SQL Server Upgrade Upgrade
Any SQL Server version to 2022 or 2025 — on-prem, Azure SQL, or a cloud VM. Jobs, linked servers, and cutover risk handled in a replica first.
SQL Server upgrade →SQL Server 2016 extended support ended 14 July 2026
Upgrade SQL Server, move a workload to or from it, or land the data — proven before production.
Upgrade Microsoft SQL Server, land a workload on it, move a workload to the engine you chose, or land the data on Fabric, Redshift, BigQuery, Snowflake, or Databricks. AI-assisted maps and tests; a named person signs before anything moves. We can run go-live.
Isolated replica
The move is proven away from production. The replica is part of the engagement.
Signed column map
AI-assisted STTM and lineage. A named person signs before anything moves.
Go-live we can run
We execute cutover, or your team runs the same tested runbook.
To SQL Server
PostgreSQL, MySQL, another SQL Server, Azure SQL, or RDS onto the SQL Server you named. Proven in a replica, then we can run production.
Migrate to SQL Server →From SQL Server
Move a SQL Server workload to Azure SQL, Amazon RDS or Aurora, Google Cloud SQL or AlloyDB, or PostgreSQL — proven in a replica, then we can run production.
Move SQL Server to another engine →Data platform
SQL Server data into Microsoft Fabric or Synapse, Amazon Redshift, Google BigQuery, Snowflake, or Databricks, with a runbook your team can operate.
SQL Server to a data platform →Mapping
Every in-scope column mapped. Lineage graphs. Fields that can be produced more than one way, flagged before go-live. Supports BCBS 239 evidence.
Source-to-target mapping and lineage →End of support
SQL Server 2016 extended support ended 14 July 2026. SQL Server 2012 and SQL Server 2014 are already out. SQL Server 2017 follows on 12 October 2027. Staying put means no security patches unless you buy Extended Security Updates — and ESU is a bridge, not a plan.
Microsoft · Amazon · Google · multi-cloud
Where the work can land
Same SQL Server source. Target is the Microsoft, Amazon, Google, Snowflake, or Databricks product you already chose. We implement that choice.
Microsoft
Azure / SQL Server
- SQL Server 2022 and 2025 (on-prem or Azure VM)
- Azure SQL Database and Azure SQL Managed Instance
- Microsoft Fabric and Azure Synapse Analytics
- Azure Databricks
Amazon
AWS
- Amazon RDS for SQL Server
- Amazon RDS for PostgreSQL and Amazon Aurora
- Amazon Redshift
Google Cloud
- Cloud SQL (SQL Server or PostgreSQL)
- AlloyDB
- BigQuery
Multi-cloud
Independent
- Snowflake
- Databricks
- PostgreSQL
Method
The method, in four steps
We rebuild the environment away from production, prove the path, then we can run go-live — or your team can, with the same artifacts.
- 01 · Intake
Inventory and a secure drop
NDA, instances, versions, jobs, and a transfer channel you control. No backups over email.
- 02 · Replica
Isolated rebuild
Schema and agreed artifacts are restored in a sandbox that is not production.
- 03 · Map and test
STTM, lineage, then the move
AI-assisted column map and lineage graph, dual-origin flags, human sign-off, then the upgrade or move in the replica — with tests tied to the map.
- 04 · Deliver
Go-live, or a runbook
We can execute production cutover from the tested path. Or we hand your team verified scripts, diffs, and a written go-live sequence. The SOW says which.
Questions
Common questions
What does an assessment cost?
The assessment request is how we start. Delivery is scoped in a written SOW — replica, mapping, and go-live as you need them. We do not publish a price list.
Do you run production go-live, or only hand over a runbook?
Either. We can execute production cutover from the path we proved in the replica, or your team can run the same tested runbook. The SOW says which.
What is AI-assisted here?
Software drafts the inventory, source-to-target map, lineage graph, and recon tests. A named person signs the map before anything moves. We do not skip your UAT, and software does not press go-live unattended.
How do you know the data will match?
In-scope columns get a source-to-target map. We graph how values flow and flag fields that can be produced more than one way. Tests are tied to that map. For banks, those artifacts support BCBS 239 evidence; we do not certify your programme.
Which platforms do you land on?
The one you chose: a supported SQL Server release, Azure SQL, Amazon RDS or Redshift, Google Cloud SQL or BigQuery, Snowflake, or Databricks — including the pipeline pattern that platform expects.
Can you migrate a workload onto SQL Server, not only off it?
Yes. PostgreSQL, MySQL, another SQL Server, Azure SQL, or RDS for SQL Server onto a SQL Server you named — isolated replica, signed map, then we can run go-live. Oracle is in scope when we can prove it in scoping.
Do I have to move to the cloud?
No. A supported SQL Server release on-prem or on an Azure, AWS, or Google VM is a valid target. If you already chose Azure SQL, Amazon RDS, or another engine, we implement that choice.
How long does this take?
Calendar time depends on instance count, Agent jobs, SSIS, and HA. The assessment reply says what we would need next. We do not publish a one-size duration.
Request an assessment
Version or source engine, target, and timeline. We reply with a path — upgrade, to SQL Server, from SQL Server, data platform — or a clear no. No public prices, no newsletter.
Request an assessment