Data Warehouse Flavors

GaussDB(DWS) provides storage-compute coupled and decoupled data warehouses. Additionally, storage-compute coupled data warehouses can also be deployed in the standalone mode. For details about the differences between them, see Data Warehouse Types.

You are advised not to use clusters with low specifications, such as clusters with 16 GB memory and 4-core vCPUs, in the production environment. Otherwise, resource overload may occur.

Flavors for Storage-Compute Coupled Clusters

A storage-compute coupled data warehouse using cloud disks with a vCPU to memory ratio of 1:8 can be elastically scaled, providing unlimited computing and storage capacity. For details, see Table 1.
A storage-compute coupled data warehouse using cloud disks with a vCPU to memory ratio of 1:8 provides high-concurrency, high-performance, and low-latency transaction processing capabilities at low costs based on large-scale data query and analysis capabilities. This type of data warehouse is ideal for HTAP hybrid load scenarios. For details about the specifications, see Table 2.
Choosing storage-compute coupled (standalone) flavors limits deployment to a single node without HA services. This choice can reduce storage costs by 50%. In standalone mode, service availability is maintained through automatic ECS rebuilding, and data reliability is ensured through the EVS multi-copy mechanism. The standalone system is more cost-effective and recommended for lightweight services. When creating a cluster, you can choose the h1 node flavor. For details about the flavor, see Table 3.
A storage-compute coupled data warehouse using local disks cannot be scaled up. You can only increase capacity by adding more nodes. For details, see Table 4.

Step indicates the interval for increasing or decreasing the disk size during cluster configuration change. You need to select a value based on the storage step of the corresponding flavor.

**Table 1** Cloud disk flavors with a vCPU to memory ratio of 1:8 for storage-compute clusters
Flavor	CPU Architecture	vCPU	Memory (GB)	Storage Capacity Per Node	Default Storage	Step (GB)	Recommended Storage	Number of DNs	Scenario
dwsx2.xlarge.m7	x86	4	32	20 GB–2,000 GB	100	10	800	1	Suitable for GaussDB(DWS) starters. These flavors can be used for testing, learning environments, or small-scale analytics systems.
dwsk2.xlarge	Arm	4	32	20 GB–2,000 GB	100	10	800	1
dwsx2.xlarge.m7n	x86	4	32	20 GB–2,000 GB	100	10	800	1
dwsk2.xlarge.km2	Arm	4	32	20 GB–2,000 GB	100	10	800	1
dwsx2.2xlarge.m7	x86	8	64	100 GB–4,000 GB	200	100	1,600	1	Suitable for internal data warehousing and report analysis in small- and medium-sized enterprises (SMEs).
dwsk2.2xlarge	Arm	8	64	100 GB–4,000 GB	200	100	1,600	1
dwsx2.2xlarge.m7n	x86	8	64	100 GB–4,000 GB	200	100	1,600	1
dwsk2.2xlarge.km2	Arm	8	64	100 GB–4,000 GB	200	100	1,600	1
dwsx2.4xlarge.m7	x86	16	128	100 GB–8,000 GB	400	100	3,200	1
dwsk2.4xlarge	Arm	16	128	100 GB–8,000 GB	400	100	3,200	1
dwsk2.4xlarge.km2	Arm	16	128	100 GB–8,000 GB	400	100	3,200	1
dwsx2.8xlarge.m7	x86	32	256	100 GB–16,000 GB	800	100	6,400	2	Recommended for the production environment. These flavors are applicable to OLAP systems that have to deal with large data volumes, BI reports, and data visualizations on large screens for most companies.
dwsk2.8xlarge	Arm	32	256	100 GB–16,000 GB	800	100	6,400	2
dwsx2.8xlarge.m7n	x86	32	256	100 GB–16,000 GB	800	100	6,400	2
dwsk2.8xlarge.km2	Arm	32	256	100 GB–16,000 GB	800	100	6,400	2
dwsk2.12xlarge	Arm	48	384	100 GB–24,000 GB	1200	100	9,600	4	These flavors can deliver excellent performance and are applicable to high-throughput data warehouse processing and high-concurrency online query.
dwsx2.16xlarge.m7	x86	64	512	100 GB–32,000 GB	1,600	100	12,800	4
dwsx2.16xlarge.m7n	x86	64	512	100 GB–32,000 GB	1,600	100	12,800	4
dwsx2.16xlarge.m7	x86	64	512	100 GB–32,000 GB	1,600	100	12,800	4
dwsk2.16xlarge	Arm	64	512	100 GB–32,000 GB	1,600	100	12,800	4
dwsx2.24xlarge.m7	x86	96	768	100 GB–48,000 GB	2,400	100	19,200	4
dwsk2.24xlarge	Arm	96	768	100 GB–48,000 GB	2,400	100	19,200	4
dwsx2.32xlarge.m7	x86	128	1,024	100 GB–48,000 GB	3,200	100	25,600	4

**Table 2** Cloud disk flavors with a vCPU to memory ratio of 1:4 for storage-compute clusters
Flavor	CPU Architecture	vCPU	Memory (GB)	Storage Capacity Per Node	Step (GB)	Number of DNs	Scenario
dwsx2.h.xlarge.4.c7	x86	4	16	20 GB–2,000 GB	20	1	Suitable for GaussDB(DWS) starters. These flavors can be used for testing, learning environments, or small-scale analytics systems.
dwsk2.h.xlarge.4.kc1	Arm	4	16	20 GB–2,000 GB	20	1
dwsk2.h.xlarge.kc2	Arm	4	16	20 GB–2,000 GB	20	1
dwsx2.h.xlarge.4.c7n	x86	4	16	20 GB–2,000 GB	20	1
dwsx2.h.2xlarge.4.c6	x86	8	32	100 GB–4,000 GB	100	1	Suitable for internal data warehousing and report analysis in small- and medium-sized enterprises (SMEs).
dwsx2.h.2xlarge.4.c7	x86	8	32	100 GB–4,000 GB	100	1
dwsk2.h.2xlarge.4.kc1	Arm	8	32	100 GB–4,000 GB	100	1
dwsk2.h.2xlarge.kc2	Arm	8	32	100 GB–4,000 GB	100	1
dwsx2.h.2xlarge.4.c7n	x86	8	32	100 GB–4,000 GB	100	1
dwsx2.h.4xlarge.4.c7	x86	16	64	100 GB–8,000 GB	100	1	Recommended for the production environment. These flavors are applicable to OLAP systems that have to deal with large data volumes, BI reports, and data visualizations on large screens for most companies.
dwsk2.h.4xlarge.4.kc1	Arm	16	64	100 GB–8,000 GB	100	1
dwsk2.h.4xlarge.kc2	Arm	16	64	100 GB–8,000 GB	100	1
dwsx2.h.4xlarge.4.c7	x86	16	64	100 GB–8,000 GB	100	1
dwsx2.h.8xlarge.4.c7	x86	32	128	100 GB–16,000 GB	100	2
dwsk2.h.8xlarge.4.kc1	Arm	32	128	100 GB–16,000 GB	100	2
dwsk2.h.8xlarge.kc2	Arm	32	128	100 GB–16,000 GB	100	2
dwsx2.h.8xlarge.4.c7n	x86	32	128	100 GB–16,000 GB	100	2
dwsk2.h.12xlarge.4.kc1	Arm	48	192	100 GB–24,000 GB	100	4	These flavors can deliver excellent performance and are applicable to high-throughput data warehouse processing and high-concurrency online query.
dwsk2.h.12xlarge.kc2	Arm	48	192	100 GB–24,000 GB	100	4
dwsx2.h.16xlarge.4.c7	x86	64	256	100 GB–32,000 GB	100	4
dwsx2.h.16xlarge.4.c7n	x86	64	256	100 GB–32,000 GB	100	4
dwsk2.h.16xlarge	Arm	64	256	100 GB–32,000 GB	100	4
dwsk2.h.24xlarge	Arm	96	384	100 GB–48,000 GB	100	4
dwsk2.h.32xlarge	Arm	128	512	100 GB–64,000 GB	100	4

**Table 3** Storage-compute coupled (standalone) cluster flavors
Flavor	CPU Architecture	vCPU	Memory (GB)	Storage Capacity Per Node	Step (GB)	Number of DNs	Scenario
dwsx2.h1.xlarge.2.c7	x86	4	8	20 GB–2,000 GB	20	1	Suitable for GaussDB(DWS) starters. These flavors can be used for testing, learning environments, or small-scale analytics systems.
dwsk2.h1.xlarge.2.kc1	Arm	4	8	20 GB–2,000 GB	20	1
dwsx2.h1.xlarge.2.c7n	x86	4	8	20 GB–2,000 GB	20	1
dwsx2.h1.2xlarge.4.c7	x86	8	32	100 GB–4,000 GB	100	1	Suitable for internal data warehousing and report analysis in small- and medium-sized enterprises (SMEs).
dwsk2.h1.2xlarge.4.kc1	Arm	8	32	100 GB–4,000 GB	100	1
dwsx2.h1.2xlarge.4.c7n	x86	8	32	100 GB–4,000 GB	100	1
dwsx2.h1.4xlarge.4.c7	x86	16	64	100 GB–8,000 GB	100	1	Recommended for the production environment. These flavors are applicable to OLAP systems that have to deal with large data volumes, BI reports, and data visualizations on large screens for most companies.
dwsk2.h1.4xlarge.4.kc1	Arm	16	64	100 GB–8,000 GB	100	1
dwsx2.h1.4xlarge.4.c7n	x86	16	64	100 GB–8,000 GB	100	1
dwsx2.h1.8xlarge.4.c7	x86	32	128	100 GB–16,000 GB	100	2
dwsk2.h1.8xlarge.4.kc1	Arm	32	128	100 GB–16,000 GB	100	2
dwsx2.h1.8xlarge.4.c7n	x86	32	128	100 GB–16,000 GB	100	2
dwsk2.h1.12xlarge.4.kc1	Arm	48	192	100 GB–24,000 GB	100	4	These flavors can deliver excellent performance and are applicable to high-throughput data warehouse processing and high-concurrency online query.
dwsx2.h1.16xlarge.4.c7	x86	64	256	100 GB–32,000 GB	100	4
dwsx2.h1.16xlarge.4.c7n	x86	64	256	100 GB–32,000 GB	100	4

**Table 4** Local disk flavors for storage-compute coupled clusters
Flavor	CPU Architecture	vCPU	Memory (GB)	Storage Capacity Per Node	Number of DNs	Scenario
dws2.olap.4xlarge.i3	x86	16	128	1,490 GB	1	Recommended for the production environment. These flavors are applicable to OLAP systems that have to deal with large data volumes, BI reports, and data visualizations on large screens for most companies. These flavors can deliver excellent performance and are applicable to high-throughput data warehouse processing and high-concurrency online query.
dws2.olap.4xlarge.ki1	Arm	16	64	2,980 GB	1
dws2.olap.8xlarge.i3	x86	32	256	2,980 GB	2
dws2.olap.8xlarge.ki1	Arm	32	128	5,960 GB	2
dws2.olap.16xlarge.i3	x86	64	512	5,960 GB	4
dws2.olap.16xlarge.ki1	Arm	64	228	11,921 GB	4

Flavors for Storage-Compute Decoupled Clusters

A storage-compute decoupled data warehouse using cloud disks can be elastically scaled, providing unlimited computing and storage capacity. For details, see Table 5.
A storage-compute decoupled data warehouse using local disks has a fixed storage capacity that cannot be expanded or modified. You can only increase capacity by adding more nodes. For details, see Table 6.

When creating a storage-compute decoupled cluster, only the second half of the flavors (for example, 4U16G.4DPU) are shown. The prefixes (dwsx3/dwsax3/dwsk3) in the flavor list indicate the storage-compute decoupled CPU architecture.

**Table 5** Storage-compute decoupled cloud disk flavors
Flavor	CPU Architecture	vCPU	Memory (GB)	Storage Capacity Per Node	Step (GB)	Number of DNs	Scenario
dwsx3.4U16G.4DPU	x86	4	16	20 GB–2,000 GB	10	1	Suitable for GaussDB(DWS) starters. These flavors can be used for testing, learning environments, or small-scale analytics systems.
dwsk3.4U16G.4DPU	Arm	4	16	20 GB–2,000 GB	10	1
dwsax3.4U16G.4DPU	x86	4	16	20 GB–2,000 GB	10	1
dwsax3.4U32G.4DPU	x86	4	32	20 GB–2,000 GB	10	1
dwsx3.8U32G.8DPU	x86	8	32	100 GB–4,000 GB	100	1	Suitable for internal data warehousing and report analysis in small- and medium-sized enterprises (SMEs).
dwsk3.8U32G.8DPU	Arm	8	32	100 GB–4,000 GB	100	1
dwsax3.8U32G.8DPU	x86	8	32	100 GB–4,000 GB	100	1
dwsax3.8U64G.8DPU	x86	8	64	100 GB–4,000 GB	100	1
dwsx3.16U64G.16DPU	x86	16	64	100 GB–8,000 GB	100	1	Recommended for the production environment. These flavors are applicable to OLAP systems that have to deal with large data volumes, BI reports, and data visualizations on large screens for most companies.
dwsk3.16U64G.16DPU	Arm	16	64	100 GB–8,000 GB	100	1
dwsax3.16U64G.16DPU	x86	16	64	100 GB–8,000 GB	100	1
dwsax3.16U128G.16DPU	x86	16	128	100 GB–8,000 GB	100	1
dwsx3.32U128G.32DPU	x86	32	128	100 GB–16,000 GB	100	2
dwsk3.32U128G.32DPU	Arm	32	128	100 GB–16,000 GB	100	2
dwsax3.32U128G.32DPU	x86	32	128	100 GB–16,000 GB	100	2
dwsax3.32U256G.32DPU	x86	32	256	100 GB–16,000 GB	100	2
dwsk3.48U192G.48DPU	Arm	48	192	200 GB–24,000 GB	100	4	These flavors can deliver excellent performance and are applicable to high-throughput data warehouse processing and high-concurrency online query.
dwsx3.64U256G.64DPU	x86	64	256	200 GB–32,000 GB	100	4
dwsk3.64U256G.64DPU	Arm	64	256	100 GB–32,000 GB	100	4
dwsax3.64U256G.64DPU	x86	64	256	100 GB–32,000 GB	100	4
dwsax3.64U512G.64DPU	x86	64	512	100 GB–32,000 GB	100	4
dwsx3.96U768G.96DPU	x86	96	768	100 GB–48,000 GB	100	4
dwsk3.96U384G.96DPU	Arm	96	384	100 GB–48,000 GB	100	4
dwsax3.96U384G.96DPU	x86	96	384	100 GB–48,000 GB	100	4
dwsax3.96U768G.96DPU	x86	96	768	100 GB–48,000 GB	100	4
dwsx3.128U1024G.128DPU	x86	128	1,024	100 GB–64,000 GB	100	4
dwsk3.128U512G.128DPU	Arm	128	512	100 GB–64,000 GB	100	4
dwsax3.128U512G.128DPU	x86	128	512	100 GB–64,000 GB	100	4
dwsax3.128U1024G.128DPU	x86	128	1,024	100 GB–64,000 GB	100	4

**Table 6** Storage-compute decoupled local disk flavors
Flavor	CPU Architecture	vCPU	Memory (GB)	Storage Capacity Per Node	Number of DNs	Scenario
dws3.16U128G.i7.16DPU	x86	16	128	2,980 GB	1	Recommended for the production environment. These flavors are applicable to OLAP systems that have to deal with large data volumes, BI reports, and data visualizations on large screens for most companies. These flavors can deliver excellent performance and are applicable to high-throughput data warehouse processing and high-concurrency online query.
dws3.16U64G.ki1.16DPU	Arm	16	64	5,960 GB	1
dws3.32U256G.i7.32DPU	x86	32	256	5,960 GB	2
dws3.32U128G.ki1.32DPU	Arm	32	128	11,920 GB	2
dws3.64U512G.i7.64DPU	x86	64	512	11,920 GB	4
dws3.64U228G.ki1.64DPU	Arm	64	228	23,840 GB	4