非关系数据库（Non relational database）.docVIP

非关系数据库（Non relational database） In fact, Mysql also has a K-V engine, the efficiency is not low, before a large network company is discussed and the Nosql, they are now used to place: big data statistics, they do not have Myql KV, instead of Cassandra, which can cause memory after the JAVA object serialization, mainly Mysql serialization of objects only Text, BloB, and they think that the Mysql report frame is too bulky to generate a lot of based on the middle table cache, etc.. Cassandra may be less efficient, but it is very convenient to use. With the introduction of Robbin: With the rise of the Internet Web2.0 website, non relational database has become an extremely popular in new areas, the development of non relational database products very quickly. The traditional relational database in Web2.0 website, especially the large scale and dynamic website Web2.0 pure SNS type high concurrency has appeared to be inadequate, exposed a lot of problems difficult to solve, for example: 1, High performance - the database needs high concurrent read and write The Web2.0 web site to generate real-time dynamic pages and provides dynamic information according to the users personalized information, so basically can not use static and dynamic pages, so database concurrent load is very high, often have to read and write requests to tens of thousands of times per second. A relational database with tens of thousands of times the SQL query also managed to withstand, but with tens of thousands of times the SQL request to write data to the hard disk IO is unbearable. In fact, for ordinary BBS sites, there are also high demand for concurrent write requests, such as JavaEye site real-time statistics for online users, recording popular post clicks, vote counting, so this is a very common demand. 2, Huge - Storage of the high efficiency of mass data storage and access needs Similar to Facebook, twitter, Friendfeed such SNS website, users every day dynamic user generated mass, in the case