cross cloud migration cost
Most cost-comparison pages give you on-demand list price math. Lumicost's Migration Analyzer pulls your real workload telemetry, queries live AWS and GCP pricing APIs (plus Azure Retail Prices), amortizes your Reserved Instances, Savings Plans and Committed Use Discounts on both source and target sides, and surfaces every AWS ↔ GCP ↔ Azure migration opportunity with explicit risk flags — so the savings number you see is the savings number you'd actually get.
Migration calculators are usually wrong because they ignore three things: live pricing (catalogues drift weekly), commitments (your effective rate is rarely on-demand), and risk (cross-region latency, special hardware, license portability). Lumicost's Migration Analyzer fixes all three. We stream the AWS Pricing API for EC2 and EBS, the GCP Cloud Billing API for Compute Engine + Persistent Disk, and Azure's Retail Prices API — then amortize your tenant commitments on both ends, score every candidate with confidence (HIGH / MEDIUM / LOW) and flag the risks you'd otherwise discover in production.
AwsLivePricingProvider streams the official AWS Pricing API (`pricing.us-east-1.amazonaws.com/offers/v1.0/aws/AmazonEC2`) for EC2 + EBS. GcpLivePricingProvider paginates the Cloud Billing API for Compute Engine + Persistent Disk. Azure uses the public Retail Prices API. Refreshed weekly via `PricingRefreshScheduler`, with seed-bundle fallback if a provider fails.
Register Reserved Instances, Savings Plans and CUDs at `/migration-analyzer/commitments`. The CommitmentDiscountResolver matches each candidate by provider + region + family + instance type, picks the best fit, and amortizes the discount onto monthly cost — for source AND target. Candidates flagged COMMITMENT_EXPIRES_SOON when within 90 days. The 'Committed' badge tells you which numbers include real discounts vs on-demand list.
Pulls the real shape of each workload (vCPU, memory, storage type and capacity) from your existing cloud connections. Scores every candidate against AWS, GCP and Azure instance families with a match score; HIGH confidence = match ≥ 0.85, same geo, no special HW. LOW = match < 0.65 or stale/unpriced. Storage cost is included in the comparison.
CROSS_REGION_LATENCY when the target is outside the source's geo group. SPECIAL_HARDWARE for GPU/Inferentia/TPU. LICENSE_PORTABILITY for Windows/SQL Server. COMMITMENT_AMORTIZATION_UNKNOWN if the source has no matching tenant commitment. EGRESS_FEES for data-heavy workloads. The Honesty Banner reminds you what the calculator can't see.
Run a single workload via `POST /api/v1/migration-analyses` or scan everything with `POST /run-all`. Summary KPIs at `/migration-analyzer`: annual savings, # of wins, # of high-confidence candidates, and which target cloud is winning most often. Filter by source provider (AWS / GCP / Azure / All) and confidence (wins / high) to focus on what's actionable this quarter.
Live, with seed fallback. AWS prices come from the public AWS Pricing API (`pricing.us-east-1.amazonaws.com`) — Linux + Shared tenancy + capacityStatus=Used + the family whitelist you configure. GCP prices come from the Cloud Billing API (Compute Engine service `6F81-5844-456A`) and we compose machine prices as `vCPU × CPU $/h + memGB × RAM $/h` per series (E2/N2/N2D/C3) by parsing SKU descriptions. If a provider fails, we fall back to the bundled seed so the calculator never returns nothing. Azure is live from Retail Prices.
Yes — and on both source and target. Register them at `/migration-analyzer/commitments` (provider, region, family, instance type, term, payment option, discount %, start / expiry dates). The analyzer caches matches per tenant for 60 seconds, picks the best fit by specificity + coverage fraction, and amortizes the discount onto each candidate's monthly cost. Candidates expiring within 90 days are flagged so you don't propose a migration whose 'savings' evaporate next quarter.
Three differences. (1) It's tenant-scoped: pulled from your actual workload telemetry, not whiteboard estimates. (2) It's continuously refreshed: pricing reloads weekly, commitments invalidate the cache on every mutation, and the analysis re-runs on demand. (3) It's honest: confidence scores, risk flags and the persistent Honesty Banner make explicit what the math can and can't see (egress fees, refactoring cost, license portability nuance).
AWS ↔ GCP ↔ Azure, any direction. EC2 + EBS on AWS, Compute Engine + Persistent Disk on GCP, equivalent compute + managed disk on Azure. The analyzer respects same-geo grouping by default and flags cross-region candidates explicitly. On the GROWTH plan you get the full analyzer; ENTERPRISE adds bulk run-all and the commitments management API.
Connect read-only credentials, get your first insights in 24 hours.