Serverless
TaurusDB serverless instances automatically scale compute and storage capabilities in real time to match fluctuating workload demands. You only pay for the resources you use, reducing database O&M costs and minimizing resource waste. This makes it an ideal choice for highly variable workloads as well as development and testing environments.
Why You Need Serverless
Database stability and reliability are critical for enterprise-grade IT systems. Any database degradation compromises the entire application system. To guarantee performance during peak hours, users adopt conservative provisioning strategies. They often allocate extra compute, memory, and storage resources and set higher limits for configuration parameters.
However, during off-peak hours, these over-provisioned resources sit idle, causing significant financial waste as costs run continuously without return. Conversely, if an unexpected traffic surge exceeds even these conservative estimates, the database can run out of resources, degrading overall performance.
Resolving this issue requires balancing performance with cost efficiency. TaurusDB serverless instances resolve this dilemma by automatically scaling out or scaling in capacity based on real-time application demands. By dynamically adjusting compute power and storage to match actual workloads, the system guarantees high performance while optimizing infrastructure spend. This eliminates the technical barriers of manual database management, allowing enterprises to focus on business innovation while quickly responding to traffic changes.
Advantages
- Cost optimization: You pay only for the resources you use. No additional infrastructure is required and resources are available immediately after purchase. Automatic resource management helps save more than 30% of compute resources, especially for workloads with fluctuating demand.
- Auto scaling: Compute resources are automatically adjusted based on real-time workloads, without manual configuration. Horizontal scaling (changing the number of read replicas) and vertical scaling (changing CPU and memory specifications) work together to provide refined and powerful compute capacity for unexpected or burst workloads. Resources are provisioned and reclaimed within seconds, ensuring uninterrupted user experience.
- Massive storage capacity: The maximum storage capacity reaches 128 TB. Storage automatically scales based on data volume, preventing interruptions caused by insufficient storage.
- Granular billing: Resources are allocated and billed in 0.5 TCU granularity. One TCU provides the compute power equivalent to 1 vCPU with 2 GB of memory. You pay only for the compute resources actually used, eliminating waste.
- Fully managed operations: Huawei Cloud expert teams handle all operational tasks, including specification scaling, automatic storage expansion, monitoring, alerting, and intelligent O&M. These background operations are transparent to users, ensuring continuous service availability and O&M-free experience.
- Robust security: Built-in data encryption, access control, and audit mechanisms ensure data security.
Use Cases
- E-commerce applications: During promotions, e-commerce platforms face massive traffic surges and fluctuating daily access volumes. The database must be able to seamlessly handle sudden spikes. TaurusDB serverless instances support critical workflows like order processing, inventory management, and user data storage. With auto scaling, the database seamlessly handles peak traffic to guarantee stable operations and automatically scales down during low-demand periods to eliminate resource waste. This feature helps enterprises optimize costs through a pay-per-use billing model.
- Mobile application backend services: As mobile user bases scale rapidly and unpredictably, backend services must guarantee low latency and highly responsive performance. Serverless instances support backend services such as user data management, message push, and content storage. Its rapid scalability automatically adapts to shifting user traffic, minimizing operational overhead, accelerating response times to workload fluctuations, and ensuring a seamless end-user experience.
- Data analysis platforms: Data analysts often process large volumes of historical data with complex and variable query patterns, causing significant fluctuations in compute demand. Serverless instances provide both data storage and analytical capabilities for such platforms and support complex queries and report generation. By allocating compute resources strictly on demand, users pay only for the capacity they consume. This reduces analytics costs, improves processing efficiency, and avoids resource waste.
How Serverless Works
TaurusDB Serverless is built on shared storage using a single-write, multi-read architecture. Both the primary node and read replicas are serverless instances. Vertical scaling (changing CPU and memory specifications) completes within seconds based on workload changes, and horizontal scaling (changing the number of read replicas) typically takes a few minutes. You only need to specify a TaurusDB compute unit (TCU) range, where 1 TCU is equivalent to roughly 1 vCPU and 2 GB of memory. The system handles the rest, continuously monitoring CPU and memory utilization to adjust resources automatically. When a client request arrives, the database proxy parses and authenticates it and then schedules it to the appropriate compute node. The compute node processes data by interacting with distributed shared storage to ensure data security. Simultaneously, a resource management module monitors load and automatically adjusts compute resources. This closed-loop design responds rapidly to peaks and troughs, ensuring maximum performance at the lowest possible operational cost.
Serverless Instances vs. Common Instances
TaurusDB serverless instances and common instances represent two distinct deployment models within the Huawei Cloud TaurusDB ecosystem. Both offer MySQL-compatible database engines, persistent storage guarantees, and high-availability architectures.
Core Differences
- Resource elasticity: Serverless instances automatically scale out, in, up, or down based on real-time workloads. Vertical scaling (changing CPU and memory specifications) completes in seconds, and horizontal scaling (changing the number of read replicas) usually takes several minutes. Common instances require manual specification adjustments and cannot adapt automatically to workload fluctuations.
- Billing mode: Serverless instances are billed strictly on actual resource consumption measured in TCUs (1 TCU ≈ 1 vCPU | 2 GB), ensuring you never pay for unconsumed capacity. Common instances are billed at a fixed rate, requiring full payment regardless of actual utilization.
- O&M complexity: Serverless instances operate on a fully hosted, O&M-free model where scaling operations, monitoring, and alarm handling are handled by Huawei Cloud expert teams. In contrast, common instances require users to perform their own O&M tasks, including performance tuning and troubleshooting.
The differences reflect their underlying architectural priorities. Serverless instances use a decoupled storage-compute architecture that pools compute resources for elastic scheduling to maximize flexibility and cost optimization. Common instances use a traditional architecture with fixed resource allocation, prioritizing absolute predictability and static stability for unvarying workloads.
| Dimension | Serverless Instance | Common Instance |
|---|---|---|
| Resource elasticity | Automated scaling; vertical scaling in seconds, horizontal scaling in minutes | Manual specification adjustments required; unable to adapt to traffic shifts |
| Billing mode | Billing based on actual usage (TCU). Users pay only for consumed resources. | Billed based on fixed specifications; fixed fee regardless of utilization |
| O&M complexity | O&M-free. All O&M tasks are handled by the cloud provider. | User-managed O&M, including performance tuning and troubleshooting |
| Application scenarios | Ideal for highly variable and unpredictable workloads | Ideal for fixed workloads with predictable resource requirements |
| Cost characteristics | Low cost under fluctuating loads, eliminating resource waste | More predictable cost for fixed workloads, simplifying budget forecasting |
Related Features and Operations
Once you understand the core mechanics and use cases of serverless instances, you can use its basic and advanced operations to optimize your workloads. For step-by-step procedures, click the corresponding hyperlinks to access the detailed Huawei Cloud documentation.
- Creating a serverless instance: Serverless instances automatically scale based on application demand. For details, see Buying a Serverless Instance.
- Modifying the compute range: You can adjust the compute range of a serverless instance to align with your workload needs. When trigger conditions are met, the instance automatically adjusts the compute capacity. For details, see Changing the Compute Range.
- Adding serverless nodes: You can add serverless read replicas as needed. For details, see Adding Serverless Read Replicas to an Instance.
- Enabling dynamic serverless for instances with fixed specifications: TaurusDB allows you to enable Dynamic Serverless for pay-per-use or yearly/monthly instances and set a compute capability range for them. When specific conditions are met, the system automatically adjusts the compute capacity of the serverless instances, ensuring flexible resource scheduling and cost optimization. For details, see Enabling Dynamic Serverless for Instances with Fixed Specifications.
What is your overall rating for this page?
Thank you very much for your feedback. We will continue working to improve the documentation.See the reply and handling status in My Cloud VOC.
For any further questions, feel free to contact us through the chatbot.
Chatbot