NoSQL

In recent years, a number of data storage systems have been developed with excell ent horizontal scaling properties. The scalability is the ability of a system, n etwork, or process to handle a growing amount of work in a capable manner or its ability to be enlarged to accommodate that growth. There are two dimensions to scalability, the first is horizontal scaling whichsupports higher transaction th roughput by splitting the database and associated load over many networked serve rswith performance improving approximately linearly with the number of machines, and the second is vertical scaling which supports higher transaction throughput by more effectively using a single server with a multi-core CPU and lots of mem ory.The highest scalability is achieved by combining these two dimensions, distr ibuting the load over many servers, and using each server very effectively.Our m ain focus is on horizontal scalability, because the number of cores that can sha re memory is limited, and horizontal scaling generally proves less expensive, us ing commodity servers. The traditional database products have comparatively little or no ability to sca le horizontally on applications. Therefore, a new data storage system was develo ped called "NoSQL" data stores which have following six key features of NoSQL ar e: a) Scale horizontally ªsimple operationsº (key lookups, reads and writes of one record or a small numberof records, simple selections). b) Replicate/distribute(partition) data over many servers. c) Simple call level interface (contrast with SQL). d) Weaker concurrency model than ACID. e) Efficient use of distributed indexes and RAM. f) Flexible schema (different records have different schema). Due to theshared-nothing horizontal scaling feature of NoSQL systems,(replicatin g and partitioning data over many servers)it is able to support a large number o f simple read/write operations per second which was failed in traditional RDMS d atabases. This simple operation load is traditionally called OLTP (on-line trans action processing). The NoSQL range in functionality from the simplest distribu ted hashing, as supported by the popular memcached open source cache, to highly scalable partitioned tables, as supported by Google's BigTable. Scalable data stores can be categorized into four groups: · Key-value stores, (including Voldemort, Riak, Redis, Membase, and Dynamo). · Document stores, (including CouchDB, MongoDB, Terrastore, and SimpleDB). · Extensible record stores, (including BigTable, HBase, HyperTable, and Cassandra) . · Scalable RDBMSs, (including MySQLCluster, VoltDB, Clustrix, and ScaleDB). An alternative, more cloud-friendly approach is to employ NoSQL. The load is able to easily grow by distributing itself over lots of ordinary, an d cheap, Intel-based servers. A NoSQL database is exactly the type of database t hat can handle the sort of unstructured, messy and unpredictable data that our s ystem of engagement requires. NoSQL is a whole new way of thinking about a database. NoSQL is not a relational database. The reality is that a relational database model may not be the best s olution for all situations. The easiest way to think of NoSQL, is that of a data base which does not adhering to the traditional relational database management s ystem (RDMS) structure. Sometimes you will also see it revered to as  not only S QL . The NoSQL taxonomy supports key-value stores, document store, BigTable, and grap h databases.