Shared Nothing Architecture

from Wikipedia, the free encyclopedia
QS IT
This article was due to content flaws on the quality assurance side of the computer science editorial added. This is done in order to bring the quality of the articles from the subject area of ​​computer science to an acceptable level. Help to eliminate the shortcomings in this article and take part in the discussion !  ( + )


Reason: Please further wikify, add individual evidence and make it easier. Graphics would be nice. - 141.62.45.66 16:42, Apr 5, 2012 (CEST)

The shared nothing architecture (or shared nothing , abbr.SN ) describes an architecture of a distributed system in which each node can independently and independently perform its tasks with its own processor and the assigned storage components such as hard disk and main memory and no specific, single node is necessary for the connection to a database. The nodes are connected to one another via a LAN or WAN network . Each node also has a copy of the database management system ; the node can delegate tasks to another underutilized node.

Shared nothing systems are often understood in contrast to systems that store a large number of status information centrally, such as database or application servers . Even if the term shared nothing only became known to a larger group in connection with the further development of the World Wide Web , the concept had been used well before that. Michael Stonebraker mentioned the term in a 1986 publication on databases at the University of California at Berkeley , but it is entirely possible that the term was coined earlier.

Because of its scalability, shared nothing is popular for web applications or parallel database systems . As was shown by Google , a shared-nothing system can be expanded almost indefinitely by adding additional nodes in the form of inexpensive computers, because there is no single network element whose limited performance reduces the speed of the entire system.

A shared nothing system can divide its data across many different nodes, for example by assigning certain nodes to certain users or to certain web queries, or it can require each node to keep its own copy of the application data, which however requires the use of a Makes coordination protocol / procedure necessary between the nodes.

There is some discussion about whether a web application can be run with different, independent network computers, but with a single, central database, e.g. B. in the form of a computer cluster should be viewed as a shared nothing system.

Other architectures with a different design are:

See also

credentials

  1. Michael Stonebraker (UC Berkeley), The Case for Shared Nothing Architecture (PDF, 19 kB). Original publication in Database Engineering , Volume 9, Number 1 (1986) (English)