Foreword:
Following are taken from source : db-engines.com . The pictures copyright belongs to me Alorya.com
I have writen it in the shortest way possible. To get more details visit db-engines.com
Types of databases and DBMS
A database is a structured set of data held in a computer or server.
Structured is the keyword here.
Structuring data makes it comprehensible/understandable,
DBMS: db management system
It is a software for creating and managing db. There are 300 systems representing 11 models of organising data.
There are X number of ways computers can store and interogate data. Each has unique strengths and weakness.There is no best database. All of them are good in one area.
DBMS power ranking:
- DBMS : slower in specialised today
- Document stores
- Key value stores
- Search engines
- Wide column store
- Graph DBMS: awesome for geospatial data
- Time series DBMS
- Native XML DBMS
- RDF Stores
- Object Oriented DBMS
- Multivalue stores
- Content Stores
- Event stores : unbeatable in the.. limited utlility
- Navigational DBMS
1) Ubiquous, used by everyone
Ex: MS Access and MySQL. 99% of apps and website use it.
a) RDBMS:Relational db management sys
RDBMS has gone further since its introduction in 1979:
- user-defined data types
- non-atomic attributes
- inheritence
- hierarchies
2) Common
source : db-engines.com
Ex: Mongo DB
a) Document stores
It is document oriented db sys. . Document stores (like spreadsheets) have no schema.
- records can be non-uniform
- records can have a nested structure
- values can be stored as arrays
It is a leading example to NoSQL database.
Ex: JSON : Javascript object name . It is a lightweight data interchange format JSON docs can be stored as a string in a key-value store or relational db (this is blurring the category /or this is exceptional)
b) Key-value stores
Ex: Redis
User: Stack Exchange
In Key-value stores, keys (of id`s) and values are stored in pairs. Each key has one and only one value. It is like RDBMS with
- one id attribute
- one other attribute
They are inadequate for complex function. But can be applied in embedded systems or as high performance in process db.
c) Search engines
Ex: Elastic search
User: Codebots, Vimeo
They are NoSQL DBMS designed for searching data only
Features:
- Complex search expression
- Full text search stemming
- Ranking and grouping of search result
- Geospatial search
- Distributed search
d)Wide column store (or extensive record store)
Ex: Cassandra
user: Netflix
It stores data in records of billions of dynamic columns. It has no Schema. But it stores data like RDBMS. It is a type of NoSQL db.
3)Big fish
source : db-engines.com
a)Graph DBMS
Ex: Neo4j, Ms Cosmos, Azure db
User: Ebay
It is based around structures of nodes and edges.
EDges define relationships between nodes. These db are designed to efficiently render visual representation of data. All graph DBMS are multi model.
These are all combined in one (= all unified in graph dbms):
- Document store
- Key value store
- Wide column store
- Graph DBMS
b) Time series DBMS
Ex: influx DB
User: Pipeline FX
It is optimised for handling time series data
c)Native XML DBMS
Ex: BaseX
Its internal data model corresponds to XML docs. It is a subclass of doc stores.
d) RDF stores
ex: Jena
It is a methodology for describing information. The RDF model represents info as triples in the form of:
a) Object Oriented DBMS
Ex: Db40
It is also known as object db. It is developed in the 80`s. It follows an object oriented data model with:
- classes
- properties
- methods
An object is always managed in its entirety. In recent years the classic RDBMS have been extended with some OO features:
- user defined data types
- structures attributes
b) Multivalue DBMS
ex: Adabas
Ir organise data as entities with attributes and records. Multivalue DBMS can assign more than one value withon a record`s attribute.
c) Content stores
ex: Jackrabbit
IT is also known as content repositories. It is specialised in managing digital content such as:
picture, text, video, metadata
d) Event stores
It implements the concept of event sourcing. This DB record all state changing events for an object and log these with a timestamp. In doing so event stores create a time series. Other DBMS store state, but event stores do not store state, it records history.
Many event store DB have performance issues necessitating snapshots.
5) Dying
source : db-engines.com
a) Navigational DBMS
In this DBMS, data is only accessible via linked records. This model was developed in 60`s. It was the first DBMS to handle large amount of data. There was subgroups, but these are superseded (replaced) by RDBMS.
No comments:
Post a Comment