Middleware (3) NoSQL

The most common explanation for NoSQL is that non-relational, or not-only SQL, can be seen from the field meaning that it refers to the generic term of a non-relational database.

Background of the birth of NoSQL

With the development of the distributed architecture of large-scale Web sites, the use of traditional relational databases seems to be inadequate in the following ways:

1, high concurrent read and write to the database

With the popularity of distributed architecture and microservices, the database load is very high, if the query requirements of 1W+/S is still barely top, but for the 1w+/s write operation, the performance of the relational database is very low.

2, storage and query of massive data

When hundreds of billions of data are generated per day, the performance is very low for relational databases in a data table with billions of records.

3, high scalability and high availability

For relational databases, upgrades and extensions are difficult.

The emergence of NoSQL is the solution to the above problems.

Categories of NoSQL

There are four types of NoSQL: The Key/value database, the document database, the Columnstore database, and the graph database. The most common in the project is the Key/value database, represented by memcached and Redis, and the document database represented by MongoDB.

Key/value Database representative: memcached and Redis.

Document-based database representative: MongoDb.

The advantages of NoSQL

Since NoSQL is a problem in bridging relational databases, its advantages are obvious:

1. Handle massive amounts of data.

2, easy to expand, due to its non-relational nature.

3, high availability, such as Redis can form a cluster, providing 7 * 24-hour service.

So, does NoSQL have no drawbacks? No, for example, it does not have a relational data transaction.

Summarize

1,nosql can not replace relational database, can only be used to compensate for the lack of relational database, the core business is to use the relational database, due to its acid characteristics.

2, for the processing of big data, we can study the better Framework es (ElasticSearch).

Middleware (3) NoSQL