NoSQL mongodb - 01 introduction to NoSQL and mongodb

NoSQL brief introduction

Reference resources ：NoSQL brief introduction

Problems with relational databases

2008 About years ago , Website 、 Forum 、 Social networks have started to develop at a high speed , Traditional relational database It is a great challenge to store and process data , It is mainly reflected in the following points ：

• It is difficult to cope with tens of thousands of highly concurrent data writes per second .

• The speed of querying hundreds of millions of data is slow .

• sub-treasury 、 It is difficult to further expand the sub database formed by tables after reaching a certain scale .

• sub-treasury 、 The rules for splitting tables may change due to changes in requirements .

• It is difficult to modify the table structure .

In many Internet application scenarios , The query demand for the data join table is not so strong , It does not need to be read immediately after the data is written , But there are very high requirements for the speed of data reading and concurrent writing .

In the era of data explosion , If you want to use the traditional relational database to meet the high concurrency of data reading and writing 、 Storage of huge amounts of data 、 Database expansion and high availability , You need to increase the specifications of software and hardware , This will significantly increase costs .

In this case , Non relational database Get rapid development .

What is? NoSQL database

RDB（Relational Database）： Relational database .

NoSQL（NoSQL = Not Only SQL）, meaning “ not only SQL”, It is also called non relational database （non-relational）.

Non relational Is to put the data directly into a big warehouse , No label 、 Do not connect 、 Simply pile it up , Thus, the requirements for high-performance storage and access of massive data are improved .

Traditional databases are burdened by various relationships , Constraints of various data forms , It is difficult to handle massive data and highly concurrent business scenarios .

In order to solve the above problems , You must have a self defeating martial art , In order to break through the bottleneck of the database system at a higher level , for example MongoDB, It can only cope with... By abandoning its inherent patterns Facebook Hundreds of millions of bits of massive data .

If you've only touched relational databases before , Such as Oracle、Mysql or SQL Server, I'm learning MongoDB You may feel uneasy , Suddenly there is a database that does not support foreign keys , Unsupported transaction , Data type conventions are not supported , It makes people feel useless .

---

On modern computing systems , A huge amount of data is generated on the network every day , A lot of this data is managed by RDBMS （RDBMS：Relational Database Manangement System） To deal with it .

1970 year E.F.Codd’s The paper that put forward the relation model “A relational model of data for large shared data banks”, This makes data modeling and application programming easier .

It is proved by practice that , The relational model is very suitable for client server programming , Far more than expected , Today it is the leading technology of structured data storage in network and business applications .

NoSQL Is a new database revolutionary movement , In the early days, someone put forward , To 2009 The annual trend is rising .NoSQL Advocates of non relational data storage , Relative to the overwhelming use of relational databases , There is no doubt that this concept is a brand new injection of thinking .

NoSQL classification

Key value database

This kind of database mainly uses the key in the data structure Key To find specific data Value.

advantage ： Do not use any mode when storing , So it's easy to add data .

This kind of database has very high read and write performance , A high access load for processing large amounts of data is appropriate .

The key value database is suitable for high access and write load scenarios of large amounts of data , For example, the logging system .

The main representatives are Redis、Flare.

Document database

This kind of database meets the storage and access needs of massive data , At the same time, the field requirements are not strict , You can add as you like 、 Delete 、 Modify fields , And there is no need to define the table structure in advance , So it is suitable for all kinds of network applications .

The main representatives are MongoDB、CouchDB.

Document databases are very similar to relational databases .

Column storage database

The main representatives are Hbase、Cassandra.

This kind of database is fast to find , High scalability , Suitable for distributed file storage system .

Graph database

The main representatives are InfoGrid、Neo4J.

Such databases use ” The graph structure “ To store the relationship information between entities , A relationship map suitable for building social networks and recommendation systems .

NoSQL And RDB How to choose

since NoSQL There are so many databases , Can it directly replace the relational database ？

NoSQL It can not completely replace the relational database ,NoSQL It is mainly used to deal with the storage and operation of large and multivariate data . Under such characteristic differences , How do we choose the right database to solve the problem of data storage and processing ？

Here are the following points as the basis for judgment ：

1、 The relevance requirements of the data model

NoSQL It is suitable for applications with low model relevance , therefore ：

• If you need multi table Association , be RDB More suitable for

• If the object entity association is less , be NoSQL Database is more suitable for

  • among MongoDB It can support data structures with relatively high complexity , The ability to embed the associated data in a document , So as to reduce the correlation operation between data

2、 Performance requirements of database

If there is a large amount of data and access speed is critical , So use NoSQL Database may be appropriate .NoSQL The database can greatly improve the storage performance through the distributed storage of data .

3、 Data consistency requirements

NoSQL Databases have one drawback ： It does not match in transaction processing and consistency RDB par .

therefore ,NoSQL It is difficult for a database to satisfy both strong consistency and high concurrency .

If the application has high performance requirements , be NoSQL The database can only achieve the final consistency of data .

4、 Data availability requirements

Considering that the unavailability of data may cause risks ,NoSQL Database provides powerful data availability （ In some applications that require quick feedback to users , Response latency is also considered a degree of high availability ）.

A project does not select only one database , It can be disassembled for design , Will need RDB The characteristics of RDB In the management , Other data is put into NoSQL In the management .

MongoDB brief introduction

Reference resources ：MongoDB brief introduction

What is? MongoDB

Official documents ：https://www.mongodb.com.

MongoDB By C++ language-written , Is an open source based on distributed file storage NoSQL Database system .

MongoDB Is a product between relational and non-relational databases , Non-relational databases are the most versatile , Most like a relational database . This will make it easier for people who have used relational databases .

MongoDB Store data as a document , Data structure consists of key value pairs （key=>value） form .

MongoDB The document is BSON（Binary Serialized Document Format Binary serialized document format ） Format , Similar to JSON object , Field values can contain other documents 、 Arrays and document arrays .

MongoDB The query function of is very powerful ：

• It not only supports single tables in most relational databases （ aggregate ） Inquire about , It also supports range query 、 Sort 、 polymerization 、MapReduce etc.

• MongoDB The query syntax of is similar to the object-oriented programming language

MongoDB The development history

• 2007 year 10 month ,MongoDB from 10gen Team development

• 2009 year 2 For the first time this month 1.0 edition

• 2011 year 9 Published in 2.0 edition , Fragmentation 、 Copy and other functions

• 2015 year 3 Published in 3.0 edition ,WiredTiger Storage engine support

• 2018 year 6 Published in 4.0 edition , Introduction ACID Transaction support , Be the first to support strong affairs NoSQL database

• …

MongoDB What scenarios apply

1、 Need to deal with a large number of low value data , It also has high requirements for data processing performance

So-called Low value Even if the data is lost, it will not have much impact .

such as , Banking transaction system belongs to high-value data .

such as , The processing of microblog data does not need to be too transactional , But there are high requirements for data access performance , This is very suitable for use MongoDB.

2、 Need to use the cache layer to process data

because MongoDB Can process data efficiently , So it is very suitable to be used as a cache layer . take MongoDB As a persistent cache layer , It can avoid the resource overload of the underlying storage .

For example, after the user logs in session Session information is stored in MongoDB in .

3、 Requires a high degree of scalability

For relational databases , When the size of the table reaches a certain order of magnitude , Its performance will decline dramatically . At this time, multiple can be used MongoDB The server builds a cluster environment , Achieve maximum expansion , And it doesn't affect performance .