HBase and RDBMS

The difference between hbase and traditional relational databases

HBase is suitable for databases with unstructured data stores. A data storage method between the map Entry and the DB row.

1. Data type: HBase only the simple string type, it only saves the string all types are handed to the user to handle. Relational databases can select types

2. Data manipulation: HBase operation only very simple insert, query and other operations, the table is separated from the table, there is no join

3. Storage mode: HBase is based on column storage, each column family is saved by several files, and the files of different column families are separated. The traditional relational database is saved based on the table structure and the row pattern.

4. Data maintenance: When the HBase update operation, the old version is still retained, and the new data is actually inserted. Traditional relational databases are substitution modifications

5. Scalability: HBase can easily increase or decrease the number of hardware

HBase Mode Design

Eg: the student selection table structure is in the RDBMS as follows:

Student Table (Student):

S_No	S_name	S_sex	S_age
School Number	Name	Gender	Age

Timetable (course):

C_no	C_name	C_credit
Course Number	Course Name	Credits

Selected timetable:

Sc_sno	Sc_cno	Sc_score
School Number	Course Number	Results

In HBase:

Student

Row key	Column Family		Column Family
	Info	Value	Course	Value
S_No	Info:s_name Info:s_sex Info:s_age	Name Sex Age	Course:c_no1 Course:c_no2 Course:c_no3	Sc_score1 Sc_score2 Sc_score3

Course

Row key	Column Family		Column Family
	Info	Value	Student	Value
C_no	Info:c_name Info:c_credit	Name Sex	Student:s_no1 Student:s_no2	Sc_score1 Sc_score2

Working principle

1) HBase Client side Connection Zookeeper

2) which server management is acquired through the Zookeeper component client-root-region

3) Client Access Management-root-server

4) Get management by-root-. Meta. Regionserver

5). Meta. All table information in HBase is recorded, based on. Meta. Get region distribution information (after getting the client will cache this information), Access Hregionserver

(Because the client caches region information, the number of accesses is so much that even if you do not access the. Meta., you can also know which hregionserver to visit)

-root-,.meta. The table relationship is as follows: