Steps to Choosing the Right Sized Virtual Machines in Azure

1. Identify Your Current Workload:
   
1. Current Compute Capability: Determine the number of CPUs and the amount of RAM you are currently using for all Microvellum and other application for each of your servers and workstations.
2. Current Storage: Assess how much and what type of storage (Standard HDD, Standard SSD, Premium SSD) you are currently using for servers, workstations, and file storage. 
3. Current Network: Determine your current network bandwidth and latency.
4. Special Situations: Determine your current GPU capabilities, high IOPS storage, or other specific features.
5. Current Database:  Determine the following:
1. Database Size: The version of SQL (SQL Server 2019, etc.), current size of the database, and expected growth based on past metrics.
2. Concurrency: Number of concurrent users and transactions.
3. Performance: Current IOPS, CPU, and memory used.
   
2. Identify Workload Requirements You Would Like to Improve
   
1. Compute Requirements: Determine if the number of CPUs and the amount of RAM you are currently using is sufficient or if you would like to increase the number of CPU’s or amount of RAM for each of your servers and workstations.
2. Storage Requirements: Determine if you need more storage or better performance and then decide how much and what type of storage (Standard HDD, Standard SSD, Premium SSD) you would like to use to improve performance of your servers, workstations, and file storage. 
3. Network Requirements: Determine if your current network bandwidth and latency meets your needs or determine what improvements you would like to see.
4. Special Requirements: Determine if you would like to improve performance in GPU capabilities (Hardware Acceleration for AutoCAD, etc.), high IOPS storage, or other specific features.
5. Database Requirements: Determine if you need to improve Database Size, Concurrency, or performance.
   
3. Understand VM Families: Azure VMs are categorized into different families based on their intended use case:
   
1. General Purpose (B, Dsv3, Dv3): Balanced CPU-to-memory ratio. Suitable for web servers, small to medium databases, and development/test environments.
2. Compute Optimized (Fsv2, Fs): High CPU-to-memory ratio. Ideal for batch processing, web servers, and application servers.
3. Memory Optimized (Es, Ev3, M): High memory-to-CPU ratio. Best for relational databases, large caches, and in-memory analytics.
4. Storage Optimized (Ls, Lsv2): High disk throughput and IOPS. Suitable for big data, SQL, and NoSQL databases.
5. GPU (NC, NV, ND): Specialized VMs with GPU capabilities. Used for machine learning, AI, and graphics rendering.
6. High Performance Compute (H): High throughput and low latency network. Ideal for high-performance computing workloads.
7. Get a more detailed overview here:  https://learn.microsoft.com/en-us/azure/virtual-machines/sizes/overview
   
4. Use Azure VM Sizing Tools:
   
1. Azure Pricing Calculator: Use the Azure Pricing Calculator to estimate costs and compare different VM sizes.
2. Azure Advisor: After the setup of Azure, this tool provides recommendations on VM sizes based on your current usage.
3. Azure Migrate: For larger scale setups this tool can help assess and right-size VMs based on on-premises workloads.
   
5. Review Performance Benchmarks:
   
1. Azure VM Performance Benchmarks: Review performance benchmarks and case studies for different VM sizes to understand how they perform under various workloads.
2. Here are some performance benchmarks done by Microsoft: https://learn.microsoft.com/en-us/azure/virtual-machines/windows/compute-benchmark-scores
   
6. Budget Constraints:
   
1. Balance performance needs with budget constraints by selecting a VM size that provides the best value for money.
2. Reserved Instances: Consider using reserved instances if you have predictable workloads, as they offer significant cost savings over pay-as-you-go pricing.
   
7. Evaluate Different Sizes:
   
1. Pilot Testing: Deploy Microvellum and your application on different VM sizes in a test environment to determine which size meets your performance and cost requirements.
   

Example Deployment of Microvellum Production Environment to Azure 

1. SQL 2022 Standard on Windows Server 2022 (SQL preloaded)
   
1. VM Family Used: E-Series (Memory Optimized)
2. Machine Size: E4bds v5
   
1. Intel Xeon Platinum 8370C 2.8GHZ (32 Cores)
2. 4 vcpus
3. 32 GiB memory
   
3. OS Disk Type: Premium SSD
4. Microvellum Product Installed: Microvellum Server (Optional, not needed unless purchased)
2. Windows 11 Pro (Microvellum Production Order Entry and Processing Only)
   
1. VM Family Used: D-Series (General Purpose)
2. Machine Size: D4s v3
1. Intel® Xeon® 8171M 2.1GHz (32 Cores)
2. 4 vcpus
3. 16 GiB memory
3. OS Disk Type: Premium SSD
4. Microvellum Product Installed: Microvellum Production (GPU not needed due to no drawing).
3. Windows 11 Pro (GPU based VM)
   
1.  VM Family Used: N-Series
2.  Machine Size: NV4as v4
1. AMD EPYC 7v12 64-core processor (64 Cores)
2. 4 vcpus
3. 14 GiB memory
4. AMD Radeon Instinct MI25 Graphics Card
5. OS Disk Type: Premium SSD
3. Microvellum Product Installed: Toolbox OEM and Standard
   

Note: We do not recommend running Microvellum in Production from a local machine connected to Azure due to performance degradation and potential data loss.

Implementing Best Practices

Consult an IT Expert:

1. Work with an IT professional experienced in Azure to implement best practices for security, infrastructure, network configurations, firewalls, VPN Setup, storage and backup, Active Directory integration, etc.
   

1. Keep detailed documentation of your Azure setup for future reference and troubleshooting.