Feature
High Performance
Supports memory tables for ultra-fast transaction processing
High Availability
- Multi process architecture
- Replica Node Support
- Redo log and checkpoint background processing
Scalability
- Scale in/out for large-scale data distribution and processing
- Versatility
Versatility
- Use table modes suitable for tasks with Memory TBS and Disk TBS
- Flexible system configurations tailored to user requirements
High Performance
Goldilocks was developed as an In-Memory Architecture.
- Data is kept resident in memory, avoiding any Disk IO operations.
- Only algorithms optimized for in-memory operations are used.
- Logging for index changes is not required, leading to reduced logging IO costs.
- In-memory transactions similarly use the WAL (Write-Ahead Logging) algorithm to safeguard against data loss.
Memory Tables– TPC-C Benchmark TEST
High Availability
Simple configuration of the application to support failover
Failover configuration (connection string)
ALTERNATE_SERVERS = (HOST=192.168.0.101:PORT=22581,
HOST=192.168.0.102:PORT=22581,
HOST=192.168.0.103:PORT=22581)
retry:
EXEC SQL UPDATE POC_DEDUCTIONSET USG_USED_AMOUNT =…
if( sqlca.sqlcode == DB_RETRY_TRANS )
{
goto tx_retry;
}
App error code configuration
GOLDILOCKS operates seamlessly in virtualization environments such as OpenStack and Kubernetes.
Scalability
Users can select from various sharding policies
for distributed processing of large-scale data.
-
Data is distributed and stored according to
the sharding policy selected by the user. -
Data can be joined and global transactions can be
processed across cluster groups. -
Data from multiple nodes can be used as if it were a single database.
-
Supported sharding policies : Hash, Range, List, Cloned
범용성
Use the table mode best suited for your workload with Memory TBS and DISK TBS.
Tools
-Logo_wine.png){kind=logo}
