Advantages
In database scenarios, Kunpeng servers feature high reliability and stability. Their multiple cores and memory channels show advantages in analytical processing (AP). Kunpeng servers can save investments for enterprises and provide better performance at the same cost. When single-instance deployment cannot fully utilize the CPU computing power, multi-instance deployment on Kunpeng servers is more advantageous, maximizing the utilization of enterprise investments for performance improvement.
Kunpeng servers have the following features:
- Multi-core architecture. High computing power of up to 384 cores enables database parallel processing. The advantage over x86 servers is more evident in OLAP scenarios.
- Multiple memory channels. Up to 24 memory channels improve the memory bandwidth and accelerate the processing of a large amount of data.
- Multi-instance deployment. When single-instance deployment cannot fully utilize the CPU computing power, multi-instance deployment on Kunpeng servers is more advantageous.
High Performance
Kunpeng servers improve the parallel computing level and fully utilize the multi-core capabilities of Kunpeng processors.
- A Kunpeng server (2 × Kunpeng 950 processor) can be configured with up to 96 cores and supports the single-node or SuperPoD architecture.
- A Kunpeng server (2 × new Kunpeng 920 processor model) can be configured with up to 80 cores.
High Security and Reliability
The hardware of the Kunpeng computing platform is designed based on high reliability and strict quality standards.
- The processors have built-in hardware accelerator, which provides higher security.
- The service performance loss caused by encryption is less than 10%.
Open Ecosystem
Kunpeng has a mature big data ecosystem in database scenarios. It supports the following components:
- Mainstream open-source databases, such as MySQL, MariaDB, Percona, PostgreSQL, MongoDB, Greenplum, and Milvus.
- Mainstream commercial databases, such as GBase, Dameng, Kingbase, Shentong Data, and GaussDB, in China
Hardware and Software Collaborative Optimization
Kunpeng databases fully utilize the multi-core computing and instruction set capabilities of Kunpeng processors to improve the overall running efficiency of databases.
- Provides system collaborative optimization based on foundational software such as the OS, compiler, network, and storage.
- Provides database acceleration suites for MySQL, Redis, RocksDB, and Milvus, and performs special optimization for typical scenarios such as transaction processing, analysis and computing, cache access, vector retrieval, and key-value storage.
Full-Lifecycle Enablement
Kunpeng provides full-lifecycle enablement capabilities for database scenarios.
Provides full-lifecycle enablement capabilities covering database migration, deployment, testing and verification, performance tuning, high availability (HA) construction, and O&M. It also provides users with continuous technical support from database rollout to production through auxiliary tools, technical documents, reference architectures, and best practices. Provides the standardized migration and deployment process and scenario-based tuning methods to reduce the complexity of database implementation and O&M on the Kunpeng platform, helping users efficiently migrate, deploy, and continuously optimize databases.