Recommend a database service tier and compute tier
5 minutes
5 Questions
When recommending database service and compute tiers in Azure, architects must evaluate workload requirements, performance needs, and budget constraints to select the optimal configuration.
For Azure SQL Database, three primary service tiers exist:
**DTU-based tiers:**
- **Basic**: Suitable for s…When recommending database service and compute tiers in Azure, architects must evaluate workload requirements, performance needs, and budget constraints to select the optimal configuration.
For Azure SQL Database, three primary service tiers exist:
**DTU-based tiers:**
- **Basic**: Suitable for small databases with light workloads, offering limited performance at low cost
- **Standard**: Ideal for most business applications requiring moderate performance and storage
- **Premium**: Designed for mission-critical applications demanding high I/O throughput and low latency
**vCore-based tiers:**
- **General Purpose**: Balances compute and storage for typical business workloads with standard availability
- **Business Critical**: Provides highest resilience with built-in high availability replicas and fastest storage
- **Hyperscale**: Supports databases up to 100TB with rapid scale-out capabilities and near-instantaneous backups
**Compute tier selection** involves choosing between:
- **Provisioned compute**: Best for predictable workloads where you specify exact vCores needed, paying per hour
- **Serverless compute**: Optimal for intermittent usage patterns with auto-scaling and per-second billing during active periods
**Key considerations for recommendations:**
1. **Performance requirements**: Analyze DTU/vCore needs based on CPU, memory, and I/O demands
2. **Storage size**: Evaluate current data volume and growth projections
3. **Availability requirements**: Higher tiers offer better SLAs and redundancy options
4. **Latency sensitivity**: Business Critical tier provides in-memory OLTP and faster storage
5. **Cost optimization**: Match tier to actual usage patterns; avoid over-provisioning
6. **Scaling needs**: Consider whether workloads require elastic pools for multiple databases
Architects should analyze existing workload metrics, conduct performance testing, and consider future growth when making tier recommendations. Starting with lower tiers and scaling up based on monitoring data often proves more cost-effective than initial over-provisioning.
Recommend a Database Service Tier and Compute Tier
Why This Is Important
Selecting the appropriate database service tier and compute tier is critical for optimizing cost, performance, and scalability in Azure solutions. As an Azure Solutions Architect, you must balance business requirements with technical constraints to ensure databases perform efficiently while staying within budget. Poor tier selection can lead to either overspending on unused resources or performance bottlenecks that impact user experience.
What Are Database Service Tiers and Compute Tiers?
Database Service Tiers define the overall capabilities and features available for your database workload. In Azure SQL Database, these include:
• Basic - Suitable for small databases with light workloads • Standard - Appropriate for most business applications with moderate performance needs • Premium - Designed for mission-critical applications requiring high I/O and low latency • Hyperscale - For very large databases up to 100 TB with rapid scale-out capabilities • Business Critical - Offers built-in high availability with local SSD storage • General Purpose - Balanced compute and storage for general workloads
Compute Tiers determine how compute resources are allocated:
• Provisioned Compute - Pre-allocated compute resources billed per hour regardless of usage • Serverless Compute - Auto-scales based on workload demand with per-second billing
How It Works
When designing database solutions, you evaluate several factors:
1. Workload Characteristics • Consistent vs. variable workloads favor provisioned vs. serverless respectively • Read-heavy workloads may benefit from read replicas • Write-intensive applications need higher IOPS
2. Performance Requirements • DTU-based models measure combined CPU, memory, and I/O • vCore-based models allow independent scaling of compute and storage • Latency-sensitive applications require Premium or Business Critical tiers
3. Availability Needs • Business Critical tier provides 99.995% SLA with zone redundancy • General Purpose offers 99.99% SLA • Consider geo-replication requirements
4. Cost Optimization • Serverless is cost-effective for intermittent usage patterns • Reserved capacity offers discounts for predictable workloads • Auto-pause capabilities reduce costs during idle periods
Key Decision Criteria
Choose Serverless when: • Workloads are unpredictable or intermittent • Development and testing environments need flexibility • Cost optimization during idle periods is essential
Choose Provisioned when: • Workloads are consistent and predictable • Applications cannot tolerate cold-start delays • Maximum performance is required at all times
Choose Hyperscale when: • Database size exceeds 4 TB • Rapid scaling of compute is needed • Near-instantaneous backup and restore is required
Exam Tips: Answering Questions on Recommend a Database Service Tier and Compute Tier
• Look for keywords: Terms like 'unpredictable workload,' 'intermittent usage,' or 'development environment' typically point to serverless compute
• Identify cost sensitivity: When questions emphasize minimizing costs for variable workloads, serverless is usually the answer