When you want to invest in the commercial databases, you need to make sure that you settle with some of the best in the market. You should have the ability to know the performance and take time to choose the best.
CouchDB – Document Store
Maps keys to data
It provides a RESTful JSON API and is written in Erlang
You can upload functions to index data and then you can call those functions
Has a very simple REST interface
Provides an innovative replication strategy – nodes can reconnect, sync, and reconcile differences after being disconnected for long periods of time
Enables new distributed types of applications and data
MongoDB – Document Store
Free-form key-value-like data store with good performance
Powerful, expansive query model
Usability rivals that of Redis
Good for complex data storage needs.
Production-quality sharding capabilities
Neo4j – GraphDB
Disk-based
Has a restricted, single-threaded model for graph traversal
Has optional layers to expose Neo4j as an RDF store
Can handle graphs of several billion nodes, relationships, or properties on a single machine
Released under a dual license – free for non-commercial use
Apache Hbase – Wide Column Store/Column Families
Built on top of Hadoop, which has functionality similar to Google’s GFS and MapReduce systems
Hadoop’s HDFS provides a mechanism that reliably stores and organizes large amounts of data
Random access performance is on par with MySQL
Has a high performance Thrift gateway
Cascading source and sink modules
Redis – Key Value/Tuple Store
Provides a rich API and does more operations in memory, using disk only periodically.
It’s extremely fast
Lets you append a value to the end of a list of items that’s already been stored on a key.
Has atomic operations, making it a best-of-breed tally server.
Memcached – Key Value/Tuple Store
High-performance, distributed memory object caching
Free and open source
Generic and agnostic to the objects/strings it caches
It’s all in-memory data
Simple yet elegant design enables easy development and deployment
Language neutral caching scheme.
Most of the large properties on the web are using it now, except for Microsoft
Project Voldemort – Eventually Consistent Key Value Store
Used by LinkedIn
Handles server failure transparently
Pluggable serialization supports rich keys and values including lists and tuples with named fields
Supports common serialization frameworks including Protocol Buffers, Thrift, and Java Serialization
Data items are versioned
Supports pluggable data placement strategies
Memory caching and the storage system are combined
Tokyo Cabinet and Tokyo Tyrant – Key Value/Tuple Store
Supports hashtable mode, b-tree mode, and table mode
It’s fast and straightforward
Good for small to medium-sized amounts of data that require rapid updating and can be easily modeled in terms of keys and values
Cassandra – Wide Column Store/Column Families
First developed by Facebook
SuperColumns can turn simple key-value architecture into an architecture that handles sorted lists, based on an index specified by the user.
Can scale from one node to several thousand nodes clustered in different data centers.
Can be tuned for more consistency or availability
Smooth node replacement if one goes down