Seller: a-platform ㅣ Available
ClustrixDB (NewSQL) is the latest relational database management system (RDBMS) providing scale-out with the same scalable performance
to even the writing task load as well as the reading task on Online Transaction Processing (OLTP), while maintaining the guarantee of ACID
on the traditional database system.
ClustrixDB is an RDBMS equipped with complete features by combining elaborate distribution architecture and simple SQL interface. It is mostly optimized to the online transactions (OLTP) and developed as an independent source. In addition, it provides MySQL compatibility.
ClustrixDB uses a non-sharing architecture. It is the only architecture that can linearly extend the node while guaranteeing ACID as an RDB.
In a non-sharing architecture, each node manages data by partitioning, and competition decreases as read and write become distributed to multiple nodes.
Because writing is possible in all nodes, writing scalability which is an existing limitation of traditional RDB becomes possible.
Also, ClustrixDB processes by automatic equal distribution of data and query execution from all nodes.
In addition, the database capacity can be easily flexed up or flexed down using ClustrixDB. For both tasks, the data within clusters are redistributed automatically from the background when the database is remaining in the online status. (Executes after being optimized to both cloud and on-premise.)
Scales out without separate replication or artificial sharding
- Writing scalability
- Reading scalability
- Storing scalability
- No change of app
- Automatic HA
- Easy installation (Easy to deploy)
Scales out TPS and connection capacity just by adding server
- 90% reading + 10% writing
- Occurrence of typical DB load
- 1 TPS = 10 SQL/sec
- 9 SELECT + 1 UPDATE
- Provides 32 nodes
High-Volume, High Concurrency OLTP
Handling High Concurrency and High Volume Transactions
- ClustrixDB has been specially created for large scale online transaction processing (OLTP). The transaction always maintains compliance of ACID even in a distributed environment.
- ClustrixDB is a full relational database created for the task loads with a high processing volume. It can scale out reading and writing by simply adding nodes.
- ClustrixDB can implement parallel processing on a level which far exceeds the level possible for a single instance database by distributing data and tasks to other nodes and cores of the clusters.
Fault-Tolerant Auto Failover (HA)
ClustrixDB was designed for fault-tolerance and can generally endure node failures without data loss. If a node becomes unavailable, it is removed from the cluster and the cluster maintains a full cluster configuration using the replicated data of another node. ClustrixDB Rebalancer automatically creates additional replica of data. This replica is generated transparently from the background without user intervention and performs auto recovery of ClustrixDB. (Conducted automatically without DBA involvement)
ClustrixDB automatically detects the loss of a node
- Automatic re-protects systems (recovery of lost slice)
- Redistribute data automatically on remaining nodes
- Prompt re-protection of data on lost nodes
- Reading and writing of the table is still possible even while re-protecting data
- Automatic self-recovery
- Cluster activates with full protection after automatic self-recovery
MySQL Compatible MariaDB with MySQL Compatibility and Easy Migration (Drop-in Replacement)
ClustrixDB uses MySQL syntax on SQL, DML, DDL, triggering, and storing procedures.
Currently, MySQL users have a high possibility of already having a current environment which is compatible with ClustrixDB.
Also, ClustrixDB supports the MySQL replication protocol which includes SBR and RBR.
A simple SQL interface to access data within ClustrixDB is provided to users. An application program sees ClustrixDB as a single instance regardless of the number of nodes in the cluster or the location of data. The application program does not have to be modified to access ClustrixDB. All nodes within ClustrixDB can process reading or writing.
Automatic Data Distribution
Resolving Data Skewness
The key component of ClustrixDB is automatic management of data distribution of cluster with the rebalancer executed continuously from the background. Through this, many replicas are maintained and managed in all clusters. An unexpected node error occurs within a cluster. If the number of replica is inadequate, the rebalancer automatically makes more replicas. Even when data is added or removed, both the data and load are evenly distributed to the cluster.
Rebalancer (Advanced Features by Default)
Rebalancer is in charge of the work to distribute data satisfactorily in a cluster environment.
- Initial data distribution: Equally distributes slices (fragment data) to all nodes
- Load balancing of data explosion: Distributes by splitting into small slices if a specific slice gets larger
- Flex-Up/Flex-Down: Redistributes and recovers slices followed by adding or deleting nodes
- Node down failure recovery: Self-recovery through the slice replication of nodes that are down
- Resolving skewed data: Equally splits and distributes again to the data node
- Hot slice balancing: Splits and distributes the loaded node slice again