Skip to content
EncryptTechBD
Home Industries Our Work Blog Pricing About Contact Discuss Your Project
বাংলা (BN)
Cloud, DevOps & FinOps 6 min read

Cloud vendor lock-in risk: cost model and calculator

The cost lines behind lock-in, a risk-adjusted formula and a calculator for deciding how much portability is worth paying for.

E
EncryptTechBD · 0 comments
MareNostrum supercomputer racks in a glass enclosure
Photo: Gemmaribasmaspoch / Wikimedia Commons

Lock-in is usually discussed as a technology preference. It is a financial exposure. The question is not whether a service is proprietary; it is what leaving would cost, how likely you are to want to leave, and how much you pay every year to keep the option open.

This guide builds that comparison. It sets out the cost lines, a risk-adjusted formula, the inputs you need and a worksheet you can take to finance. It does not assume a particular cloud provider or quote rates, because egress, compute and support prices change and vary by contract.

The cost lines behind an exit

  • Data transfer and egress. Moving application data, backups, logs, media and archives out of the current provider. Volume and request patterns matter as much as the headline rate.
  • Extraction and rewriting. Exporting from proprietary databases, queues, identity systems and managed services into equivalents elsewhere.
  • Migration labour. Engineering, testing and project management time, including the internal people who know the system.
  • Parallel running. Running old and new environments at the same time while data is reconciled and traffic is cut over.
  • Downtime and business interruption. Lost orders, support load and contractual penalties if the migration overruns.
  • Retraining. Your team must learn new consoles, identity models, deployment tools and failure modes.
  • Contract commitments. Minimum spend, early termination, reserved capacity and support fees that survive the move.
  • Lost leverage. The discounts and credits that were tied to the current provider may not transfer.
  • Opportunity cost. Product work delayed while engineers migrate infrastructure.

The formula

One-off exit cost = data transfer + extraction and rewrite + migration labour + parallel running + downtime + termination costs + retraining.

Annual portability premium = abstraction and tooling + duplicate skills + portability testing + slower access to some managed features.

Risk-adjusted exposure = (likelihood of exit multiplied by impact of exit) + annual portability premium.

Use a 1–5 scale for likelihood and impact, with a written rationale for each score. The output is not a precise forecast. It is a way to compare two architectures, two contracts or two providers on the same basis.

Calculator inputs

  1. Data volume and growth. Current terabytes or gigabytes, plus a realistic annual growth rate. Include backups, logs and cold archives, not just the production database.
  2. Monthly egress. The data you already move out, and the data you would need to move during an exit. Check the provider’s current egress terms.
  3. Proprietary services in use. Count managed databases, queues, identity, serverless functions, AI APIs and observability tools that have no direct equivalent.
  4. Contract term and notice. Minimum commitment, renewal date, termination notice and any early-exit charge.
  5. Recovery time objective. How long the business can operate without the workload during a move.
  6. Environments. Production, staging, disaster recovery and analytics copies all add migration work.
  7. Team familiarity. How many people can operate the target platform without training, and how long training will take.
  8. Compliance and residency. Data-location, retention and audit requirements that constrain the destination.

Where the bill hides

Managed data services. A proprietary warehouse or document store may export data cheaply but require a rewrite of queries, pipelines and permissions.

Identity and permissions. Roles, policies and single sign-on integrations are easy to overlook and slow to reproduce safely.

Observability. Dashboards, alerts and log retention often use provider-specific agents and query languages.

AI and API services. A model endpoint may be replaceable, but prompts, evaluation data and latency expectations may not transfer cleanly.

Support and marketplace contracts. Third-party tools bought through the cloud marketplace may carry their own commitments.

Portability investments and what they buy

An export test. Restore a production-sized copy into a neutral environment once a year. This measures exit cost more honestly than an architecture diagram.

Open formats. Store data in formats that another system can read without a proprietary driver. Keep the schema and transformation logic documented.

Containerisation. Package workloads so the runtime is portable. This reduces migration effort but does not remove data, identity or network dependencies.

An abstraction layer. Hide provider-specific calls behind an internal interface. It adds code and maintenance, so use it where switching is plausible, not everywhere.

A second-provider pilot. Run one non-critical workload elsewhere for a quarter. The pilot exposes the real friction before an emergency.

Budget worksheet

  • Exit cost baseline: basis — current data volume, egress terms and service inventory. Owner — infrastructure lead. Evidence — costed migration plan.
  • Migration labour: basis — engineering days multiplied by loaded rate. Owner — engineering manager. Evidence — work breakdown and estimate.
  • Parallel run: basis — months of duplicate run cost. Owner — finance lead. Evidence — vendor quotes and run plan.
  • Downtime exposure: basis — hours of interruption multiplied by an agreed business-impact rate. Owner — business owner. Evidence — impact assessment.
  • Portability premium: basis — tooling, training and abstraction maintenance per year. Owner — architecture lead. Evidence — annual budget line.
  • Exit reserve: basis — agreed percentage of the exit estimate. Owner — finance lead. Evidence — approved reserve.

When not to spend on portability

Early-stage products with little data, short-lived workloads and commodity compute may get better value from speed than from optionality. A business with a clear regulatory reason to stay, a long contract and stable pricing may also choose to accept the risk and monitor it. The point is to make that choice explicitly, with a number and an owner, rather than discovering it during a renewal negotiation.

90-day plan

  1. Weeks 1–2: inventory data, services and contracts.
  2. Weeks 3–4: estimate exit cost and score likelihood and impact.
  3. Weeks 5–6: test an export of one production-sized dataset.
  4. Weeks 7–8: review contract notice, renewal and termination terms.
  5. Weeks 9–10: price the portability options you would actually use.
  6. Weeks 11–12: decide, document and assign an annual review.

Metrics to track

  • Exit cost as a share of annual cloud spend.
  • Number of proprietary services with no tested alternative.
  • Time to export and restore a production-sized dataset.
  • Contract notice period and renewal date.
  • Share of the team trained on a second platform.

If you need help measuring exit cost and designing a proportionate portability plan, start with Managed Technology & Cybersecurity. The pricing page explains how assessments and ongoing support are scoped. For the architectural choices behind the number, read four cloud vendor lock-in models compared.

Sources

Services that deliver this

Want this implemented properly?

Pricing →

Every engagement starts with a written scope. If a requirement is outside it, we quote it separately before starting.

Responses (0)

No comments yet. Add the first one below.

Leave a response

Your email address will not be published. Required fields are marked *.

Keep reading

More field notes