DocsICP FilingConsole
官方文档
首页

Instance types

更新时间：
复制 MD 格式
产品详情
我的收藏

The PolarDB-X 1.0 compute layer offers various instance families and types to meet the performance requirements of different business scenarios.

Families and types

FamilyInstance type (CPU and memory)Single-node specificationsNumber of nodesFeatures
Starter Edition8-core 32 GB4-core 16 GB2Designed for initial business development or testing scenarios. Does not provide acceleration for complex queries.
Standard Edition16-core 64 GB8-core 32 GB2Offers a wide range of cost-effective instance types. Designed for online business scenarios that require ultra-high concurrency, complex queries, and lightweight analytics. The parallel query feature is enabled by default to improve the execution efficiency of complex queries, such as table joins, aggregations, and sorting.
32-core 128 GB8-core 32 GB4
48-core 192 GB8-core 32 GB6
64-core 256 GB8-core 32 GB8
96-core 384 GB8-core 32 GB12
128-core 512 GB8-core 32 GB16
Enterprise Edition32-core 128 GB16-core 64 GB2Provides large-scale resources for enterprise-level business scenarios that require ultra-high concurrency, complex queries on large-scale data, and accelerated analytics. The parallel query feature is enabled by default to significantly improve the execution efficiency of complex queries and report analysis on massive datasets.
64-core 256 GB16-core 64 GB4
96-core 384 GB16-core 64 GB6
128-core 512 GB16-core 64 GB8
192-core 768 GB16-core 64 GB12
256-core 1024 GB16-core 64 GB16
348-core 1536 GB16-core 64 GB24
512-core 2048 GB16-core 64 GB32
768-core 3072 GB16-core 64 GB48
1024-core 4096 GB16-core 64 GB64
Note

Sysbench

Sysbench is an open source, modular, cross-platform, and multi-threaded performance testing tool. You can use it to perform performance tests on databases, CPUs, memory, threads, and I/O. It supports databases such as MySQL, Oracle, and PostgreSQL. The following sections describe the performance of PolarDB-X 1.0 in Sysbench online transaction processing (OLTP) and SELECT scenarios.

  • Test design
    • PolarDB-X 1.0 compute layer (four instance types): Starter Edition 8-core 32 GB, Standard Edition 16-core 64 GB, Enterprise Edition 32-core 128 GB, and Enterprise Edition 64-core 256 GB
    • ECS stress testing instance (1): A compute-optimized, network-enhanced instance with 32 cores and 64 GB of memory, running the Alibaba Cloud Linux 2.1903 (64-bit) operating system.
    • RDS (12 instances): 16-core 64 GB, MySQL 5.7 Dedicated
    Note PolarDB-X 1.0 compute layer, ECS, and RDS instances are located in the same zone and VPC.
  • OLTP test results456789
    SpecificationsConcurrencyReads/Writes per second
    Starter Edition 8-core 32 GB10020807.12
    Standard Edition 16-core 64 GB23049667.48
    Enterprise Edition 32-core 128 GB45090693.70
    Enterprise Edition 64-core 256 GB900177506.48
  • SELECT test results456789
    Instance typeConcurrencyReads/Writes per second
    Starter Edition 8-core 32 GB20041401
    Standard Edition 16-core 64 GB30098182.26
    Enterprise Edition 32-core 128 GB600180500.00
    Enterprise Edition 64-core 256 GB1200366863.48

TPC-C

TPC-C is an industry-standard benchmark published by the Transaction Processing Performance Council (TPC) to evaluate the OLTP capabilities of a database. It involves 10 tables and includes the following business transaction models:
  • NewOrder: Creates a new order.
  • Order payment
  • OrderStatus: Queries the status of a recent order.
  • Delivery
  • StockLevel: Analyzes the stock level for out-of-stock items.
Note The TPC-C implementation in this topic is based on TPC-C benchmarking. The results cannot be compared with published TPC-C benchmark results because the tests in this topic do not meet all the requirements of the TPC-C benchmark.

TPC-C uses the tpmC (transactions per minute) value to measure the maximum qualified throughput (MQTh) of a system. The transactions are based on the NewOrder transaction. The final unit of measurement is the number of new orders processed per minute.

  • Test design
    • Enterprise Edition staging environment: PolarDB-X 1.0 compute layer Enterprise Edition 32-core 128 GB (single node: 16-core 64 GB), four RDS for MySQL 5.7 instances (8-core 32 GB Dedicated)
    • Standard Edition staging environment: PolarDB-X 1.0 compute layer Standard Edition 16-core 64 GB (single node: 8-core 32 GB), four RDS for MySQL 5.7 instances (4-core 32 GB Dedicated)
    • Extra-large instance type staging environment: PolarDB-X 1.0 compute layer Enterprise Edition 256-core 1024 GB (single node: 16-core 64 GB), 12 RDS for MySQL 5.7 instances (32-core 128 GB Dedicated)
  • Test results
    ConcurrencyStandard Edition instance tpmCEnterprise Edition instance tpmCExtra-large instance type tpmC
    1 client × 1000 concurrent connections65735.14101620.8N/A
    6 clients × 1000 concurrent connectionsN/AN/A821547.97
  • tpmC curve at different concurrencies3

TPC-H

TPC-H is an industry-standard benchmark published by the TPC to evaluate the analytical query capabilities of a database. TPC-H queries involve eight data tables and 22 complex SQL queries. Most queries include table joins, subqueries, and Group-by aggregations.

Note The TPC-H implementation in this topic is based on TPC-H benchmarking. The results cannot be compared with published TPC-H benchmark results because the tests in this topic do not meet all the requirements of the TPC-H benchmark.
  • Test design
    • Enterprise Edition staging environment: PolarDB-X 1.0 compute layer Enterprise Edition 32-core 128 GB (single node: 16-core 64 GB), four RDS for MySQL 5.7 instances (8-core 32 GB)
    • Standard Edition staging environment: PolarDB-X 1.0 compute layer Standard Edition 16-core 64 GB (single node: 8-core 32 GB), four RDS for MySQL 5.7 instances (4-core 32 GB)
  • Test results5
    QueryEnterprise Edition (Unit: seconds)Standard Edition (Unit: seconds)
    Q0155.82111.84
    Q026.1211.54
    Q0315.9930
    Q0417.7136.56
    Q0510.8923.01
    Q068.0616.76
    Q0717.0934.80
    Q0813.4426.09
    Q0953.81101.51
    Q108.7319.67
    Q1118.2519.74
    Q128.8018.60
    Q1314.1531.33
    Q1417.4942.43
    Q1520.6242.79
    Q162.134.15
    Q171.934.07
    Q1811.0122.82
    Q1912.9727.61
    Q2027.7749.25
    Q2138.8468.08
    Q225.2711.29
    Total386.77754.65
Previous：Service architectureNext：Scenarios