— Lakehouse migration · Warehouse to lakehouse

Databricks migration services that move the warehouse, not the problems.

Snowflake, Teradata and on-prem warehouses to a Databricks lakehouse, and Hive metastore to Unity Catalog. Our Databricks migration services translate what is worth keeping, run old and new side by side, and cut over only when the numbers match.

Databricks Migration Services: Migrating Snowflake, Teradata and Hive metastore workloads to the Databricks lakehouse with dual-run and valid

30+ years inside IBM · EY · J&J · McKesson

Lean Six Sigma Master Black Belt

AIMSolve from $2,500

60-Day Ship Guarantee

— The pattern · What breaks

Most migrations copy the mess to a new address.

Everything translated.

Years of unused objects, duplicate logic and reports nobody opens get translated faithfully. The bill moves, the confusion stays.

Symptom: the new platform costs more to run than the old one.

Cutover before reconciliation.

The cutover date arrives before reconciliation is finished, and the first month on the new platform is spent explaining why revenue differs.

Symptom: two revenue figures and a steering committee.

No way back.

The old system is switched off on day one. When something breaks, the only option is to fix forward under pressure.

Symptom: every incident becomes a weekend.

— The work · What we do

What our Databricks migration services actually cover.

Estate survey

Lakebridge, the free open-source Databricks Labs toolkit, surveys the existing estate so we can see what is used before anything is moved.

SQL translation, reviewed

Lakebridge translation treated as a first draft that engineers review. It is a Labs project, not a GA product with an SLA, and we plan accordingly.

Snowflake and Teradata

Workloads moved from Snowflake, Teradata and on-prem data warehouses onto the lakehouse, redesigned where the old design was the problem.

Hive metastore to Unity Catalog

UCX, the Databricks Labs tool, assesses the workspace and upgrades Hive metastore tables into Unity Catalog, where governance and lineage now live.

Dual run and reconciliation

Old and new run in parallel, reconciled table by table, with the results written down before anyone signs off.

Rollback plan

A written rollback plan held for every wave until it is signed off, so a bad night is a revert, not a crisis.

Databricks Migration Services: two engineers comparing validation reports side by side during a data platform cutover

— The approach · Retire before you translate

Decide what to retire before anything is translated.

The cheapest object to migrate is the one you switch off. We start with an inventory of every table, job and report, mark what is actually used, and agree what to retire, what to redesign and what to move as is.

Only then does translation start, one workload at a time, reconciled against the source before the next wave begins. The lakehouse you land on is smaller and cleaner than the warehouse you left.

— The method · AIM-IT

Five phases. Every build, every time.

Assess, Innovate, Model, Implement, Track. Each phase ends with something you can inspect and sign off, not a slide.

01 · ASSESS

Assess

Object and usage inventory

Every object, job and report inventoried, with what is actually used marked.

02 · INNOVATE

Innovate

Retire, redesign, translate

A decision per workload: retire it, redesign it, or translate it as is.

03 · MODEL

Model

One workload, reconciled

One workload migrated end to end and reconciled against the source.

04 · IMPLEMENT

Implement

Waves with dual run

The rest moved in waves, dual-running each until the numbers match.

05 · TRACK

Track

Sign-off and retirement

The old platform retired only after sign-off, with the rollback plan kept until then.

— The deliverable · What you get

A migrated lakehouse you can prove is right.

AIMSolve

Migration assessment and a first workload, fixed scope.

AIMSolve is the fixed-scope start to a migration. You get:

  • An inventory of what exists, what is used and what can be retired.
  • A first workload running on the lakehouse, governed in Unity Catalog.
  • Table-level reconciliation from a dual run, written down.
  • A rollback plan held until each wave is signed off.
  • Your team able to run it. We are not trying to become permanent.
Entry: AIMSolve · migration assessment and first workload, fixed scope. Starting at $2,500.

$2,500+

Starting price

1st

Workload migrated and reconciled

5 phases

AIM-IT, end to end

60 days

Ship guarantee

60-Day
Ship Guarantee

On every Sprint engagement: if we do not deliver the agreed working artifact in 60 days, you do not pay the final invoice. That is what AIM-IT is for.

— Unity Catalog · What changes on 30 September 2026

Hive metastore to Unity Catalog, on your timetable.

From 30 September 2026, Databricks provisions new workspaces without Hive metastore, DBFS root and mounts, no-isolation shared clusters, or runtimes below 13.3 LTS. Existing workspaces are not affected, and no end date has been set for their Hive metastore.

So there is no forced deadline for most teams. There is a good reason to move anyway: Unity Catalog is where governance, lineage and access control now live on the lakehouse, and every new Databricks feature assumes it.

Databricks Migration Services: migration plan printed on a table with a phased timeline and checklists

— Related reading · From the blog

Go deeper on lakehouse migration.

Snowflake

A cutover plan with dual run and rollback built in.

Teradata

What to retire, redesign and translate from Teradata.

Unity Catalog

The UCX path and the parts that break.

Governance

What Unity Catalog governs and why it matters.

Concepts

What actually changes when a lake becomes a lakehouse.

Cost

Cost ranges by scope, before you ask for a quote.

Cost

Where the money goes in a warehouse exit.

Platform

The platform build the migration lands on.

— Straight answers · FAQ

Questions we get asked first.

Are you a Databricks partner or certified?

No, and we will not imply otherwise. Certification is in preparation, and we are not in the Databricks partner programme. What we bring is thirty years of delivering the enterprise systems this data comes out of, inside IBM, Ernst & Young, Johnson & Johnson and McKesson, plus Lean Six Sigma at Master Black Belt level. Ask us for the work, not the badge.

Should we move from Snowflake to Databricks?

Not always. It depends on your workload mix, your team and what your BI layer already reads. We work on both platforms and will tell you when staying on Snowflake is the better answer.

Does Lakebridge support Teradata?

Yes. Lakebridge supports Teradata among its source dialects, alongside Snowflake, Oracle, SQL Server and others. It speeds up survey and translation. It does not replace the decisions about what to retire and how to prove the numbers match.

Is Databricks ending Hive metastore support?

Not for existing workspaces. From 30 September 2026 Databricks provisions new workspaces without Hive metastore, and no end date has been set for existing ones. Moving to Unity Catalog is still where governance, lineage and access control now live.

What do Databricks migration services cost?

AIMSolve starts at $2,500 for a fixed-scope assessment and first workload. Full migrations are quoted after the assessment, because an honest number needs to see your estate first.

Start with the inventory, not the cutover date.

One call, your current warehouse on the table, and a straight read on what should move to the lakehouse and what should be retired. If we are not the right team, we will say so.