Resources Case Studies

Upgrading 3,500 SQL Server databases in 128 days

Hundreds of SQL Server 2012 databases were weeks from end of life, across an estate of 3,500 of them and the 300 applications sitting on top. The deadline was not movable, so the plan had to be right before any of it was touched.

Upgrade and migration

128 days

SQL Server 2012 to SQL Server 2019, data encrypted, and every application coordinated through the move.

Databases upgraded and migrated
3,500
Applications moved with them
300
Service lost during the move
None

In short

What happened

A large North American insurance company found that hundreds of its Microsoft SQL Server 2012 databases were approaching end of life, which meant no more security patches, no updates and no extended support from Microsoft. Fortified Data ran a data tier assessment, established which parts of the estate carried the outage risk, and prepared a side-by-side upgrade along a critical path of system integrations, dependencies, application supportability and encryption. 3,500 databases and 300 applications were upgraded to SQL Server 2019 and migrated to the cloud in 128 days, with no loss of service for users and no application failure.

3,500

SQL Server databases upgraded from SQL Server 2012 to SQL Server 2019 and migrated to the cloud.

Fortified Data case study, 2025

128 days

From the start of the engagement to the last database migrated, with no loss of service for users and no application failure.

Fortified Data case study, 2025

300

Applications migrated alongside the databases, each with an owner to coordinate and its own dependencies to clear first.

Fortified Data case study, 2025

What end of life meant at this scale

When a SQL Server version reaches end of life, the software stops receiving security patches, updates and extended support contracts from Microsoft. For this insurer that meant upgrading its SQL Server 2012 environments to SQL Server 2019, migrating the data, encrypting it, and coordinating the move with more than 300 unique applications in the data ecosystem. Because end of life was imminent, all of it had to happen quickly.

The scope was the difficulty. 3,500 SQL Server databases, hundreds of applications, many of them mission-critical and complex, and 300 application owners to coordinate. The insurer set a window of three months, which left very little room for error in testing and in the upgrades themselves.

An insurance organization handles large amounts of highly sensitive data supporting several mission-critical systems, and thousands of users depend on the applications sitting on it. An upgrade that took a database offline unexpectedly would not have been a database problem for long.

The engagement, in order

Why the order mattered

In any data environment there are areas of strength and areas needing improvement. The work is isolating the few that will have the most impact on what the client is trying to do, and doing them in the sequence that keeps the rest of the estate running.

  1. Assess the whole estate first

    Fortified Data completed a thorough data tier assessment before committing to a plan. Creating a detailed plan for execution that prioritizes the right actions is what makes it possible to deliver better results more quickly, and at 3,500 databases there was no version of this that could be improvised.

    The plan every later decision was prioritized against.
  2. Establish priority across the environment

    The assessment established priority for the data environment specifically to mitigate the risk of critical outages. Which systems could tolerate a window and which could not is the first thing a three-month deadline needs answered, because it decides what runs in parallel and what does not.

    Outage risk ranked before any database was touched.
  3. Prepare the side-by-side upgrade

    The upgrade ran side by side rather than in place, along a stated critical path: system integrations first, then external and application dependencies, then application supportability, then encryption support. Each one gates the next, which is why it is a path rather than a checklist.

    A critical path four items long, with encryption on it rather than after it.
  4. Design the right service for the right application

    With the path set, the team designed the right service for each application, right-sized the environment for capacity planning, and improved the encryption across it. Right-sizing at this point rather than after the migration is what turns an upgrade into a reduction in what there is to patch and maintain.

    Capacity planned against measured need rather than against existing allocation.
  5. 128 days

    Upgrade, encrypt and migrate

    Fortified Data's DBAs upgraded every server to SQL Server 2019 and migrated 3,500 databases and 300 applications to the cloud. Data clean-up and improved encryption went with them, so performance and security both improved across the environment rather than being restored to where they had been.

    3,500 databases and 300 applications migrated, with no loss of service for users and no application failure.

3,500

What the sequence moved

SQL Server databases upgraded from 2012 to 2019, encrypted, and migrated to the cloud alongside 300 applications, in 128 days. The insurer set a three-month window; the assessment is what made a schedule of that order arguable at all.

What changed

  • Every server on SQL Server 2019The end-of-life exposure closed, and with it the security patches, updates and support contracts that stop arriving when a version goes out of support.
  • 3,500 databases and 300 applications in the cloudMigrated in 128 days, with no loss of service for users and no application failure during the move.
  • Performance and security both improvedData clean-up and improved encryption ran as part of the migration rather than after it, so the environment came out of the project in better condition than it went in.
  • A higher return on the estateThere is less to patch and less to maintain, and the upgraded architecture freed cost that was available for capacity planning.

Three ways to work with us

On-Call DBA

A guaranteed 15-minute emergency response plus monthly on-demand hours that bank when unused.

Read more

Is a version going end of life on you?

The useful first question is not how long the upgrade takes, it is which parts of the estate carry the outage risk and what depends on them. That is what a data tier assessment answers, and it is what made a 128-day schedule arguable here rather than optimistic.

Let us show you what's possible.

Two ways to start

Bring the version and the deadline.