NoSQL ( Not Only SQL for English ) refers to databases that track a non-relational approach, thus breaking with the long history of relational databases. These data memory require no fixed table schemas, and try to avoid joins, they scale it horizontally. In the academic environment, they are often referred to as " structured data storage " means ( engl. structured storage ).
The term NoSQL (in the sense of no SQL) was first used for a 1998's light-weight open-source database, which did not provide SQL access option. Carlo Strozzi, the developer of this database, but the NoSQL database from the NoSQL movement is different in that the former is a database, which eliminates the use of the SQL language, while the latter more is a concept that takes distance from the relational model.
The term NoSQL (in the sense of emergency only SQL) was introduced for a meeting on distributed structured data store in early 2009 by Johan Oskarsson. The name was an attempt of a common term determination for the growing number of non-relational, distributed data storage systems, which usually waived on ACID properties.
This topic is not entirely new. The effort to save data without the limitations of the relational model, was previously known under the title documents oriented database. In this respect, all the representatives of this subject are to be regarded as NoSQL systems.
Although the popularity of NoSQL systems continue to increase, the market is still clearly dominated by relational systems.
Relational databases usually suffer from performance problems for data-intensive applications such as indexing of large quantities of documents, web pages with high load traffic, and streaming media applications. Relational databases are efficient only if they are optimized for frequent but small transactions or for large batch transactions with rare write accesses. You can hardly be expected to deal with simultaneously high data requirements and frequent data changes. NoSQL on the other hand can deal with many read and write requests. Examples can be found at Digg, Facebook, and eBay.
NoSQL architectures usually offer only weak guarantees regarding consistency such as eventual consistency or to individual records restricted transactions. Some systems also support ACID, for example, by the addition of special middleware such as CloudTPS.
Many NoSQL implementations support distributed databases with redundant data on many servers, for example by using a distributed hash table. Thus, the systems can easily scale and survive failures of single server.
NoSQL implementations can be broken down as follows: