Hello all,
Do you know the features of GaussDB T database? This post will answer this question.
High Performance
● Query optimization
Query optimization maximizes the query performance of GaussDB T and improves database administrators' efficiency. In addition, rule-based optimizer (RBO) and cost-based optimizer (CBO) are built in GaussDB T to provide high query performance.
● Parallel query
GaussDB T processes a query with a large number of rows in parallel through proper utilization of CPU resources, thereby maximizing running efficiency of a server and improving query performance. In a parallel query of GaussDB T, the server generates several working threads based on the maximum number of parallel rows specified in hints of the SELECT statement initiated by an application. Then, the server splits the statement and allocates the pieces to the working threads for processing, thereby fully utilizing CPU resources.
● MVCC
GaussDB T supports multi-version concurrency control (MVCC) based on timestamps and UNDO. In this way, MVCC prevents data reading and modification from blocking each other, isolating new data from old data. This greatly improves the capability of concurrent data modification and access.
● Concurrency control
GaussDB T implements high concurrency through mechanisms such as locks, MVCC, and database isolation levels. This ensures data integrity when multiple transactions are executed by more than one user. In this way, data is secure and consistent after operations are performed through concurrent threads in the database.
● SQL statement execution plans can be cached to shared pool memory, which greatly improves SQL statement parsing efficiency
● Range partitioning, hash partitioning, list partitioning, and interval partitioning are supported. When partitioned tables are queried, partition pruning improves the query performance by over 10 times.
High Reliability
● Flashback and recycle bin
Generally, it is difficult to recover table data that is updated, deleted, truncated, or dropped by mistake. In some cases, you can restore data to the time point before the mistaken operation, which is called Point-In-Time Recovery (PITR). However, the recovery may make the entire database unavailable, and some table operations that users do not want to undo will also be undone. GaussDB T supports flashback and recycle bin. You can perform a flashback query or flash back a table based on timestamps to recover the data that is deleted or updated by mistake. By flashing back truncated or dropped table objects from the recycle bin, you can restore table data to the state before the mistaken operation, which greatly improves the reliability of user data.
● Logical replication
GaussDB T supports logical replication to synchronize data between GaussDB T and other heterogeneous databases. Specifically, the changes of specific database objects (usually data tables) can be synchronized to another database, and the data of multiple database instances can be aggregated to one instance for data analysis. In addition, logical replication helps users with in-service upgrade, monitors and displays data changes in key tables, and discloses backup data to different user groups. Figure 1 shows the process.
Figure 1 Logical replication
High Availability
● Rolling upgrade, flashback, and online redefinition ensure that the database is not stopped and services are not interrupted.
● Full backup, incremental backup, and standby database synchronization ensure data security.
Large Capacity
In GaussDB T, you can create a bigfile tablespace. In this way, there will be a tablespace with a single, but very large data file in GaussDB T. This reduces the number of data files and simplifies O&M.
Low Costs
● Proper compatibility with SQL
GaussDB T supports hundreds of syntaxes in DQL, DML, DDL, and DCL. GaussDB T fully supports SQL:2003 and partially supports SQL:2006, SQL: 2008, SQL:2011, and SQL:2016. In addition, GaussDB T provides syntaxes (for example, MERGE INTO) suitable for different ecosystems. The multiecosystem SQL syntaxes greatly reduce the cost of migration from other commercial databases to GaussDB T.
● Standard interfaces are supported, such as ODBC/JDBC and CLI tools. Applications can be quickly migrated and launched.
● The database is deployed on x86 servers or Kunpeng-powered servers in HA or distributed mode
Thank you.
