How do you do large TCO ? How long will take?

Google Cloud Compute Infrastructure
1

have a question regarding TCO estimation for a large-scale lift-and-shift migration scenario.

If you are provided with only one year of AWS or Azure billing invoices, without access to the customer’s architecture, utilization metrics, or workload details, what is the expected scope and methodology for delivering a Google Cloud TCO?

Specifically:

  1. Should the TCO be based purely on a 1:1 service mapping (e.g., EC2 → Compute Engine, Azure SQL → Cloud SQL) using equivalent instance types and configurations?

  2. Or is optimization (such as right-sizing, architecture improvements, or alternative services) expected, even though invoice data alone does not provide CPU, memory, or storage utilization metrics?

  3. Does the TCO deliverable typically include migration effort estimates (such as timeline, resource requirements, and migration complexity), or is it limited strictly to projected Google Cloud run-rate costs?

  4. For large environments with $5M–$10M annual cloud spend, what level of accuracy is expected from an invoice-only TCO, and how long is typically allocated to produce such an estimate?

  5. Is it standard practice to treat invoice-only TCO as a directional estimate, followed by a deeper discovery phase to produce an accurate and optimized TCO?

  6. Can we get it done this kid of TCO in few hours ? Is it realistic?

I would like to ensure alignment with Google Cloud best practices and expected delivery standards for large enterprise migration assessments.

Thank you.

2

Hi Mohamed,

If you only have invoice data and no visibility into architecture or utilization, the TCO would generally be treated as a directional estimate rather than a fully optimized one.

Typically, you’d begin with a 1:1 service mapping (for example, EC2 to Compute Engine) using comparable configurations. Without CPU, memory or storage utilization metrics, meaningful right-sizing or architectural improvements would involve assumptions and should be clearly framed that way.

In most cases, an invoice-based TCO focuses on projected Google Cloud run-rate costs. Migration effort, timelines and complexity usually require a proper discovery phase with architectural input.

For large environments, this kind of estimate is expected to be high-level. A few hours may be sufficient for a preliminary directional model, but not for a validated and optimization-driven assessment.

Thanks

3

Thanks for your reply. Just to confirm, do you mostly use AI tools to perform a 1-to-1 mapping and generate a high-level GCP run cost estimate?

In many cases, this can produce inaccurate or unfavorable cost comparisons. If you map Azure resources directly to GCP without optimization, GCP may appear more expensive especially since Azure includes bundled software licensing that is priced differently in GCP. In that scenario, the outcome can depend heavily on assumptions rather than actual optimization.

Meanwhile, other cloud providers may appear more cost-competitive in a simple 1-to-1 comparison.

Is that a fair understanding?

4

i would help but as a ne wdev its hard to understand can you illeterate im always happy to help :smiley::smiley: