Whether the end of the relational database has come

Recently, a large number of new relational databases have sprung up in clouds outside the cloud. One of the key messages released here is: "If you want to get rich and on-demand scalability, you need a database that is not relational." ”

If this is true, then is this not a sign that the once powerful relational database finally cracks up in its armor? Is the date of the relational database over? Should Cain retire? In this article, we examine the current trend of getting rid of relational databases in specific situations and analyze what this means for the future of relational databases.

The relational database has been in the past. During this period, a brief outbreak of the so-called end of the relationship database Revolution. Ultimately, of course, they failed, and did not waver in the dominant position of the relational database.

Let's get some background

A relational database is basically a collection of tables (entities). A table is made up of Lie (a set of variables). These tables have constraints that define relationships with each other. A relational database uses SQL to query, and the result set is generated by a query that accesses one or more tables. Multiple tables that are accessed in a single query are "Connected" by using the paradigm defined in the Table relational column. Normalization is a data structure model used by relational database, which can ensure the consistency and eliminate data redundancy.

Relational databases are facilitated with the help of a relational database management system (RDBMS). Most of the database systems we use today are RDBMS, including Oracle,sql Server, Mysql,sybase,db2,teradata, and so on.

The reason why a relational database dominates is not trivial. They continue to provide the best combination of simplicity, robustness, flexibility, and performance, and bring common data management compatibility.

However, in order to achieve all this, the internal of the relational database has to be incredibly complex. For example, a relatively simple SELECT statement might have hundreds of execution paths that the optimizer must evaluate at run time. All this is hidden, invisible to the user, and the RDBMS uses a cost-like algorithm to determine the "execution plan" of the best response request.