To verify the real machine capacity (actual hardware resources vs. advertised specifications), follow these structured steps:

• Source: Check vendor documentation, cloud provider instance details, or purchase orders.
• Key Metrics:
  • CPU: Cores, threads, clock speed (GHz), hyper-threading.
  • Memory (RAM): Total capacity (GB).
  • Storage: Disk size (GB/TB), type (SSD/HDD), IOPS (for cloud).
  • Network: Bandwidth (Gbps), interface speed.

Verify Physical Hardware (On-Premises Machines)

Use OS tools to confirm actual hardware:

Linux:

• CPU:

  lscpu          # Core count, threads, architecture
  cat /proc/cpuinfo  # Detailed CPU info

• Memory:

  free -h        # Total RAM and usage
  dmidecode --type memory  # Detailed RAM specs

• Storage:

  lsblk          # Disk partitions and sizes
  fdisk -l       # Disk geometry
  smartctl -a /dev/sda  # Disk health (requires `smartmontools`)

• Network:

  ip a           # Network interfaces and speeds
  ethtool eth0   # Interface speed/duplex

Windows:

• CPU:

  systeminfo | findstr /C:"Processor"
  taskmgr → Performance tab → CPU

• Memory:

  systeminfo | findstr /C:"Total Physical Memory"
  taskmgr → Performance tab → Memory

• Storage:

  wmic diskdrive get size, model, interfaceType

• Network:

  ipconfig /all   # Interface details
  Get-NetAdapter -Name "Ethernet" | Select-Object LinkSpeed

Test Actual Performance (Benchmarking)

Use benchmarks to validate if hardware delivers expected performance:

• CPU:

  sysbench cpu --cpu-max-prime=20000 run  # Linux
  Geekbench (cross-platform) # Download from geekbench.com

• Memory:

  sysbench memory run --memory-block-size=1K --memory-total-size=10G  # Linux

• Disk I/O:

  dd if=/dev/zero of=testfile bs=1G count=1 oflag=direct  # Sequential write
  fio --name=randread --ioengine=libaio --iodepth=16 --rw=randread --bs=4k --direct=1 --size=1G  # Random read

• Network:

  iperf3 -c server_ip -t 60  # Client test (requires server)

Check Virtualization Overhead (VMs/Containers)

• Hypervisor Layer:
  • VMware: esxcli hardware cpu list (ESXi), vmware-cmd (vCenter).
  • KVM: virsh dominfo VM_NAME (CPU/memory limits).
  • Hyper-V: Get-VMProcessor -VMName "VM" (CPU limits).
• Guest OS: Use tools from Section 2 to detect allocated resources (not physical capacity).
• Cloud-Specific Checks:
  • AWS: curl http://169.254.169.254/latest/meta-data/instance-type (instance type).
  • Azure: az vm show -n VM_NAME -g RG_NAME --query hardwareProfile (CPU/RAM).

Detect Resource Contention

• Monitor Real-Time Usage:
  • Linux: htop, vmstat, iostat, nethogs.
  • Windows: Performance Monitor (perfmon), Resource Monitor.
• Cloud Tools:

  AWS CloudWatch, Azure Monitor, Google Cloud Monitoring.

• Key Metrics:
  • CPU/Memory utilization >80% indicates bottleneck.
  • Disk latency >10ms (SSD) or >50ms (HDD).
  • Network packet loss/jitter.

Validate Against Cloud Promises

• Cloud Providers:
  • AWS: Use ec2:InstanceTypes API or CloudWatch metrics.
  • Azure: Check Microsoft.Compute/virtualMachines API or Azure Monitor.
• Burstable Instances (e.g., AWS T2/T3):
  • Monitor CPU credits: aws cloudwatch get-metric-statistics --metric-name CPUUtilization --namespace AWS/EC2.
• Storage IOPS:
  • Cloud: Use fio or provider-specific tools (e.g., AWS io1 volumes).

Common Pitfalls

• Hyper-Threading: OS reports 2x cores (logical vs. physical). Use lscpu --topology (Linux) or Task Manager (Windows).
• Disk Cache: Benchmark tests may include cache effects. Use direct=1 in dd/fio to bypass cache.
• Network Throttling: Cloud providers cap bandwidth. Check instance type limits.
• Hardware Degradation: Use smartctl (Linux) or Windows Disk Health to check for failing drives.

Summary Workflow

1. Advertised Specs → Vendor docs/cloud console.
2. Physical Check → OS tools (lscpu, free, lsblk).
3. Benchmarking → sysbench, iperf3, fio.
4. Virtualization → Hypervisor APIs/cloud metadata.
5. Contention Check → Real-time monitoring (htop, CloudWatch).
6. Cloud Validation → Provider-specific APIs/tools.

By following these steps, you can accurately determine whether a machine’s real capacity matches its advertised specifications.