This article provides standardized VM SKU size for use when deploying new Azure VMs.

Key Governance Requirements

Villanova University has established a set of standard Azure Virtual Machine (VM) sizes to be used across the Azure tenant. These standards are intended to simplify provisioning, improve cost management, reduce over provisioning, and ensure long-term supportability by leveraging current-generation Azure VM families.

• Standard VM sizes must be used for all new Azure deployments whenever possible.
• Application owners should select the smallest VM size that reasonably meets business and technical requirements.
• Resource utilization should be monitored and adjusted as necessary after deployment.
• Requests for non-standard VM sizes require technical justification.
• Approval is required before deploying custom or specialized VM SKUs.

The standard sizing model defines four approved workload classifications:

• Production/Stage General Purpose Workloads – Use Das_v7-Series virtual machines for balanced compute and memory requirements.
• Production/Stage CPU Optimized Workloads – Use Fas_v7-Series virtual machines for processor-intensive applications.
• Production/Stage Memory Optimized Workloads – Use Eas_v7-Series virtual machines for memory-intensive applications and databases.
• Development/Test Workloads – Use cost-optimized Bas_v2-Series burstable virtual machines.

For each workload classification, administrators should select one of four standard sizes (Small, Medium, Large, or XL) based on the application's resource requirements. The approved sizes range from 2 vCPUs to 16 vCPUs and are designed to address the vast majority of Villanova's application hosting needs.

Standard Sizing Matrix

Unless a documented exception is approved, all Azure virtual machines should be deployed using one of the standard sizes defined in this document. This approach provides consistency, operational simplicity, predictable costs, and alignment with Villanova's Azure governance strategy.

 

SKU Size Technical Breakdowns

 

---

Production & Stage General Purpose Workloads

General purpose workloads should utilize AMD-based D-Series v6 or v7 virtual machines. These VMs provide balanced CPU and memory resources suitable for the majority of enterprise applications.

Standard Sizes

Typical Use Cases

Small

• Web servers
• Lightweight application servers
• DNS servers
• Small utility servers

Medium

• Departmental applications
• Middleware servers
• Small business applications

Large

• Enterprise application servers
• Application middleware platforms
• Medium-sized database workloads

XL

• Heavily utilized application servers
• Large middleware environments
• Enterprise services platforms

Custom/Other

• Requires written business and technical justification from the application owner.

---

Production & Stage CPU Optimized Workloads

Compute-intensive workloads should utilize F-Series virtual machines. These VMs provide a higher CPU-to-memory ratio and are intended for workloads where processor performance is the primary requirement.

Standard Sizes

Typical Use Cases

Small

• Lightweight batch processing
• Utility services
• Processing jobs

Medium

• Application processing engines
• API services
• Compute-heavy workloads

Large

• ETL processing
• Reporting engines
• Search and indexing services

XL

• Large-scale batch processing
• Analytics workloads
• High-performance compute applications

Custom/Other

• Requires written business and technical justification from the application owner.

---

Production & Stage Memory Optimized Workloads

Memory-intensive workloads should utilize E-Series virtual machines. These VMs provide a higher memory-to-CPU ratio and are intended for databases, caching platforms, and applications with significant RAM requirements.

Standard Sizes

Typical Use Cases

Small

• Small SQL Servers
• In-memory applications
• Application caching

Medium

• Departmental databases
• Reporting databases
• Application data services

Large

• Enterprise SQL workloads
• Large database servers
• Analytics platforms

XL

• High-performance database workloads
• Enterprise data platforms
• Memory-intensive business applications

Custom/Other

• Requires written business and technical justification from the application owner.

Development and Test Workloads

Development, testing, QA, sandbox, and non-production workloads should utilize Burstable B-Series virtual machines to minimize costs while providing adequate performance during periods of increased demand.

Standard Sizes

Typical Use Cases

Small

• Single application servers
• Utility servers
• Small web applications
• Lightweight Linux workloads

Medium

• Standard development environments
• Departmental applications
• Application testing environments

Large

• Multi-tier application testing
• Integration environments
• Large development databases

XL

• Enterprise application QA environments
• Performance testing environments

Custom/Other

• Requires written business and technical justification from the application owner.