup needs careful capacity planning.

Verdict. A sensible pick for very large live data volumes.

02

KeyDB

Best for: Multithreaded Redis-compatible caching

KeyDB is designed for Multithreaded Redis-compatible caching. KeyDB is a multithreaded fork of Redis that keeps full protocol and command compatibility, so existing Redis clients, scripts and cluster setups work without changes. Each connection gets a thread, which lifts throughput on multi-core machines for parallel workloads.

Choose KeyDB if you want Redis compatibility with more cores in use and multi-master replication at no license cost. KeyDB is free under a BSD license and you run it on your own servers, with no official managed cloud service from the project itself. That keeps the license cost at $0 but puts setup, monitoring, backup and failover on your team. KeyDB starts at: $0 for self-hosted open source. You can start on KeyDB without paying.

Points against: No official managed cloud service.

Pros

  • Active Replication and FLASH storage included
  • Multithreaded engine uses all CPU cores

Cons

  • Smaller community than Redis

Price: $0 for self-hosted open source, as listed by the vendor (KeyDB website). Confirm the current plan price before you buy.

Skip it if this limit matters to you: smaller community than Redis.

Verdict. A sensible pick for multithreaded Redis-compatible caching.

03

Apache Ignite

Best for: In-memory SQL and compute at scale

Apache Ignite makes the most sense for in-memory SQL and compute at scale. Apache Ignite is an in-memory computing system that works as a distributed database, a cache and a compute grid at the same time.

The main drawbacks: heavier to operate than a straightforward cache.

It keeps data in RAM across plenty of nodes, supports SQL through JDBC and ODBC drivers plus native APIs for Java, C#, C++ and Python, and runs ACID transactions across the cluster. Also, choose Ignite if your workload mixes SQL queries, transactions and compute in memory. What you pay to start with Apache Ignite: $0 for self-hosted open source. Apache Ignite has a free plan.

Pros

  • Key-value and compute APIs in one system
  • SQL

Cons

  • Enterprise features and support need a GridGain contract

Price: $0 for self-hosted open source, as listed by the vendor (Apache Ignite website). Confirm the current plan price before you buy.

Skip it if this limit matters to you: enterprise features and support need a GridGain contract.

Verdict. A sensible pick for in-memory SQL and compute at scale.

04

Dragonfly Cloud

Best for: High throughput Redis workloads

For high throughput Redis workloads, Dragonfly Cloud is a natural candidate. Dragonfly is a multithreaded in-memory datastore that stays compatible with Redis and Memcached clients, so most Redis code runs against it without changes. Its shared-nothing design spreads work across CPU cores on a single node.

Choose Dragonfly if one node must serve very high throughput. Worth adding: self-hosted Dragonfly is free up to 100 vCPUs. Dragonfly Cloud Flex opens at $36 per month for a completely managed service on AWS, GCP or Azure, with one-click provisioning, scaling and a monitoring dashboard. The starting price for Dragonfly Cloud: $36/month. Free plan details: Free trial.

Where it is weaker: smaller ecosystem than Redis Cloud.

Pros

  • Multithreaded engine uses numerous cores on one node
  • Flex begins at $36 per month

Cons

  • Business level begins at $2,000 per month

Price: $36/month (Flex), as listed by the vendor (Dragonfly Cloud website). Confirm the current plan price before you buy.

Skip it if this limit matters to you: business level begins at $2,000 per month.

Verdict. A sensible pick for high throughput Redis workloads.

05

Redis Cloud

Best for: Most teams starting a managed cache

Redis Cloud makes the most sense for most teams starting a managed cache. One limit: the level ladder jumps from 5 GB to 12 GB. So a dataset just over 5 GB pays for headroom it does not use.

The Free plan costs $0 and hands one 30 MB database on shared cloud, with 30 connections, 5 GB of monthly bandwidth and a 100 ops/sec ceiling. As for the details, Pro begins at $0.014 an hour with a $200 per month minimum, and it layers on dedicated deployment and Active-Active replication. A 1 GB database costs $18 per month. More specifically, essentials opens at $0.007 an hour, shown as $5 per month for 250 MB, with 256 connections, 100 GB of bandwidth and 1,000 ops/sec. Entry price for Redis Cloud: $5/month for 250 MB. Redis Cloud has a free plan.

The main drawbacks: Free level caps at 100 ops/sec and 30 connections.

Pros

  • Pro layers on dedicated deployment and Active-Active replication
  • Free level with 30 MB

Cons

  • No 6 GB Essentials level, so a 6 GB dataset needs the 12 GB plan

Price: $5/month for 250 MB (Essentials), as listed by the vendor (Redis Cloud website). Confirm the current plan price before you buy.

Skip it if this limit matters to you: no 6 GB Essentials level, so a 6 GB dataset needs the 12 GB plan.

Verdict. A sensible pick for most teams starting a managed cache.

06

Memcached

Best for: Straightforward distributed caching at no license cost

Memcached is built with straightforward distributed caching at no license cost in mind. The default memory allocation is 64 MB.

You run it on your own servers, and you add capacity by adding nodes. It is free and open source under a BSD license. The starting price for Memcached: $0 for self-hosted open source. A free plan is available with Memcached.

Drawbacks to consider: keys expire by LRU eviction when memory fills.

Pros

  • Very light and straightforward to run
  • Broad client support in every major language

Cons

  • No persistence and no built-in replication

Price: $0 for self-hosted open source, as listed by the vendor (Memcached website). Confirm the current plan price before you buy.

Skip it if this limit matters to you: no persistence and no built-in replication.

Verdict. A sensible pick for straightforward distributed caching at no license cost.

07

Upstash Redis

Best for: Serverless apps with spiky traffic

Upstash Redis is built with serverless apps with spiky traffic in mind. Upstash Redis is a serverless Redis-compatible database that you reach over TCP or HTTP, which fits serverless functions and edge runtimes that cannot hold a long TCP connection. The usage plan bills $0.20 per 100K commands, with the first 1 GB of storage free and then $0.25 per GB.

From a buyer's view, the Free level costs $0 and bundles 250 MB of data, 500K commands per month and 10 GB of monthly bandwidth. Also, choose Upstash if your traffic is uneven and you want a $0 idle cost. For everyday use, one limit: a Global database counts each replicated write as a command in every read region, so 100K writes with one read region cost $0.40 instead of Worth adding: fixed plans run from $10 per month for 250 MB to $1,500 per month for 500 GB, with no per-command billing. On cost, Upstash Redis starts like this: $0.20 per 100K commands. Upstash Redis has a free plan.

Honest limits: Prod Pack costs $200 per month per database.

Pros

  • Idle usage-plan database costs $0
  • Budget cap stops runaway bills

Cons

  • Replicated writes count again in each read region

Price: $0.20 per 100K commands (pay-as-you-go), as listed by the vendor (Upstash Redis website). Confirm the current plan price before you buy.

Skip it if this limit matters to you: replicated writes count again in each read region.

Verdict. A sensible pick for serverless apps with spiky traffic.

What to check before you buy

FAQ

Is Aerospike the right pick for everyone comparing in memory databases?

No single tool suits every team. Aerospike comes first because it fits very large real-time data volumes, while KeyDB suits Multithreaded Redis-compatible caching and Apache Ignite suits in-memory SQL and compute at scale. Match the tool to your own size, budget and workflow before deciding.

How much does Redis Cloud cost?

Redis Cloud Free costs $0 for a 30 MB database with 100 ops/sec and 30 connections. Essentials begins at $5 per month for 250 MB, a 1 GB database costs $18 per month. The 12 GB level costs $216 per month.

How are the in memory databases ordered?

Aerospike sits first and Upstash Redis last because the order follows typical fit for in memory databases, then price clarity and free access. It is a judgement call, so read each profile for fit and limits.

Which is a good match for very large live data?

The first pick on this list is Aerospike, because it suits very large live data volumes. KeyDB is better for Multithreaded Redis-compatible caching, and Apache Ignite for in-memory SQL and compute at scale. Check price and limits against your own needs before choosing.

Which in memory databases have a free plan?

6 of the 7 tools have a free plan, including Aerospike, KeyDB, Apache Ignite and Redis Cloud. Free plans usually cap users, projects or features, so check the limits on the vendor page. One other tool offers a trial instead.

Do these tools publish their prices?

Not all of them. Aerospike publishes no starting price, so expect a quote request for it. The rest, KeyDB included, show a starting price, though seats and billing terms still change the final cost.

Sources we checked

  1. Aerospike (official site)
  2. KeyDB (official site)
  3. Apache Ignite (official site)
  4. Dragonfly Cloud (official site)
  5. Redis Cloud (official site)
  6. Memcached (official site)
  7. Upstash Redis (official site)

Change log. First published 21 Nov 2026.