Types of the databases: detailed overview of the DBMS types

Often, in the reviews of the database types, there is information about relational and “other” types, “NoSQL” etc. or the most basic types of DBMSs (databases), the rare ones are not mentioned t all. In this article we will try to describe all the types of the databases and give examples of the specific implementations. Of course, the article does not pretend to be all-embracing, databases can be classified in different ways, including types of optimal load, etc. Moreover, all the types of the databases will be described briefly. But we really hope that the article will give basic understanding of the DBMS types as well as of the principles of their operation.

In this article we are going to talk about the following types of the databases:

  • Relational
  • Key-value
  • Document-oriented
  • Time series databases
  • Graph Databases
  • Search Engines
  • Object-oriented databases
  • RDF (Resource Description Framework)
  • Wide Column Stores
  • Multimodal DBMS
  • Native XML DBMS
  • GEO/GIS (spatial) and specialized DBMSs
  • Event DBMS (state transition databases)
  • Content DBMSs
  • Navigational DBMSs
  • Vector databases

Let’s start with the most common type – relational DBMSs.

Relational DBs

The most well-known relational databases are Open Source projects PostgreSQL, MySQL and SQLite, as well as proprietary solutions such as Oracle, Microsoft SQL Server and IBM Db2Relational.

The essence of relational databases consists in data storage in the linked tables:

It is also worth saying that the relational databases come with row-based (PostgreSQL) and column-based (ClickHouse, Vertica) data storage. Column databases are better for analytics purposes, while row orientation database is better for transactional workloads.

Relational DBMSs are the most common type of the databases. The table with more than 150 relational variants is given below. The source of this list is a special site-agregator.

Full list of 166 relational DBMSs

  1. Oracle – relational, multimodal
  2. MySQL – relational, multimodal
  3. Microsoft SQL Server – relational, multimodal
  4. PostgreSQL – relational, multimodal
  5. IBM Db2Relational – multimodal
  6. Microsoft Access – relational
  7. SQLite – relational
  8. Snowflake – relational
  9. MariaDB – relational, multimodal
  10. Microsoft Azure SQL Database – relational, multimodal
  11. Hive – relational
  12. Databricks – multimodal
  13. Teradata – relational, multimodal
  14. Google BigQuery – relational
  15. FileMaker – relational
  16. SAP HANA – relational, multimodal
  17. SAP Adaptive Server – relational
  18. Microsoft Azure Synapse Analytics – relational
  19. Firebird – relational
  20. Informix – relational, multimodal
  21. Amazon Redshift – relational
  22. Impala – relational, multimodal
  23. Spark SQL – relational
  24. ClickHouse – relational, multimodal
  25. Netezza – relational
  26. Vertica – relational
  27. Presto – relational
  28. dBASE – relational
  29. Apache Flink – relational
  30. Greenplum – relational, multimodal
  31. Amazon Aurora – relational, multimodal
  32. H2 – relational, multimodal
  33. Oracle Essbase – relational
  34. Microsoft Azure Data Explorer – relational, multimodal
  35. Microsoft Azure Data Explorer – relational, multimodal
  36. CockroachDB – relational
  37. Derby – relational
  38. Interbase – relational
  39. Trino – relational, multimodal
  40. SingleStore – relational, multimodal
  41. SAP SQL Anywhere – relational
  42. Ingres – relational
  43. HyperSQL – relational
  44. Ignite – multimodal
  45. SAP IQ – relational
  46. Virtuoso – multimodal
  47. OpenEdge – relational
  48. Oracle NoSQL – multimodal
  49. Google Cloud Spanner – relational
  50. YugabyteDB – relational, multimodal
  51. MaxDB -relational
  52. TiDB – relational, multimodal
  53. Apache Druid – multimodal
  54. InterSystems Caché – multimodal
  55. InterSystems IRIS – multimodal
  56. DuckDB – relational
  57. SAP Advantage Database Server – relational
  58. HEAVY.AI – relational, multimodal
  59. 4D – relational
  60. Percona Server for MySQL – relational
  61. EDB Postgres – relational
  62. Apache Drill – multimodal
  63. EXASOL – relational
  64. Apache Phoenix – relational
  65. Citus – relational, multimodal
  66. Datomic – relational
  67. Empress – relational
  68. GridGain – multimodal
  69. OceanBase – relational, multimodal
  70. MonetDB – relational, multimodal
  71. VoltDB – relational
  72. Tibero – relational
  73. TimesTen – relational
  74. IBM Db2 warehouse -relational
  75. SQLBase – relational
  76. Firebolt – relational
  77. Fauna – multimodal
  78. mSQL – relational
  79. MatrixOne – relational
  80. DataEase – relational
  81. Oracle Rdb – relational
  82. Altibase – relational
  83. PlanetScale – relational, multimodal
  84. NonStop SQL – relational
  85. Cubrid – relational
  86. Infobright – relational
  87. Apache Kylin – relational
  88. GBase – relational
  89. Apache HAWQ – relational
  90. NuoDB – relational
  91. Dolt – relational, multimodal
  92. solidDB – relational
  93. FoundationDB – multimodal
  94. 1010data – relational
  95. openGauss – relational, multimodal
  96. HFSQL – relational
  97. Actian Vector – relational
  98. SQL.JS -relational
  99. OpenBase – relational
  100. Vitess – relational, multimodal
  101. Kognitio – relational
  102. StarRocks – relational
  103. TDSQL for MySQL – relational, multimodal
  104. DBISAM – relational
  105. FrontBase – relational
  106. TypeDB -multimodal
  107. NexusDB – relational
  108. Datacom/DB – relational
  109. Kinetica – relational, multimodal
  110. eXtremeDB – multimodal
  111. ScaleArc – relational
  112. VistaDB – relational
  113. Yellowbrick – relational
  114. Splice Machine – relational
  115. Postgres-XL – relational, multimodal
  116. Alibaba Cloud MaxCompute – relational
  117. AlaSQL – multimodal
  118. Apache Pinot – relational
  119. Alibaba Cloud AnalyticDB for MySQL – relational, multimodal
  120. SQream DB – relational
  121. Sequoiadb – multimodal
  122. Kingbase – relational, multimodal
  123. Trafodion – relational
  124. R:BASE – relational
  125. Apache Doris – relational
  126. Transbase – relational
  127. Lovefield – relational
  128. Raima Database Manager – multimodal
  129. Alibaba Cloud AnalyticDB for PostgreSQL – relational
  130. Tajo – relational
  131. Mimer SQL – relational
  132. Kyligence Enterprise – relational
  133. YDB – multimodal
  134. Databend – relational
  135. Actian PSQL – relational
  136. Alibaba Cloud ApsaraDB for PolarDB – relational
  137. Brytlyt – relational
  138. XtremeData – relational
  139. TransLattice – relational
  140. ElevateDB – relational
  141. Comdb2 – relational
  142. Linter – relational, multimodal
  143. AntDB – relational
  144. FeatureBase – relational
  145. LeanXcale – multimodel
  146. PipelineDB – relational
  147. GeoSpock – relational, multimodal
  148. Faircom DB – multimodal
  149. Tibco ComputeDB – relational
  150. Valentina Server – relational
  151. PieCloudDB -relational
  152. BigObject – relational
  153. Edge Intelligence – relational
  154. Fujitsu Enterprise Postgres – relational, multimodal
  155. EsgynDB – relational
  156. Transwarp KunDB – relational
  157. JethroData – relational
  158. MyScale – multimodal
  159. OushuDB – relational
  160. AgensGraph – multimodal
  161. SmallSQL – relational
  162. ActorDB – relational
  163. DaggerDB – relational
  164. EdgelessDB – relational
  165. K-DB – relational
  166. Sadas Engine – relational

Among the “domestic” players we can highlight the following:

  • Postgres Pro – PostgreSQL finalized for corporate tasks;
    • Jatoba – similar to variant mentioned above as it is based on PostgreSQL;
    • Quantum Hybrid – one more PostgreSQL variant;
    • Red – a DBMS based on Interbase/Firebird;
    • ProximaBD – based on PostgreSQL;
    • Arenadata DB – enterprise solution based on Greenplum;
    • YDB – serverless solution by the Yandex company. This is an Open Source product which is available for on-premise installations and as a managed service with dedicated/serverless consumption model.

Relational databases cover a wide range of tasks: from transactional databases to analytical ones, but they are not a “silver bullet” for all the tasks. Consider other types of the databases as well.

Key-value DBs

Key-value database type is designed to perform fast, almost instantaneous queries for such tasks as cache, balance display etc. The high speed is realized due to the data storage in accordance with the key-value principle, and in most cases it is effected by operations in RAM.

Dictionaries contain a collection of objects or records; objects contain many different fields, containing data (each of them). Records are stored and retrieved by using a key that uniquely identifies the record and is used for fast data retrieval.

The main purpose is to speed up data mapping for the end users and reduce loads, including I/O on organizations’ infrastructure.

The most well-known and widely used Key-Value solutions are Redis and Memcached.

The full list of 72 Key-Value DBMSs is given below:

  1. Redis – Key-value, multimodal
  2. Amazon DynamoDB – multimodal
  3. Microsoft Azure Cosmos DB – multimodal
  4. Memcached – Key-value
  5. etcd – Key-value
  6. Hazelcast – Key-value, multimodal
  7. Aerospike – multimodal
  8. Ehcache – Key-value
  9. Riak KV – Key-value
  10. Ignite – multimodal
  11. OrientDB – multimodal
  12. Google Cloud Bigtable – multimodal
  13. GemFire – Key-value, multimodal
  14. ArangoDB – multimodal
  15. Infinispan – Key-value
  16. Oracle NoSQL – multimodal
  17. RocksDB Key-value
  18. Oracle Berkeley DB – multimodal
  19. InterSystems Caché – multimodal
  20. InterSystems IRIS – multimodal
  21. LevelDB – Key-value
  22. LMDB – Key-value
  23. Geode – Key-value
  24. Amazon SimpleDB – Key-value
  25. Geode – Key-value
  26. Amazon SimpleDB – Key-value
  27. Oracle Coherence – Key-value
  28. GridGain – multimodal
  29. Tarantool – Key-value, multimodal
  30. GT.M – Key-value
  31. ZODB – Key-value
  32. FoudationDB – multimodal
  33. NCache – Key-value
  34. WebSphere eXtreme Scale – Key-value
  35. Hibari – Key-value
  36. MapDB – Key-value
  37. BoltDB – Key-value
  38. Graph Engine – multimodal
  39. Scalaris – Key-value
  40. KeyDB – Key-value
  41. Project Voldemort – Key-value
  42. Upscaledb – Key-value
  43. Cloudflare Workers KV – Key-value
  44. Elliptics –  Key-value
  45. Tokyo – Tyrant – Key-value
  46. LeanXcale – multimodal
  47. Immudb – Key-value, multimodal
  48. STSdb – Key-value
  49. TomP2P – Key-value
  50. ArcadeDB – multimodal
  51. Speedb – Key-value
  52. Faircom DB – multimodal
  53. Kyoto Tycoon – Key-value
  54. Skytable – Key-value
  55. HyperLevelDB – Key-value
  56. YTsaurus – multimodal
  57. InfinityDB – Key-value
  58. Tigris – multimodal
  59. Badger – Key – value
  60. Dragonfly – Key – value
  61. LedisDB – Key-value
  62. TerarkDB – Key-value
  63. Cachelot.io – Key-value
  64. ScaleOut StateServer – Key-value
  65. JaguarDB – Key-value
  66. Resin Cache – Key-value
  67. Faircom EDGE – multimodal
  68. SwayDB – Key-value
  69. BergDB – Key-value
  70. CortexDB – multimodal
  71. Helium – Key-value
  72. Tkrzw – Key-value

Among the domestic DBMSs firsr of all we should highlight Tarantool, distributed under the Simplified BSD license, it has a separate version for large corporate clients. Tarantool implements a hybrid data schema: key-value, document-oriented, relational and spatial.

Document-oriented DBs

If you need to store a lot of files or documents and you don’t want to think  about storage structure, hierarchy, relationships etc. we suggest choosing one of the document-oriented databases. The main advantage is the wide scalability.

Document-oriented databases are created for the hierarchical data structures (documents) storage. The basis of document-oriented DBMSs are document storages with a tree or forest structure. Trees start from the root node and may contain several internal and leaf nodes. Leaf nodes contain the data, which are entered into the indexes while adding a document; this makes it possible to find the path to the searched data even with a rather complex structure. Unlike the key-value type storages, the selection on the request to the document storage can contain parts of a large number of documents without full loading of these documents into the RAM.

The most popular document-oriented database is MongoDB.

Full list of 56 document-oriented database solutions is given below:

  1. MongoDB – document-oriented, multimodal
  2. Amazon DynamoDB – multimodal
  3. Databricks – multimodal
  4. Microsoft Azure Cosmos DB – multimodal
  5. Couchbase – document-oriented, multimodal
  6. Firebase Realtime Database – document-oriented
  7. CouchDB – document-oriented, multimodal
  8. Google Cloud Firestore – document-oriented
  9. MarkLogic – multimodal
  10. Realm – document-oriented
  11. Aerospike – multimodal
  12. Google Cloud Datastore – document-oriented
  13. Virtuoso – multimodal
  14. OrientDB – multimodal
  15. ArangoDB – multimodal
  16. RavenDB – document-oriented, multimodal
  17. Oracle NoSQL – multimodal
  18. IBM Cloudant – document-oriented
  19. RethinkDB – document-oriented, multimodal
  20. InterSystems IRIS – multimodal
  21. PouchDB – document-oriented
  22. CloudKit – document-oriented
  23. Apache Drill – multimodal
  24. Amazon DocumentDB – document-oriented
  25. Mnesia – document-oriented
  26. LiteDB – document-oriented
  27. Fauna – multimodal
  28. Datameer – document-oriented
  29. GigaSpaces – multimodal
  30. FoundationDB – multimodal
  31. AllegroGraph – multimodal
  32. HPE Ezmeral Data Fabric – multimodal
  33. CrateDB – multimodal
  34. LokiJS – document-oriented
  35. BigchainDB – document-oriented
  36. AlaSQL – multimodal
  37. SurrealDB – multimodal
  38. Sequoiadb – multimodal
  39. Percona Server for MongoDB – document-oriented
  40. HarperDB – document-oriented
  41. EJDB – document-oriented
  42. YDB – multimodal
  43. ArcadeDB – multimodal
  44. Bangdb – multimodal
  45. XTDB – document-oriented
  46. YTsaurus – multimodal
  47. OrigoDB – multimodal
  48. WhiteDB – document-oriented
  49. ToroDB – document-oriented
  50. SenseiDB – document-oriented
  51. Acebase – document-oriented
  52. iBoxDB – document-oriented
  53. RaptorDB – document-oriented
  54. NosDB – document-oriented
  55. CortexDB – multimodal
  56. JasDB – document-oriented

Among the domestic solutions, the Yenisei DBMS can be referred to the document-oriented databases.

And within one of the modalities, YDB and YTsaurus can be referred to such databases as well.

Time-series DBs

If you have time-ordered or time-stamped data, such as metrics from infrastructure or sensor data, it may be reasonable to use one of the time series databases.

General characteristics of the time-series databases:

  • Time-series data are always collected over a period of time;
    • Data from workloads are new and are recorded as inserts. Existing data are not updated by replacement;
    • When data are written, they are automatically assigned to the last time interval.

Time-series databases are often used to monitor various metrics (CPU utilization, sensor performance etc.).

The most popular time-series databases are Prometheus, InflubDB, Graphite.

Full list of 43 time-series DBMSs  is given below:

1.InfluxDB Time Series, multimodal
2.Kdb multimodal
3.PrometheusTime Series
4.GraphiteTime Series
5.TimescaleDBTime Series, multimodal
6.DolphinDBTime Series, multimodal
7.RRDtoolTime Series
8.Apache Druidmultimodal
9.TDengine Time Series, multimodal
10.QuestDB Time Series, multimodal
11.OpenTSDBTime Series
12.GridDB Time Series, multimodal
14.VictoriaMetrics Time Series
15.Amazon TimestreamTime Series
16.M3DB Time Series
17.HeroicTime Series
18.eXtremeDB multimodal
19.CrateDB multimodal
20.Apache IoTDBTime Series
21.KairosDBTime Series
22.ITTIATime Series,multimodal
23.Raima Database Manager multimodal
24.AxibaseTime Series
25.Riak TSTime Series
26.MachbaseTime Series
27.CnosDBTime Series
30.IRONdbTime Series
31.QuasardbTime Series
32.SiteWhereTime Series
33.NSDbTime Series
34.Alibaba Cloud TSDBTime Series
35.IBM Db2 Event Storemultimodal
36.Tigris multimodal
37.GreptimeDB Time Series
38.Hawkular MetricsTime Series
39.SiriDBTime Series
40.BluefloodTime Series
41.Warp 10Time Series
42.YanzaTime Series
43.NewtsTime Series

Graph DBs

If you need to analyze data relationships, their connections or just simplify queries with Join, it makes sense to use graph databases.

Data and their relationships are represented as vertices and edges of the graph, respectively.

Thus you can easily represent money transfers (to identify various fraudulent schemes), social network connections and the graph of communication between mobile network operators.

Graph databases are usually used in the financial sphere in order to detect various fraudulent schemes, but they also will be convenient for any other task where it is necessary to work with relations of the objects.

The most well-known Open Source solution is Neo4J, but there are a number of other high-quality alternatives.

Full list of 39 graph databases is given below:

1.Neo4j graph
2.Microsoft Azure Cosmos DB multimodal
3.Virtuoso multimodal
5 .ArangoDB multimodal
6.Memgraph graph
7.GraphDB multimodal
8.Amazon Neptunemultimodal
10.Stardog multimodal
11.NebulaGraph graph
16.AllegroGraph multimodal
18.TypeDB multimodal
19.Graph Enginemultimodal
25.AnzoGraph DBmultimodal
34.TerminusDBgraph, multimodal
39.Transwarp StellarDBgraph

Search Engines

If you need to search large amounts of data, especially unstructured data, it will be very useful to work with  a database that would combine the functionality of text search with the functionality of information storage.

Let’s imagine that you have N petabytes of logs (or other text data). Ordinary word search is no longer suitable in this case.

Indexing is a great solution. If we consider it in a very exaggerated way, we can visualize it as follows: we assign an index to each word/term/n-gram and record these indices in a special table, where the rows are the document and the columns are the indices. A similar system is used for plagiarism search, but in this case shingles (indexes with layering) are used.

It is much faster to search by index than by matching words in the documents.

Of course, modern search engines offer wider functionality.

The most popular solutions of this kind are Elasticsearch (and its version – OpenSearch, Splunk (I wrote about it in one of the previous articles) and Sphinx.

Full list of search engines is given below:

1.ElasticsearchSearch engine, multimodal
2.SplunkSearch engine
3.SolrSearch engine, multimodal
4.OpenSearch Search engine, multimodal
6.AlgoliaSearch engine
7.Microsoft Azure SearchSearch engine
8.SphinxSearch engine
9.Virtuoso multimodal
10.ArangoDB multimodal
11.CoveoSearch engine
12.Amazon CloudSearchSearch engine
13.MeilisearchSearch engine
14.XapianSearch engine
15.SearchBloxSearch engine
16.TypesenseSearch engine
18.MarqoSearch engine
19.Alibaba Cloud Log Service Search engine
20.Manticore SearchSearch engine, multimodal
21.ExorbyteSearch engine
22. Tigris мультимодальная 
23.IndicaSearch engine
24.RizhiyiSearch engine, multimodal

Object-oriented DBs

Object-oriented databases are databases where information is represented as objects, just like in object-oriented programming languages.

Object-oriented databases have appeared as a way of a code native communication which is written using object-oriented languages with a database.

Object-oriented databases have the following advantages:

  • There is no problem of data model mismatch between the data model in the database and the application because the database stores the data in the same form;
    • There is no need to maintain the data model on the database side separately;
    • All the objects at the source level are strictly typed.

The best known object-oriented DBMS is Db4o.

Full list of these DBs is given below:

  1. InterSystems Caché – multimodal
  2. Db4o – object-oriented
  3. Actian NoSQL Database – object-oriented
  4. ObjectStore – object-oriented
  5. Matisse – object-oriented
  6. Perst – object-oriented
  7. GigaSpaces – multimodal
  8. ObjectBox – object-oriented, multimodal
  9. ObjectDB – object-oriented
  10. atoti – object-oriented
  11. GemStone/S – object-oriented
  12. Objectivity/DB – object-oriented
  13. Starcounter – object-oriented
  14. Jade – object-oriented
  15. Actian Fast – object-oriented
  16. Eloquera – object-oriented
  17. Siaqodb – object-oriented
  18. OrigoDB – multimodal
  19. DataFS – object-oriented, multimodal
  20. WakandaDB – object-oriented
  21. VelocityDB – multimodal

RDF (Resource Description Framework)

RDF databases are partly similar to graph databases. Their main function is based on the concept of formulating resource-related statements as subject-predicate-object expressions. The subject denotes the resource and the predicate denotes the features of the resource and defines the relationship between the subject and the object.

Here’s what Wikipedia says about RDF:

Resource Description Framework (RDF, “resource description framework”) is a model developed by the World Wide Web Consortium for representing data, especially metadata. RDF represents statements about resources in a form suitable for machine processing. RDF is part of the semantic web concept.

A resource in RDF can be any entity, whether informational (e.g. a website or an image) or non-informational (e.g. a person, a city or some abstract concept). A statement made about a resource has the form “subject – predicate – object” and is called a triplet. The statement “the sky is blue” in RDF terminology can be represented as follows: subject – “the sky”, predicate – “has a colour”, object – “blue”. URIs are used to denote subjects, relations and objects in RDF.

A set of RDF statements forms an oriented graph in which the vertices are subjects and objects and the edges represent relations.

RDF itself is not a file format, but an abstract data model, i.e. it describes the proposed structure, processing and interpretation of data. A number of record formats exist for storing and transmitting information stacked in the RDF model.

For processing RDF-data it is proposed to implement query languages such as SPARQL (W3C standard), RQL, RDQL.”

Full list of the DBs is given below:

  1. MarkLogic – multimodal
  2. Apache Jena TDB – RDF
  3. Virtuoso – multimodal
  4. GraphDB – multimodal
  5. Amazon Neptune – multimodal
  6. Stardog – multimodal
  7. AllegroGraph – multimodal
  8. Blazegraph – multimodal
  9. RDF4J  – RDF
  10. Redland – RDF
  11. Strabon – RDF
  12. 4store – RDF
  13. Mulgara – RDF
  14. CubicWeb – RDF
  15. RedStore – RDF
  16. AnzoGraph DB – multimodal
  17. BrightstarDB – RDF
  18. Dydra – RDF
  19. RDFox – multimodal
  20. SparkleDB RDF

Wide Column Stores

Wide Column Stores is a NoSQL database that stores data in flexible columns that can be distributed across multiple servers or database nodes, using multidimensional mapping to reference data by column, row and timestamps.

The advantages of Wide Column Stores databases include query speed, scalability and a flexible data model.

The main difference in comparison with relational databases is the following:

A relational database management system stores data in a table with rows that span multiple columns. If one row requires an additional column, that column must be added to the entire table with NULL or default values for all other rows. If you need to query this DBMS table for a value that is not indexed, scanning the table to find these values will be extremely slow.

Wide Column DBMSs have the concept of rows like relational databases, but reading or writing a row of data consists of reading or writing individual columns. A column is written only if there is a data element for it. Each data item can be referenced by a row key, the value query is optimized in the same way as an index query in a DBMS.

The first column model is an entity/attribute/value table. Inside each entity (columns) there is a table of values/attributes.

The Wide Column DBMS concept allows building solutions for processing really large amounts of data.

The most well-known implementations are Cassandra and HBase DBMSs.

Full list is given below:

  1. Cassandra – Wide column, multimodal
  2. HBase – Wide column
  3. Microsoft Azure Cosmos DB – Multimodal
  4. Datastax Enterprise – Wide column, multimodal
  5. ScyllaDB – Wide column, multimodal
  6. Microsoft Azure Table Storage – Wide column
  7. Google Cloud Bigtable – Multimodal
  8. HPE Ezmeral Data Fabric – multimodal
  9. Amazon Keyspaces – Wide column
  10. Elassandra – Wide column, multimodal
  11. Alibaba Cloud Table Store – Wide column
  12. SWC-DB – Wide column, multimodal
  13. Accumulo – Wide column

Multimodal DBMSs

It is worth saying that each multimodal DBMS can ultimately be reduced to one or more of the types on which they are based. Namely, it can be based on a relational model or on a document-oriented model etc. But they also allow emulating additional types in terms of user interaction with them.

Full list of the DBMSs is given below:

  1. Adabas – multimodal
  2. UniData – multimodal
  3. jBASE – multimodal
  4. Northgate Reality – multimodal
  5. Model 204 – multimodal
  6. D3 – multimodal
  7. SciDB – multimodal
  8. OpenInsight – multimodal
  9. Rasdaman – multimodal
  10. OpenQM – multimodal

Native XML DBMSs

The Native XML Database type is based on the use of internal XML representation in contrast to XML add-ons over existing relational databases (XML enabled DB), which implement XML-SQL according to the SQL-2003 access standard.

Native XML DBMSs are designed to manage a large number of XML documents.

XML native DB (NXD) uses an internal representation  of the DOM XML in the database and has the following features:

  • The minimal model includes: elements, attributes, PCDATA sections and document list;
    • An XML document is represented as a fundamental part of the repository (like a table in a relational database);
    • To access the XML native DB information store the XQuery query language should be used.

Full list of the Native XML DBMSs is given below:

  1. MarkLogic – multimodal
  2. Virtuoso – multimodal
  3. Oracle Berkeley DB – multimodal
  4. BaseX – Native XML
  5. Sedna – Native XML
  6. eXist-db – Native XML
  7. searchxml – multimodal

GEO/GIS and specialized DBMSs

If you are working at a map service or a specialized application that uses location data, it is logical to use a specialized DBMS.

Unlike other types of databases designed for processing numeric and symbolic information, spatial databases have the ability to work with complete spatial objects, combining both traditional and geometric types of data. Spatial databases allow us to perform analytical queries containing spatial operators for analyzing spatial and logical relations of objects (“intersects…”, “touches…”, “is contained in…”, “contains…”, “is at a given distance from…”, “coincides…” and others).

Full list of the DBMSs is given below:

  1. PostGIS – Spatial DBMS, multimodal
  2. Aerospike – multimodal
  3. SpatiaLite – Spatial DBMS, multimodal
  4. GeoMesa – Spatial DBMS
  5. H2GIS – Spatial DBMS, multimodal
  6. SpaceTime – Spatial DBMS, multimodal

Event DBMSs

Let’s imagine that the data in your DBMS has changed and you want to understand the event that changed it as well as to view the history of changes and what led to it.

Event databases store not only the data itself, but also the flow/sequence of its changes and the events that caused these changes.

This format can be useful both for incident investigation and for researching business and information processes occurring in an enterprise.

Here is a list of well-known event databases:

  1. EventStoreDB – Event
  2. NEventStore – Event
  3. IBM Db2 Event Store – multimodal

Content DBMSs

Content databases are the implementations of content repositories. They can be used to store and manage different content (photos, texts, code repositories, etc.).

Here are some examples of content databases:

  1. Jackrabbit – content
  2. ModeShape content

Navigational DBMSs

Warning! Navigational databases have nothing to do with spatial databases. A navigational database is a type of database in which records or objects are found primarily by links from other objects. In some ways, it is close to the concept of graph databases.

A navigational database is a combination of a hierarchical and network model of database interfaces. Navigation methods use “pointers” and “paths” to navigate among the data records. The opposite model is the relational model, which uses “declarative” methods in which you ask the system what you want rather than how to get to it.

Here are some examples of navigational databases:

  1. IMS – navigational
  2. IDMS – navigational

Vector DBs

If you are engaged in machine learning, the main thing you have to work with is vectors. These can be rows from a table corresponding to the features of an object, or an embedding obtained at the output of a neural network that characterizes the processed object and is subsequently used for face identification or text translation.

Vector databases have been developed especially for this kind of tasks, allowing to work directly with vectors represented as numerical arrays.

The main advantage of such kind of the databases is the speed of search and processing of vectors.

Vector databases are a relatively new type of DBMS, it is difficult to name a leader in this category.

List of the vector DBMSs:

  1. Kdb – multimodal
  2. Pinecone – vector
  3. Chroma – vector
  4. Milvus – vector
  5. Weaviate – vector
  6. Vald – vector
  7. Qdrant – vector
  8. Deep Lake – vector
  9. Vespa – multimodal
  10. MyScale – multimodal

Final thoughts

In this article we have looked through the different types of database management systems, starting from the most popular types, such as relational databases, and ending by the very exotic ones, such as navigational databases. Each type has its own specialization and purposes. The main purpose of the article was to give you a general idea of the variety of the existing DBMS, more details will be given later.