Suggestions on Instance Selection

Overview

Before purchasing a GaussDB instance, consider key factors such as cost, performance, workload capacity, and your intended use case. This section describes the billing modes, editions, instance types, deployment models, storage types, and specifications of GaussDB to help you quickly select the right instance.

Billing Modes

GaussDB provides two billing modes: yearly/monthly and pay-per-use. Choose the billing mode that best fits your usage patterns and workload requirements. For details, see Table 1.

**Table 1** Billing modes
Billing Mode	Yearly/Monthly	Pay-per-use
Payment	Prepaid You pay upfront for a subscription term in exchange for a discounted rate. The longer the subscription term, the greater the discount.	Postpaid You can start using the database instance first and then pay as you go.
Billing Cycle	Billed by the subscription term you purchase	Billed by the second and settled by the hour
Billing Item	Instance specifications (vCPUs and memory) and storage	Instance specifications (vCPUs and memory), storage space, and backup space
Billing Mode Change	You can change the billing mode from yearly/monthly to pay-per-use. The change takes effect only after the yearly/monthly subscription expires. For details, see Yearly/Monthly to Pay-per-Use.	You can change the billing mode from pay-per-use to yearly/monthly. For details, see Pay-per-Use to Yearly/Monthly.
Use Case	Applicable to scenarios where the resource usage duration is predictable and intended for long-term use. This mode is more cost-effective.	Applicable to scenarios where compute resource requirements fluctuate. You can enable and delete resources at any time.

Editions

GaussDB provides two editions: basic edition and enterprise edition. Choose an edition that best fits your usage patterns and workload requirements. For details, see Table 2.

**Table 2** Differences between the basic and enterprise editions of GaussDB instances
Edition Type	Basic edition	Enterprise edition
Supported Instance Type	Centralized Distributed	Centralized Distributed
Supported Deployment Model	Centralized 1 primary + 2 standby 1 primary + 1 standby + 1 log Single Distributed Combined	Centralized 1 primary + 2 standby 1 primary + 1 standby + 1 log Single Distributed Independent
Available Instance Specifications	Supported specification types: General-purpose Dedicated Kunpeng dedicated	Supported specification types: Dedicated Kunpeng dedicated
	Minimum specifications: Centralized 4 vCPUs \| 16 GB (General-purpose, Dedicated, and Kunpeng dedicated) Distributed Combined deployment: 16 vCPUs \| 64 GB (Dedicated and Kunpeng dedicated) NOTE: For a combined deployment, all components must be deployed on a single ECS, which demands higher CPU resources to maintain optimal performance. As a result, a minimum of 16 vCPUs is necessary for a combined deployment.	Minimum specifications: Centralized 4 vCPUs \| 16 GB (Dedicated and Kunpeng dedicated) Distributed Independent deployment: 8 vCPUs \| 64 GB (Dedicated and Kunpeng dedicated)
	Maximum specifications: x86: 128 vCPUs \| 1,024 GB Arm: 128 vCPUs \| 1,024 GB	Maximum specifications: x86: 128 vCPUs \| 1,024 GB Arm: 128 vCPUs \| 1,024 GB
Supported Advanced Feature	Advanced compression	Advanced compression, dynamic data masking, and anti-tampering
Performance (Benchmark: Enterprise Edition)	Centralized instance: same as the enterprise edition Distributed instance in a combined deployment: 50% of the performance of the enterprise edition, when the same specifications are configured	-
Price (Benchmark: Enterprise Edition)	When the specifications are the same, the cost of a centralized instance in the basic edition is about 30% of the cost of the enterprise edition. When the specifications are the same, the cost of a distributed instance in a combined deployment of the basic edition is about 7.5% of the cost of the enterprise edition.	-

Instance Types

Table 3 lists the instance types provided by GaussDB. Choose a type that best fits your usage patterns and workload requirements.

**Table 3** Instance types
Instance Type	Scale-out Supported	Use Case	Component	Service Processing
Distributed	Yes	The data volume is large, and large data capacity and high concurrency are required.	OM CM GTM ETCD CN DN	An application sends a SQL query request to a CN. The CN uses the optimizer of the database to generate an execution plan and sends the plan to DNs. Each DN processes data based on the execution plan. After the processing is complete, DNs return the result sets to the CN for summary. Finally, the CN returns the summary result to the application.
Centralized	No	The data volume is small and stable, and data reliability and service availability are important.	OM CM ETCD DN	An application sends a task directly to DNs, and the DNs then process and return the results to the application.

Deployment Models

Table 4 describes the deployment models supported by GaussDB. Choose a deployment model that best fits your usage patterns and workload requirements.

**Table 4** Deployment models
Deployment Model		Characteristic	Use Case
Distributed	Independent deployment	The CN, DN, CMS, and GTM components of the database are deployed on different nodes.	Applicable to scenarios where there is a need to process a large amount of data, improve the storage capacity and concurrency of instances through scale-out, and the performance requirements cannot be met by deploying a distributed instance in combined deployment.
Distributed	Combined deployment	Different database components are deployed on the same node.	Applicable to scenarios with large data volumes, potential future data volume growth, moderate performance requirements, and a need for a distributed database.
Centralized	1 primary + 2 standby	The 1 primary + 2 standby deployment model is a 3-node deployment where there is one shard.	Applicable to scenarios with small and stable volumes of data, where data reliability and service availability are important.
	1 primary + 1 standby + 1 log	The 1 primary + 1 standby + 1 log deployment model is a 3-node deployment where there is one shard. The log node is only used to store logs and does not incur fees. This reduces costs and resource consumption.	Applicable to the same scenario as the centralized model of 1 primary + 2 standby, but with slightly lower reliability.
	Single (also known as single-replica)	The single-replica deployment model is a single-node deployment where there is only one CMS and one DN.	Applicable only to test scenarios with small data volume.

Storage Types

GaussDB supports ultra-high I/O and extreme SSD storage types. Choose a type that best fits your usage patterns and workload requirements. For details, see Table 5.

**Table 5** Storage types
Storage Type	Use Case
Ultra-high I/O	Ultra-high performance cloud disks are excellent for enterprise mission-critical services as well as workloads demanding high throughput and low latency.
Extreme SSD	Superfast disks are ideal for workloads demanding ultra-high bandwidth and ultra-low latency. NOTE: Extreme SSD storage is only available for authorized users. To apply for the permissions needed, submit a service ticket.

Specifications

Table 6 lists the specifications supported by GaussDB. Choose a specification that best fits your usage patterns and workload requirements.

**Table 6** Specifications
Specification		CPU Architecture	Supported Edition	Description
General-purpose	General-purpose (1:4)	x86	Basic edition	This specification type uses general-computing ECSs, which prioritize resource sharing and offer better cost-effectiveness. However, if the database is heavily loaded, it can lead to resource contention and fluctuations in computing performance. Consequently, this specification type is not suitable for large services, and it is not offered in the enterprise edition.
Dedicated	Dedicated (1:4) Dedicated (1:8)	x86	Basic edition Enterprise edition	This specification type uses general computing-plus x86 ECSs. There is no resource contention between ECSs, and the performance is stable.
Kunpeng dedicated	Kunpeng dedicated (1:4) Kunpeng dedicated (1:8)	Arm	Basic edition Enterprise edition	This specification type uses general computing-plus ECSs with Kunpeng CPUs. There is no resource contention between ECSs, and the performance is stable.