InfraAtlas.

The instrument that started it all. Cross-reference Amazon EC2 instance types against AWS regions. Filter by family, architecture, vendor. See which regions offer your instance type — and which instance types live in your region.

Every region from every provider on a single chart. Where is closest to your users — really? Surface clusters, compare hyperscaler geographies, and see the actual market overlap at a glance.

Azure VM sizes by region — the EC2 Observatory equivalent for the Microsoft cloud. Filter by series (Dsv5, Easv5, Lasv3), arch (Intel · AMD · Ampere · Cobalt), GPU class (T4 · A10 · A100 · H100 · H200), and confidential-compute capability.

Google Cloud machine types by region — E2 (cost) · N2/N4 (Intel) · N2D/T2D (AMD) · T2A/C4A (ARM) · C2/C3/C3D/C4 (compute) · H3 (HPC) · M1–M3/X4 (memory) · A2/A3/A4/G2 (NVIDIA T4 → L4 → A100 → H100 → B200).

Oracle Cloud's shape catalogue — E5/E4 Flex (AMD) · A1/A2 Flex (ARM Ampere) · Standard3 (Intel) · DenseIO (NVMe) · BM bare metal across all vendors · HPC2 + HPC.E5 with RDMA · GPU progression V100 → A10 → A100 → H100 → B200.

The European alternative. Public Cloud (b3/c3/r3/i1/t-series), Bare Metal (HGR/Advance/Rise/Scale), HPC, and GPU H100 — mapped across 15 regions from Gravelines to Sydney. EU-sovereign by default.

The cross-cloud translator. Pick any instance type — EC2, Azure VM, GCP, OCI or OVH — and see the closest match on every other cloud, scored on vCPU, memory and category. Built for migration planning and honest price comparison.

The cross-vendor comparator. Apigee · AWS API Gateway · Azure APIM · Mulesoft · Kong · Gravitee · IBM API Connect, all on one grid. OAuth flavours, mTLS, GraphQL, rate-limiting, self-hosted, FedRAMP — with the asterisks intact and a footnote on every nuance.

AWS API Gateway, demystified. REST vs HTTP vs WebSocket, the authorizer zoo, deployment modes, pricing — and the gotchas nobody explains until you hit them in production (29-second timeout, anyone?).

Apigee X, Hybrid, and Edge demystified. Policy library, Shared Flows + KVMs, environment topology, pricing tiers, migration paths — and a 12–24 month plan for getting off Edge before it sunsets.

Mulesoft Anypoint demystified. CloudHub 2.0 vs RTF vs Flex Gateway, the connector ecosystem, DataWeave (the secret weapon), pricing in vCore vs Flex Replica — and the integration platform DNA that sets it apart.

The three self-hostable contenders compared head-to-head. OSS scope, Kubernetes nativity, plugin ecosystems, total cost (license + ops + people) — and a decision tree for when to pick which (or none).

What AWS, Azure and Google Cloud charge to move a gigabyte, on one grid. Internet egress, cross-AZ, cross-region, NAT gateway processing and private interconnect — every rate sourced, dated, and read with the footnotes the pricing calculators bury.

A Rosetta stone for cloud networking — every core primitive mapped across AWS, Azure, Google Cloud and OCI. VPCs, subnets, peering, firewalls, load balancers and private links — with the semantic caveats spelled out per cell.

Which compliance programs each hyperscaler actually holds — FedRAMP, IRAP, C5, ENS, ISO 27001, SOC 2, PCI DSS and more — across AWS, Azure, GCP, OCI and OVHcloud. Scoped to region and service, never blanket-claimed.

Hardware-backed trusted execution, mapped across the major clouds — Intel SGX, AMD SEV-SNP, Intel TDX, AWS Nitro Enclaves and confidential GPU. Which cloud, which technology, which families, which regions.

Four hyperscaler IAM systems on one grid — AWS IAM, Azure (Entra ID + RBAC), GCP IAM and OCI IAM. Policy model, evaluation logic, workload-identity federation, short-lived credentials, the account hierarchy and ABAC — with the model spelled out per cell.

Which providers are actually sovereign — and at which layer. Operational vs legal sovereignty, the US CLOUD Act reaching US-parent “sovereign” regions, EU-native providers, and the regulatory frame — Data Act, EUCS, Schrems II / DPF — with the asterisks intact.

Managed Kubernetes, vendor by vendor — EKS · AKS · GKE · OKE · OVH. Control-plane SLA and cost, version cadence, node-pool limits, networking, autoscaling — fourteen capabilities with the footnotes that decide a platform choice.

The major cloud generative-AI platforms on one grid — Amazon Bedrock, Azure AI Foundry, Vertex AI, OCI Generative AI and OVHcloud AI Endpoints. Which model families each hosts — Claude, GPT, Gemini, Llama, Mistral — and how they compare on fine-tuning, RAG, agents and guardrails.

Cloud-native stacks and managed platforms compared across metrics, logs, traces, OTEL support, Kubernetes integration and pricing tiers. CloudWatch, Azure Monitor, GCP, OCI alongside Datadog, Grafana Cloud, New Relic, Honeycomb and Dynatrace.

Self-hosted open-source observability stacks compared across signal coverage, OpenTelemetry, Kubernetes, multi-tenancy and licensing. Prometheus, Grafana stack, Elastic Stack, OpenSearch, Jaeger and VictoriaMetrics — the third buying decision.

The Internal Developer Portals compared on what they ship, not what they pitch — Backstage, Spotify Portal, Roadie, Port, Cortex, OpsLevel and Atlassian Compass. Catalog, scaffolder, scorecards, plugins, RBAC — plus deployment and licensing as plain facts, every cell read off the official docs.

The default quotas you hit while scaling — on-demand vCPUs, function concurrency, VPCs, elastic IPs, buckets, IAM roles, cluster counts — across AWS, Azure, GCP and OCI, each cell sourced, with adjustable-vs-hard marked and where to raise each one.

Managed relational, NoSQL, and object storage compared across AWS, Azure, GCP and OCI — which service in Cloud B is the closest structural match to the one in Cloud A, and where the match breaks down.

What object storage really costs across S3, Azure Blob, GCS, OCI and OVHcloud — storage by access tier plus the asterisks nobody charts neutrally: retrieval fees, minimum-duration penalties, restore latency, egress and per-request charges. Every cell sourced.