DBMS > HEAVY.AI vs. Oracle Berkeley DB vs. RocksDB vs. Sphinx vs. Splice Machine

System Properties Comparison HEAVY.AI vs. Oracle Berkeley DB vs. RocksDB vs. Sphinx vs. Splice Machine

Editorial information provided by DB-Engines

Name

Description

A high performance, column-oriented RDBMS, specifically developed to harness the massive parallelism of modern CPU and GPU hardware

Widely used in-process key-value store

Embeddable persistent key-value store optimized for fast storage (flash and RAM)

Open source search engine for searching in data from different sources, e.g. relational databases

Open-Source SQL RDBMS for Operational and Analytical use cases with native Machine Learning, powered by Hadoop and Spark

Primary database model

Relational DBMS

Key-value store

supports sorted and unsorted key sets
Native XML DBMS

in the Oracle Berkeley DB XML version

Key-value store

Search engine

Relational DBMS

Secondary database models

Spatial DBMS

DB-Engines Ranking measures the popularity of database management systems
Trend Chart

Score	1.64
Rank	#145	Overall
	#67	Relational DBMS

Score	2.01
Rank	#126	Overall
	#21	Key-value stores
	#3	Native XML DBMS

Score	3.41
Rank	#86	Overall
	#11	Key-value stores

Score	5.95
Rank	#55	Overall
	#5	Search engines

Score	0.54
Rank	#252	Overall
	#115	Relational DBMS

Website

github.com/heavyai/heavydb
www.heavy.ai

www.oracle.com/database/technologies/related/berkeleydb.html

rocksdb.org

sphinxsearch.com

splicemachine.com

Technical documentation

docs.heavy.ai

docs.oracle.com/cd/E17076_05/html/index.html

github.com/facebook/rocksdb/wiki

sphinxsearch.com/docs

splicemachine.com/how-it-works

Developer

HEAVY.AI, Inc.

Oracle

originally developed by Sleepycat, which was acquired by Oracle

Facebook, Inc.

Sphinx Technologies Inc.

Splice Machine

Initial release

2016

1994

2013

2001

2014

Current release

5.10, January 2022

18.1.40, May 2020

9.2.1, May 2024

3.5.1, February 2023

3.1, March 2021

License

Commercial or Open Source

Open Source

Apache Version 2; enterprise edition available

Open Source

commercial license available

Open Source

BSD

Open Source

GPL version 2, commercial licence available

Open Source

AGPL 3.0, commercial license available

Cloud-based only

Only available as a cloud service

DBaaS offerings (sponsored links)

Database as a Service

Providers of DBaaS offerings, please contact us to be listed.

Implementation language

C++ and CUDA

C, Java, C++ (depending on the Berkeley DB edition)

C++

Java

Server operating systems

Linux

AIX
Android
FreeBSD
iOS
Linux
OS X
Solaris
VxWorks
Windows

Linux

FreeBSD
Linux
NetBSD
OS X
Solaris
Windows

Linux
OS X
Solaris
Windows

Data scheme

yes

schema-free

yes

Typing

predefined data types such as float or date

yes

XML support

Some form of processing data in XML format, e.g. support for XML data structures, and/or support for XPath, XQuery or XSLT.

yes

only with the Berkeley DB XML edition

Secondary indexes

yes

full-text index on all search fields

yes

SQL

Support of SQL

yes

SQL interfaced based on SQLite is available

SQL-like query language (SphinxQL)

yes

APIs and other access methods

JDBC
ODBC
Thrift
Vega

C++ API
Java API

Proprietary protocol

JDBC
Native Spark Datasource
ODBC

Supported programming languages

All languages supporting JDBC/ODBC/Thrift
Python

.Net

Figaro is a .Net framework assembly that extends Berkeley DB XML into an embeddable database engine for .NET
others

Third-party libraries to manipulate Berkeley DB files are available for many languages
C
C#
C++
Java
JavaScript (Node.js)

3rd party binding
Perl
Python
Tcl

C
C++
Go
Java
Perl
Python
Ruby

C++

unofficial client library
Java
Perl

unofficial client library
PHP
Python
Ruby

unofficial client library

C#
C++
Java
JavaScript (Node.js)
Python
R
Scala

Server-side scripts

Stored procedures

yes

Java

Triggers

yes

only for the SQL API

yes

Partitioning methods

Methods for storing different data on different nodes

Sharding

Round robin

none

horizontal partitioning

Sharding

Partitioning is done manually, search queries against distributed index is supported

Shared Nothhing Auto-Sharding, Columnar Partitioning

Replication methods

Methods for redundantly storing data on multiple nodes

Multi-source replication

Source-replica replication

yes

none

Multi-source replication
Source-replica replication

MapReduce

Offers an API for user-defined Map/Reduce methods

Yes, via Full Spark Integration

Consistency concepts

Methods to ensure consistency in a distributed system

Immediate Consistency

Foreign keys

Referential integrity

yes

Transaction concepts

Support to ensure data integrity after non-atomic manipulations of data

ACID

yes

ACID

Concurrency

Support for concurrent manipulation of data

yes

yes, multi-version concurrency control (MVCC)

Durability

Support for making data persistent

yes

The original contents of fields are not stored in the Sphinx index.

yes

In-memory capabilities

Is there an option to define some or all structures to be held in-memory only.

yes

User concepts

Access control

fine grained access rights according to SQL-standard

Access rights for users, groups and roles according to SQL-standard

More information provided by the system vendor

We invite representatives of system vendors to contact us for updating and extending the system information,
and for displaying vendor-provided information such as key customers, competitive advantages and market metrics.