Compare VM instance families and naming conventions across providers.
Last verified: April 2026
Balanced compute, memory, and networking for a wide range of workloads including web servers, small databases, and dev/test environments.
m6i.xlargeD4s_v5n2-standard-4High-performance processors for compute-intensive tasks such as batch processing, HPC, gaming servers, and media encoding.
c6i.2xlargeF8s_v2c2-standard-8High memory-to-CPU ratios for in-memory databases, real-time big data analytics, and caching workloads.
r6i.2xlargeE8s_v5n2-highmem-8High sequential read/write access to very large datasets on local storage, ideal for data warehousing and distributed file systems.
i3en.xlargeL8s_v3n2-standard-8 + local SSDGPU-accelerated instances for machine learning training, inference, graphics rendering, and scientific simulations.
p4d.24xlargeNC6s_v3a2-highgpu-1gCost-effective instances that provide a baseline CPU with the ability to burst above it, suited for variable workloads.
t3.microB2s_v2e2-microARM processor-based instances offering better price-performance for scale-out and cloud-native workloads.
m7g.xlargeDps_v5t2a-standard-4Very high memory instances for large in-memory databases such as SAP HANA and other memory-intensive enterprise applications.
x2idn.24xlargeM128s_v2m2-ultramem-416Purpose-built accelerators for ML inference, training, and specialized compute including custom chips from each provider.
inf2.xlargeNP10sv5litepod-4[family][generation][attributes].[size]AWS instance names encode the family, generation, processor/attribute flags, and a T-shirt size separated by a dot.
mFamily: General Purpose (M), Compute (C), Memory (R), etc.6Generation: higher is newer (5, 6, 7...)iAttribute: i = Intel, a = AMD, g = Graviton (ARM), n = network optimizedxlargeSize: nano, micro, small, medium, large, xlarge, 2xlarge, ..., metal[Family][vCPUs][Addons]_v[Generation]Azure VM sizes use a letter family prefix, vCPU count, optional addons (s = premium storage, d = local disk), and a version suffix.
DFamily: D = General Purpose, F = Compute, E = Memory, L = Storage, N = GPU4vCPU count: the number directly following the family lettersAddon: s = premium storage capable, d = local temp disk, a = AMD, p = ARMv5Version: v2, v3, v4, v5 indicating hardware generation[family]-[type]-[vCPUs]GCP machine types use a family prefix, a type describing the memory profile, and the vCPU count separated by hyphens.
n2Family: n2 = general purpose, c2 = compute, e2 = cost-optimized, m2 = memory, t2a = ARMstandardType: standard, highmem, highcpu, ultramem, megamem4vCPU count: the number of virtual CPUs[
{
"category": "General Purpose",
"description": "Balanced compute, memory, and networking for a wide range of workloads including web servers, small databases, and dev/test environments.",
"aws": {
"family": "M-series (M5, M6i, M7i)",
"naming": "[family][generation][processor].[size] e.g. m6i.xlarge",
"example": "m6i.xlarge"
},
"azure": {
"family": "D-series (Ds_v5, Ds_v4)",
"naming": "[Family][vCPUs][addons]_v[generation] e.g. D4s_v5",
"example": "D4s_v5"
},
"gcp": {
"family": "N2 / N2D / N4",
"naming": "[family]-standard-[vCPUs] e.g. n2-standard-4",
"example": "n2-standard-4"
}
},
{
"category": "Compute Optimized",
"description": "High-performance processors for compute-intensive tasks such as batch processing, HPC, gaming servers, and media encoding.",
"aws": {
"family": "C-series (C5, C6i, C7i)",
"naming": "[family][generation][processor].[size] e.g. c6i.2xlarge",
"example": "c6i.2xlarge"
},
"azure": {
"family": "F-series (Fs_v2)",
"naming": "[Family][vCPUs][addons]_v[generation] e.g. F8s_v2",
"example": "F8s_v2"
},
"gcp": {
"family": "C2 / C2D / C3",
"naming": "[family]-standard-[vCPUs] e.g. c2-standard-8",
"example": "c2-standard-8"
}
},
{
"category": "Memory Optimized",
"description": "High memory-to-CPU ratios for in-memory databases, real-time big data analytics, and caching workloads.",
"aws": {
"family": "R-series (R5, R6i, R7i)",
"naming": "[family][generation][processor].[size] e.g. r6i.2xlarge",
"example": "r6i.2xlarge"
},
"azure": {
"family": "E-series (Es_v5, Es_v4)",
"naming": "[Family][vCPUs][addons]_v[generation] e.g. E8s_v5",
"example": "E8s_v5"
},
"gcp": {
"family": "N2-highmem / M3",
"naming": "[family]-highmem-[vCPUs] e.g. n2-highmem-8",
"example": "n2-highmem-8"
}
},
{
"category": "Storage Optimized",
"description": "High sequential read/write access to very large datasets on local storage, ideal for data warehousing and distributed file systems.",
"aws": {
"family": "I-series (I3, I3en, I4i)",
"naming": "[family][generation][variant].[size] e.g. i3en.xlarge",
"example": "i3en.xlarge"
},
"azure": {
"family": "L-series (Ls_v3, Ls_v2)",
"naming": "[Family][vCPUs][addons]_v[generation] e.g. L8s_v3",
"example": "L8s_v3"
},
"gcp": {
"family": "N2-standard + local SSD",
"naming": "[family]-standard-[vCPUs] with --local-ssd e.g. n2-standard-8 + local SSD",
"example": "n2-standard-8 + local SSD"
}
},
{
"category": "GPU",
"description": "GPU-accelerated instances for machine learning training, inference, graphics rendering, and scientific simulations.",
"aws": {
"family": "P-series / G-series (P4d, P5, G5)",
"naming": "[family][generation][variant].[size] e.g. p4d.24xlarge",
"example": "p4d.24xlarge"
},
"azure": {
"family": "NC-series (NCv3, NCas_T4_v3, NC_A100_v4)",
"naming": "NC[vCPUs][GPU-variant]_v[generation] e.g. NC6s_v3",
"example": "NC6s_v3"
},
"gcp": {
"family": "A2 / G2 (A100, L4)",
"naming": "[family]-[profile]-[GPUs]g e.g. a2-highgpu-1g",
"example": "a2-highgpu-1g"
}
},
{
"category": "Burstable",
"description": "Cost-effective instances that provide a baseline CPU with the ability to burst above it, suited for variable workloads.",
"aws": {
"family": "T-series (T3, T3a, T4g)",
"naming": "[family][generation][processor].[size] e.g. t3.micro",
"example": "t3.micro"
},
"azure": {
"family": "B-series (Bs_v2, B-series)",
"naming": "[Family][vCPUs][addons]_v[generation] e.g. B2s_v2",
"example": "B2s_v2"
},
"gcp": {
"family": "E2-micro / E2-small / E2-medium",
"naming": "e2-[size] e.g. e2-micro",
"example": "e2-micro"
}
},
{
"category": "ARM-based",
"description": "ARM processor-based instances offering better price-performance for scale-out and cloud-native workloads.",
"aws": {
"family": "Graviton (M7g, C7g, R7g)",
"naming": "[family][generation]g.[size] e.g. m7g.xlarge",
"example": "m7g.xlarge"
},
"azure": {
"family": "Ampere (Dps_v5, Eps_v5)",
"naming": "[Family]p[vCPUs]s_v[generation] e.g. Dps_v5",
"example": "Dps_v5"
},
"gcp": {
"family": "Tau T2A",
"naming": "t2a-standard-[vCPUs] e.g. t2a-standard-4",
"example": "t2a-standard-4"
}
},
{
"category": "High Memory",
"description": "Very high memory instances for large in-memory databases such as SAP HANA and other memory-intensive enterprise applications.",
"aws": {
"family": "X-series / u-series (X2idn, u-24tb1)",
"naming": "[family][generation][variant].[size] e.g. x2idn.24xlarge",
"example": "x2idn.24xlarge"
},
"azure": {
"family": "M-series (Ms_v2, Mv2)",
"naming": "M[vCPUs][addons]_v[generation] e.g. M128s_v2",
"example": "M128s_v2"
},
"gcp": {
"family": "M2-ultramem / M3-ultramem",
"naming": "[family]-ultramem-[vCPUs] e.g. m2-ultramem-416",
"example": "m2-ultramem-416"
}
},
{
"category": "Accelerated Computing",
"description": "Purpose-built accelerators for ML inference, training, and specialized compute including custom chips from each provider.",
"aws": {
"family": "Inf / Trn (Inf2, Trn1)",
"naming": "[family][generation].[size] e.g. inf2.xlarge, trn1.32xlarge",
"example": "inf2.xlarge"
},
"azure": {
"family": "NP-series (NP-series for FPGAs)",
"naming": "NP[vCPUs][addons] e.g. NP10s",
"example": "NP10s"
},
"gcp": {
"family": "TPU VM (v4, v5e, v5p)",
"naming": "tpu-[version]-[topology] e.g. v5litepod-4",
"example": "v5litepod-4"
}
}
]The Multi-Cloud VM Compare tool provides a side-by-side comparison of virtual machine instance families and naming conventions across AWS (EC2), Azure (Virtual Machines), and GCP (Compute Engine). It maps equivalent instance types across providers, explains naming schemes, and compares features like burstable instances, spot/preemptible pricing, and discount programs.
AWS uses family+generation+size (m6i.xlarge), Azure uses Family_Version (Standard_D4s_v5), and GCP uses family-type-vcpus (n2-standard-4). Each encodes CPU, memory, and feature information differently. This tool maps equivalent configurations across all three naming schemes.
AWS offers Reserved Instances (1/3 year, up to 72% off) and Savings Plans (flexible compute commitment). Azure has Reserved VM Instances with similar discounts. GCP provides Committed Use Discounts (1/3 year) and automatic Sustained Use Discounts (up to 30% off). The best option depends on commitment flexibility and workload patterns.
AWS Spot Instances offer up to 90% off with 2-minute interruption warning. Azure Spot VMs offer up to 90% off with configurable eviction policies. GCP Spot VMs (replacing preemptible) offer 60-91% off with 30-second warning and no 24-hour time limit. All three are best for fault-tolerant workloads.
Your company is negotiating a multi-cloud strategy and needs to justify why running the same 50-server workload might cost differently on each provider. You enter your standard configuration (8 vCPU, 32 GB RAM, general purpose) and the tool shows: AWS m6i.2xlarge at $2,803/month, Azure Standard_D8s_v5 at $2,748/month, and GCP n2-standard-8 at $2,220/month (after SUDs). But with 1-year commitments, the picture changes: AWS drops to $1,793 with Savings Plans, Azure to $1,648 with RIs, and GCP to $1,399 with CUDs. The comparison gives your procurement team concrete numbers for vendor negotiations.
The comparison tool maintains a mapping database of equivalent instance families across AWS, Azure, and GCP based on vCPU count, memory-to-CPU ratio, and workload category (general purpose, compute optimized, memory optimized). It normalizes pricing to a common per-hour and per-month basis, applying default discounts where applicable (GCP SUDs). Feature comparisons cover burstable behavior, spot/preemptible pricing, network bandwidth, and storage options for each equivalent instance across all three providers.
Instance type equivalence across clouds is approximate, not exact. AWS m6i.xlarge (4 vCPU, 16 GB) maps to Azure Standard_D4s_v5 (4 vCPU, 16 GB) and GCP n2-standard-4 (4 vCPU, 16 GB) — the specs look identical but CPU clock speeds, memory bandwidth, and network performance differ. Always benchmark your actual workload on each provider rather than assuming performance parity from matching specs.
GCP's automatic Sustained Use Discounts (up to 30% off for running all month) apply without any commitment, making GCP effectively cheaper than list price for always-on workloads. When comparing against AWS and Azure, use GCP's effective rate (after SUDs) not list price — otherwise you'll overestimate GCP costs by 20-30%.
For multi-cloud organizations, standardize on a common sizing taxonomy (small=2vCPU/8GB, medium=4vCPU/16GB, large=8vCPU/32GB) rather than provider-specific instance names. This makes cost comparison and workload migration straightforward and prevents teams from accidentally over-provisioning when they can't find an 'exact match' across providers.
Was this tool helpful?
Disclaimer: This tool runs entirely in your browser. No data is sent to our servers. Always verify outputs before using them in production. AWS, Azure, and GCP are trademarks of their respective owners.