Azure Local Hardware Requirements & Planning

Table of Contents

  1. Overview
  2. Hardware Requirements Overview
    1. Minimum vs. Recommended Specifications
    2. Server Form Factor
  3. Physical Topology and Network Architecture
    1. Cluster Hardware Topology
  4. Approved Hardware Partners and Systems
    1. Tier 1 OEM Partners
    2. Tier 2 System Builders
    3. Benefits of Validated Hardware
  5. CPU, Memory, and Storage Specifications
    1. Processor Requirements
    2. Memory Requirements
    3. Storage Requirements
  6. Network Adapter Requirements
    1. Minimum Requirements
    2. Recommended Configuration
    3. RDMA Technologies
    4. Popular Network Adapters
    5. Network Configuration Best Practices
  7. Redundancy and High Availability Hardware Setup
    1. Node Redundancy
    2. Storage Redundancy
    3. Network Redundancy
    4. Power Redundancy
    5. Cooling and Environmental
  8. Sizing Guidance Based on Workload Types
    1. General Virtualization
    2. VDI (Virtual Desktop Infrastructure)
    3. Database Workloads
    4. AI/ML and GPU Workloads
    5. File Server / Storage-Heavy
    6. Edge / Branch Office
  9. Power and Cooling Requirements
    1. Power Consumption
    2. Power Requirements
    3. Cooling Requirements
  10. Rack and Deployment Considerations
    1. Rack Requirements
    2. Cable Management
    3. Physical Security
  11. Upgrade Paths and Future-Proofing
    1. Upgrade Strategies
    2. Technology Trends
    3. Future-Proofing Recommendations
  12. Cost Analysis of Hardware Investment
    1. Initial Hardware Costs
    2. Software Licensing
    3. Ongoing Costs
    4. Total Cost of Ownership (TCO) - 5 Years
  13. Deployment Checklist
    1. Pre-Deployment (2-4 weeks)
    2. Deployment Week 1
    3. Deployment Week 2
    4. Deployment Week 3
    5. Deployment Week 4
  14. Next Steps
⏱️ Reading Time: 15-20 min 🎯 Key Topics: Validated hardware, sizing, vendor selection 📋 Prerequisites: Azure Local Overview

Overview

Selecting the right hardware is critical for Azure Local success. This page provides comprehensive guidance on hardware requirements, validated partners, sizing strategies, and deployment planning.

Key Topics:

  • Hardware specifications and requirements
  • Validated hardware partners and systems
  • Sizing guidance based on workload types
  • Network adapter and storage requirements
  • Deployment planning checklist

← Back to Azure Local Overview

Hardware Requirements Overview

Minimum Configuration (Testing/POC Only):

  • 1 physical server
  • 8-core CPU
  • 128 GB RAM
  • 4 drives minimum
  • 2x 10 GbE NICs
  • Not recommended for production

Production Recommended:

  • 2-4 physical servers (for HA)
  • 16+ core CPU per server
  • 384+ GB RAM per server
  • 8+ drives per server (mix of NVMe, SSD, HDD)
  • 2x 25 GbE or faster NICs (RDMA-capable)
  • Validated hardware from Microsoft partners

Server Form Factor

Rack-Mount Servers (Most Common):

  • 1U or 2U form factor
  • Standard 19-inch racks
  • Hot-swap components
  • Redundant power supplies
  • Remote management (iDRAC, iLO, etc.)

Blade Servers:

  • Higher density
  • Shared infrastructure (power, cooling)
  • More complex networking
  • Limited expandability

Tower Servers:

  • Suitable for branch offices
  • Lower cost entry point
  • Limited scalability
  • Not recommended for multi-node clusters

Physical Topology and Network Architecture

Cluster Hardware Topology

Approved Hardware Partners and Systems

Microsoft maintains a catalog of validated Azure Local solutions from OEM and system builder partners.

Tier 1 OEM Partners

Dell Technologies:

  • Product Line: PowerEdge R650, R750
  • Configuration: 2U, 2-socket, Intel or AMD
  • Strengths: Comprehensive support, global availability
  • URL: Dell Azure Local Solutions

Hewlett Packard Enterprise (HPE):

  • Product Line: ProLiant DL380, DL385
  • Configuration: 2U, 2-socket, Intel or AMD
  • Strengths: iLO management, InfoSight integration
  • URL: HPE Azure Local Solutions

Lenovo:

  • Product Line: ThinkSystem SR650, SR850
  • Configuration: 2U, 2-socket, scalable to 4-socket
  • Strengths: Competitive pricing, XClarity management
  • URL: Lenovo Azure Local Solutions

Fujitsu:

  • Product Line: PRIMERGY RX2530, RX2540
  • Configuration: 2U, 2-socket, Intel Xeon
  • Strengths: European manufacturing, strong EMEA presence
  • URL: Fujitsu Azure Local Solutions

Tier 2 System Builders

Supermicro:

  • Custom-built solutions
  • Wide range of configurations
  • Cost-effective options
  • Strong for specialized workloads (GPU, high-density)

DataON:

  • Azure Local-optimized systems
  • Pre-configured solutions
  • Direct support relationship with Microsoft
  • Focus on storage performance

QCT (Quanta Cloud Technology):

  • ODM solutions
  • Competitive pricing
  • Hyperscale heritage
  • Strong in Asia-Pacific

Wortmann AG:

  • European system integrator
  • TERRA servers
  • Strong DACH region presence
  • Local support

Benefits of Validated Hardware

Pre-Tested and Certified:

  • Microsoft-validated configurations
  • Guaranteed compatibility
  • Firmware and driver validation
  • Performance benchmarking

Simplified Support:

  • Single point of contact
  • Coordinated troubleshooting
  • Faster resolution times
  • Clear escalation paths

Optimized Performance:

  • Tuned for Azure Local workloads
  • RDMA configuration validated
  • Storage performance validated
  • Power efficiency optimized

Reference: Azure Local Solutions Catalog

CPU, Memory, and Storage Specifications

Processor Requirements

Supported Processors:

  • Intel: Xeon Scalable (Skylake or newer)
  • AMD: EPYC 7002 series or newer
  • 64-bit x86 architecture only

Required Features:

  • Virtualization Extensions: Intel VT-x or AMD-V
  • Second Level Address Translation (SLAT): Intel EPT or AMD RVI
  • Hardware-Assisted Virtualization: Mandatory
  • Intel VT-d / AMD-Vi: For device passthrough

Core Count Recommendations:

Small Deployments (< 50 VMs):

  • Minimum: 8 cores per CPU, 16 cores per node
  • Recommended: 12-16 cores per CPU, 24-32 cores per node

Medium Deployments (50-200 VMs):

  • Recommended: 16-24 cores per CPU, 32-48 cores per node

Large Deployments (200+ VMs):

  • Recommended: 24-32 cores per CPU, 48-64 cores per node

Special Workloads:

  • VDI: 32+ cores per node
  • AI/ML: High core count + GPUs
  • HPC: Maximum cores (64+)

Memory Requirements

Minimum: 128 GB DDR4 ECC
Recommended Production: 384 GB - 1 TB
Large Deployments: 1 TB - 6 TB

Memory Sizing Formula:

Total Memory = (Base OS + Storage Spaces Direct + VMs + Overhead)

Where:
- Base OS: 8-16 GB per node
- Storage Spaces Direct: 2-4 GB per TB of cache + 1 GB per TB of capacity
- VMs: Sum of VM memory allocations
- Overhead: 20% buffer for spikes

Example Calculation (4-node cluster):

  • Base OS: 12 GB × 4 = 48 GB
  • S2D (10 TB cache, 100 TB capacity): 40 GB + 100 GB = 140 GB
  • VMs: 200 VMs × 8 GB average = 1,600 GB
  • Overhead: 20% = 320 GB
  • Total: 2,108 GB (divide by 4 nodes = 528 GB per node)

Memory Type:

  • DDR4 or DDR5: ECC (Error-Correcting Code) required
  • Speed: 2666 MHz or faster recommended
  • Configuration: Populate all memory channels for maximum bandwidth
  • NUMA: Be aware of NUMA node boundaries for VM placement

Storage Requirements

Drive Types:

OS Drives (Required):

  • 2x 240 GB+ SATA SSD (mirrored)
  • Used for host OS and boot
  • Separate from Storage Spaces Direct pool
  • M.2 or SATA form factor

Cache Tier (Recommended):

  • NVMe SSD for optimal performance
  • 2-8x drives per node
  • 800 GB - 3.2 TB per drive
  • PCIe 3.0 x4 or better
  • Enterprise-grade (high endurance)

Capacity Tier:

  • SSD or HDD based on performance needs
  • 4-16x drives per node
  • 1-16 TB per drive
  • SATA or SAS interface

Minimum Drive Configuration:

  • 4 drives per node (minimum)
  • All nodes must have same drive configuration
  • Mix of drive types (NVMe + SSD or NVMe + HDD)

Recommended Drive Configuration:

Performance-Optimized:

  • 2-4x NVMe SSD (cache): 1.6-3.2 TB each
  • 6-10x SATA SSD (capacity): 4-8 TB each
  • Use Case: Databases, VDI, latency-sensitive apps

Balanced:

  • 2x NVMe SSD (cache): 1.6 TB each
  • 4-6x SATA SSD (capacity): 4 TB each
  • 4-6x SATA HDD (capacity): 8-12 TB each
  • Use Case: General virtualization

Capacity-Optimized:

  • 2x NVMe SSD (cache): 800 GB each
  • 10-12x SATA HDD (capacity): 12-16 TB each
  • Use Case: File servers, archives, backup

Storage Sizing Formula:

Usable Capacity = (Total Raw Capacity × Efficiency Factor × Resiliency Factor)

Efficiency Factors:
- Deduplication/Compression: 1.5-3x (varies by data type)
- Without dedupe: 1x

Resiliency Factors:
- Two-way mirror: 0.5 (50% efficiency)
- Three-way mirror: 0.33 (33% efficiency)
- Parity (erasure coding): 0.5-0.8 (depends on configuration)

Example:

  • 4 nodes × 10x 4TB SSD = 160 TB raw
  • Two-way mirror: 160 TB × 0.5 = 80 TB usable
  • Three-way mirror: 160 TB × 0.33 = 53 TB usable

Network Adapter Requirements

Minimum Requirements

Adapter Count: 2 adapters (for redundancy)
Speed: 10 GbE minimum
Features: SR-IOV, RDMA (for storage traffic)

Management/Compute:

  • 2x 10 GbE or 25 GbE adapters
  • Bonded/teamed for redundancy
  • Standard Ethernet (no special requirements)

Storage:

  • 2x 25 GbE or faster (40/50/100 GbE)
  • RDMA-capable (iWARP, RoCE v2, or InfiniBand)
  • Dedicated for Storage Spaces Direct traffic
  • DCB-capable (for RoCE)

RDMA Technologies

iWARP (Internet Wide Area RDMA Protocol):

  • Works over standard Ethernet
  • Easier to configure (no DCB required)
  • Slightly higher CPU usage
  • Recommendation: Good choice for most deployments

RoCE v2 (RDMA over Converged Ethernet):

  • Higher performance than iWARP
  • Lower CPU usage
  • Requires DCB (Data Center Bridging)
  • More complex network configuration
  • Recommendation: Best for large, high-performance clusters

InfiniBand:

  • Highest performance
  • Separate network fabric required
  • Less common in Azure Local
  • Higher cost
  • Recommendation: Specialized use cases only

Mellanox/NVIDIA:

  • ConnectX-5, ConnectX-6
  • 25/40/50/100 GbE
  • Excellent RDMA performance
  • Strong software support

Intel:

  • E810-C series
  • 25/40/100 GbE
  • Good iWARP support
  • Broad compatibility

Broadcom:

  • BCM57xxx series
  • 25/50/100 GbE
  • Enterprise-grade reliability
  • Good OEM support

Network Configuration Best Practices

NIC Teaming/Bonding:

  • Use Switch Independent mode
  • Dynamic load balancing
  • Redundant paths to switches

Jumbo Frames:

  • Enable for storage network (MTU 9000+)
  • Reduces CPU overhead
  • Improves throughput

Quality of Service (QoS):

  • Prioritize storage traffic
  • Bandwidth reservation
  • Prevent starvation

VLAN Segmentation:

  • Separate management, storage, compute traffic
  • Improves security
  • Reduces broadcast domains

Redundancy and High Availability Hardware Setup

Node Redundancy

2-Node Clusters:

  • Tolerates 1 node failure
  • Requires file share witness (for quorum)
  • 50% capacity overhead
  • Most cost-effective HA configuration

3-Node Clusters:

  • Tolerates 1 node failure
  • Better performance than 2-node
  • 33% capacity overhead
  • Good balance of cost and resilience

4-Node Clusters (Recommended):

  • Tolerates 1 node failure
  • 25% capacity overhead
  • Better performance distribution
  • Most common configuration

5-16 Node Clusters:

  • Tolerates 1-2 node failures (depends on resiliency)
  • Lower overhead percentage
  • Higher aggregate capacity
  • For large deployments

Storage Redundancy

Drive-Level Protection:

  • Hot-spare capability
  • Automatic rebuild
  • Multiple drive failures tolerated
  • RAID not used (Storage Spaces Direct handles this)

Node-Level Protection:

  • Data replicated across nodes
  • Two-way or three-way mirror
  • Erasure coding (parity) option
  • Configurable per volume

Network Redundancy

Adapter-Level:

  • NIC teaming (bonding)
  • Active-passive or active-active
  • Automatic failover

Switch-Level:

  • Dual top-of-rack switches
  • Each node connected to both switches
  • Eliminates single point of failure

Path-Level:

  • Multiple paths to storage
  • Multipath I/O (MPIO)
  • Automatic path failover

Power Redundancy

Node-Level:

  • Dual power supplies per server
  • Each PSU on separate circuit
  • Automatic failover on PSU failure

Facility-Level:

  • Dual power feeds to rack
  • UPS for graceful shutdown
  • Generator for extended outages
  • Power distribution units (PDUs)

Cooling and Environmental

Cooling:

  • Hot aisle / cold aisle configuration
  • Adequate airflow (front-to-back)
  • Temperature monitoring
  • Redundant cooling units

Environmental Monitoring:

  • Temperature and humidity sensors
  • Water leak detection
  • Smoke detection
  • Integration with facility management

Sizing Guidance Based on Workload Types

General Virtualization

Typical Profile:

  • Mixed Windows and Linux VMs
  • Moderate I/O requirements
  • Standard enterprise applications

Recommended Configuration (per node):

  • CPU: 2x 16-core (32 cores total)
  • Memory: 512 GB
  • Storage: 2x NVMe (1.6 TB) + 6x SSD (4 TB)
  • Network: 2x 25 GbE (RDMA)

Expected Capacity:

  • 40-60 VMs per node
  • 160-240 VMs per 4-node cluster

VDI (Virtual Desktop Infrastructure)

Typical Profile:

  • Many small VMs (2-4 GB each)
  • Burst I/O patterns
  • Graphics requirements (optional)

Recommended Configuration (per node):

  • CPU: 2x 24-core (48 cores total)
  • Memory: 768 GB - 1 TB
  • Storage: 4x NVMe (3.2 TB) + 6x SSD (4 TB)
  • Network: 2x 25 GbE (RDMA)
  • GPU: Optional (NVIDIA T4 or similar)

Expected Capacity:

  • 100-150 VDI sessions per node
  • 400-600 VDI sessions per 4-node cluster

Database Workloads

Typical Profile:

  • High IOPS requirements
  • Low latency critical
  • High memory usage

Recommended Configuration (per node):

  • CPU: 2x 20-core (40 cores total)
  • Memory: 768 GB - 1.5 TB
  • Storage: 4-8x NVMe (3.2 TB) - All-flash
  • Network: 2x 25 GbE or faster (RDMA)

Expected Capacity:

  • 10-20 database VMs per node
  • 500K+ IOPS per cluster
  • Sub-millisecond latency

AI/ML and GPU Workloads

Typical Profile:

  • GPU-accelerated compute
  • Large memory requirements
  • High storage bandwidth

Recommended Configuration (per node):

  • CPU: 2x 24-core (48 cores total)
  • Memory: 1 TB - 2 TB
  • Storage: 4x NVMe (3.2 TB) + high-bandwidth network storage
  • Network: 2x 25 GbE or faster
  • GPU: 2-4x NVIDIA A100 or H100

Expected Capacity:

  • 5-10 ML training jobs concurrent
  • 50-100 inference workloads

File Server / Storage-Heavy

Typical Profile:

  • High capacity requirements
  • Moderate performance needs
  • Deduplication and compression

Recommended Configuration (per node):

  • CPU: 2x 16-core (32 cores total)
  • Memory: 384 GB
  • Storage: 2x NVMe (800 GB) + 12x HDD (12-16 TB)
  • Network: 2x 25 GbE (RDMA)

Expected Capacity:

  • 500 TB - 1 PB usable (with parity)
  • 3-5 GB/s throughput

Edge / Branch Office

Typical Profile:

  • Small footprint (1-2 nodes)
  • 10-30 VMs
  • Local services (AD, DNS, DHCP, file shares)

Recommended Configuration (per node):

  • CPU: 1-2x 12-core (12-24 cores total)
  • Memory: 256-384 GB
  • Storage: 2x NVMe (800 GB) + 4x SSD (2 TB)
  • Network: 2x 10 GbE

Expected Capacity:

  • 15-25 VMs per node
  • 30-50 VMs for 2-node cluster

Power and Cooling Requirements

Power Consumption

Per Node Estimates:

Small Node:

  • Idle: 150-250 watts
  • Average: 300-500 watts
  • Peak: 600-800 watts

Medium Node:

  • Idle: 250-350 watts
  • Average: 500-800 watts
  • Peak: 1000-1200 watts

Large Node (with GPUs):

  • Idle: 350-500 watts
  • Average: 800-1200 watts
  • Peak: 1500-2000 watts

4-Node Cluster (Medium):

  • Idle: 1-1.4 kW
  • Average: 2-3.2 kW
  • Peak: 4-4.8 kW
  • Plus networking gear: +300-500 watts

Power Requirements

Electrical:

  • 208V or 240V (most efficient)
  • Dedicated circuits recommended
  • PDUs with monitoring
  • Surge protection

UPS (Uninterruptible Power Supply):

  • Size for 15-30 minutes runtime
  • For 4-node cluster: 5-10 kVA UPS
  • Includes networking gear
  • Allows graceful shutdown

Generator (Optional):

  • For extended outages
  • Size for sustained load
  • Automatic transfer switch (ATS)
  • Regular testing required

Cooling Requirements

Heat Output:

  • Power consumption = heat output (BTU/hr = Watts × 3.412)
  • 4-node cluster average: 2.5 kW = 8,500 BTU/hr
  • Plus switch gear and UPS losses

Cooling Capacity:

  • Size for peak load, not average
  • Include overhead (15-20%)
  • 4-node cluster: 10,000-12,000 BTU/hr cooling required

Airflow:

  • Front-to-back or back-to-front (depends on system)
  • 250-500 CFM per server
  • Hot aisle / cold aisle configuration
  • Ensure no recirculation

Temperature:

  • Maintain 18-27°C (64-80°F)
  • Below 32°C (90°F) maximum
  • Monitor inlet and outlet temperatures
  • Alert on thermal excursions

Humidity:

  • Maintain 20-80% relative humidity
  • 45-55% optimal
  • Monitor and alert

Rack and Deployment Considerations

Rack Requirements

Standard 42U Rack:

  • 19-inch width (standard)
  • 600-800mm depth (depends on server depth)
  • 1000mm width (for cable management)
  • Cable management arms
  • PDU mounting

Space Requirements:

  • 1-2U per node (typical)
  • 1-2U per switch
  • 2-4U for UPS (or separate rack)
  • 2-4U for patch panels and cable management
  • Total: 8-16U for 4-node cluster

Cable Management

Best Practices:

  • Use cable management arms
  • Label all cables clearly
  • Separate power and data cables
  • Use color-coding for cable types
  • Document cable routing

Cable Types:

  • Power: C13/C14 or C19/C20
  • Management: Cat6 or better
  • Storage: DAC (Direct Attach Copper) or fiber
  • Compute: DAC or fiber (for 25G+)

Physical Security

Rack Security:

  • Lockable rack doors
  • Side panels secured
  • Serial number tracking
  • Tamper-evident seals

Facility Security:

  • Badge access
  • Video surveillance
  • Visitor logs
  • Environmental monitoring

Upgrade Paths and Future-Proofing

Upgrade Strategies

Vertical Scaling (Scale-Up):

  • Add memory to existing nodes
  • Upgrade CPUs (same generation)
  • Add drives to storage pool
  • Upgrade network adapters
  • Advantage: No new hardware purchase
  • Limitation: Socket and slot constraints

Horizontal Scaling (Scale-Out):

  • Add nodes to cluster (up to 16)
  • Automatic rebalancing
  • Increase aggregate capacity
  • Advantage: More flexibility
  • Limitation: 16-node max

Cluster Refresh:

  • Replace entire cluster with newer hardware
  • Migrate VMs to new cluster
  • Decommission old cluster
  • Advantage: Latest technology
  • Limitation: Significant investment

Processor Evolution:

  • Higher core counts per socket
  • More memory channels
  • PCIe 5.0 and beyond
  • CXL (Compute Express Link)

Storage Evolution:

  • NVMe over Fabric (NVMe-oF)
  • PCIe 5.0 SSDs (higher bandwidth)
  • Persistent memory (PMem)
  • QLC and PLC NAND (higher density)

Networking Evolution:

  • 100 GbE and 200 GbE adapters
  • Faster RDMA (RoCE v3)
  • SmartNICs and DPUs
  • Quantum networking (long-term)

Future-Proofing Recommendations

1. Plan for Growth:

  • Size for 3-5 year horizon
  • Leave room for expansion
  • Choose upgradeable platforms

2. Standardize on Latest Generation:

  • Don’t buy previous-gen hardware
  • Invest in PCIe 4.0 or 5.0
  • Use DDR5 if available

3. Modular Design:

  • Separate clusters by workload
  • Use Azure Arc for unified management
  • Enable workload mobility

4. Software-Defined Everything:

  • SDN for networking flexibility
  • Software-defined storage
  • Policy-based management
  • Reduces hardware lock-in

Cost Analysis of Hardware Investment

Initial Hardware Costs

Small Deployment (2-node, branch office):

  • Servers: $15,000 - $25,000 per node = $30,000 - $50,000
  • Networking: $5,000 - $10,000
  • Rack and infrastructure: $5,000 - $10,000
  • Total: $40,000 - $70,000

Medium Deployment (4-node, production):

  • Servers: $25,000 - $50,000 per node = $100,000 - $200,000
  • Networking: $15,000 - $30,000
  • Rack and infrastructure: $10,000 - $20,000
  • Total: $125,000 - $250,000

Large Deployment (8-node, enterprise):

  • Servers: $40,000 - $80,000 per node = $320,000 - $640,000
  • Networking: $50,000 - $100,000
  • Rack and infrastructure: $20,000 - $40,000
  • Total: $390,000 - $780,000

Software Licensing

Windows Server Datacenter:

  • Per-core licensing
  • ~$6,000 per 16-core license
  • 2-socket server = $12,000 - $24,000

Azure Local Subscription:

  • Per-core pricing
  • ~$10 per core per month
  • 4-node cluster (32 cores each): $1,280/month

Ongoing Costs

Annual Maintenance (Hardware):

  • 15-20% of hardware cost per year
  • Includes parts replacement
  • On-site support

Power and Cooling:

  • Electricity: $0.10 - $0.20 per kWh
  • 4-node cluster: 2.5 kW average = $2,200 - $4,400/year
  • Cooling: ~1.5x power consumption

Staffing:

  • Varies greatly by organization
  • Plan for 0.5-2 FTE per cluster
  • Training and certification costs

Total Cost of Ownership (TCO) - 5 Years

Medium Deployment (4-node) Example:

  • Initial hardware: $150,000
  • Licensing (one-time): $50,000
  • Azure Local subscription (5 years): $76,800
  • Maintenance (5 years): $150,000
  • Power (5 years): $15,000
  • Total 5-Year TCO: $441,800

Per VM Cost:

  • Assuming 200 VMs
  • $441,800 / 200 VMs / 5 years = $442/VM/year
  • Compare to public cloud (varies widely by workload)

Break-Even Analysis:

  • Typically 2-3 years vs. public cloud
  • Faster for 24/7 workloads
  • Slower for sporadic workloads

Deployment Checklist

Pre-Deployment (2-4 weeks)

Planning:

  • Define workload requirements
  • Size cluster appropriately
  • Select hardware vendor
  • Design network topology
  • Plan IP addressing
  • Document architecture

Procurement:

  • Order hardware
  • Order network gear
  • Procure licenses
  • Order rack and PDUs
  • Plan delivery logistics

Facility Preparation:

  • Verify power capacity
  • Verify cooling capacity
  • Prepare rack location
  • Install power circuits
  • Install network drops

Deployment Week 1

Physical Installation:

  • Rack servers
  • Install network switches
  • Connect power cables
  • Connect network cables
  • Install management station
  • Verify all connections

Initial Configuration:

  • Configure BIOS settings
  • Enable virtualization features
  • Configure RAID for OS drives
  • Configure BMC/iDRAC
  • Verify hardware inventory

Deployment Week 2

Operating System:

  • Install Windows Server
  • Apply latest updates
  • Configure networking
  • Join to domain
  • Install Hyper-V role
  • Install required features

Cluster Creation:

  • Validate cluster hardware
  • Create failover cluster
  • Configure cluster networking
  • Test cluster failover

Deployment Week 3

Storage Spaces Direct:

  • Enable Storage Spaces Direct
  • Verify all drives
  • Create storage pools
  • Create volumes
  • Test storage performance
  • Configure resiliency

Azure Integration (if Connected Mode):

  • Register with Azure Arc
  • Configure monitoring
  • Set up backup
  • Apply policies
  • Test connectivity

Deployment Week 4

Validation:

  • Run validation tests
  • Performance benchmarking
  • Failover testing
  • Backup and restore test
  • Security hardening
  • Documentation review

Handover:

  • Train operations team
  • Provide runbooks
  • Transfer to operations
  • Schedule follow-up

Next Steps

Continue Learning:

External Resources:

Last Updated: October 2025