Skip to content
DBM
Open menu

SQL Server engineering studio — available for new engagements

DBM — Duomenų Bazių Magija (“Database Magic”)

DBM makes SQL Server fast, reliable, and shippable.

DBM helps engineering teams cut query times, run migrations without surprises, and ship database changes through real CI/CD pipelines — with measurable before-and-after numbers, not vibes. A SQL Server engineering studio, currently a one-person practice, so you work directly with the engineer doing the work.

Book a free 30-min consult Browse services
Experience
17+ yrs
SQL Server
2012 → 2022
Open source
tSQLt
cpu-hotspots.sql 
-- Find the queries actually burning CPU on your server
SELECT TOP 5
    qs.execution_count,
    qs.total_worker_time / 1000 AS cpu_ms,
    qs.total_logical_reads      AS reads,
    SUBSTRING(qt.text, qs.statement_start_offset / 2 + 1,
        (CASE qs.statement_end_offset
              WHEN -1 THEN DATALENGTH(qt.text)
              ELSE qs.statement_end_offset
         END - qs.statement_start_offset) / 2 + 1) AS query_text
FROM sys.dm_exec_query_stats qs
CROSS APPLY sys.dm_exec_sql_text(qs.sql_handle) qt
ORDER BY qs.total_worker_time DESC;

Services

Three things DBM does well

The scope is tight on purpose — you get a senior engineer who has done the work, not a generalist Googling along with you.

Performance tuning & query optimization

Slow queries, blocking, runaway CPU, missing or bloated indexes — diagnosed and fixed with measurable before/after numbers.

Learn more

Migrations & upgrades

SQL Server version upgrades, server consolidations, and schema re-organisations — planned and executed without surprises.

Learn more

Database development & CI/CD

Schema as code, automated tests, and pipelines that ship database changes as confidently as application code.

Learn more

Why work with DBM

A senior engineer who shows numbers, not opinions

  • Hands-on, not slideware. DBM writes the queries, sets up the pipeline, and reruns the benchmark. You see the change live.
  • Measurable outcomes. Every engagement starts with a baseline and ends with before/after numbers.
  • Knowledge transfer by default. DBM leaves runbooks and dashboards so your team can keep going on its own.
  • Fixed-scope phases. No open-ended retainer dragging on — clear deliverables, clear stop points.

Writing

From the blog

performance t-sql benchmarks storage sparse-columns

Sparse columns: cheap NULLs, taxed non-NULLs, and where the bet actually pays

Benchmarked regular NULLable columns vs SPARSE on SQL Server 2025 across seven data types and six NULL fractions. Sparse stores a NULL for zero bytes but charges every non-NULL — so it only pays above a per-type break-even. I measured those break-evens (they come in BELOW Microsoft's documented thresholds for a wide block of sparse columns), the high-NULL storage payoff, the non-NULL tax, and the insert/read cost: a big cold-IO win where the table shrinks, a warm-scan decode penalty everywhere.

sql-server in-memory-oltp migration troubleshooting ssdt

Getting rid of an In-Memory OLTP filegroup you can't delete: four ways out

You inherited a database with a MEMORY_OPTIMIZED_DATA filegroup, zero in-memory tables, and no way to ALTER it away. Here are four production scenarios for actually removing it — a scripted rebuild, the SSDT/DACPAC shortcut, why Always On can't save you, and a minimal-downtime logical cutover — each reproduced end to end with receipts.

performance t-sql benchmarks storage compression

ROW vs PAGE compression: it depends on your data, and it can make inserts faster

Benchmarked NONE / ROW / PAGE data compression on SQL Server 2025 across five data shapes. There is no single compression ratio — ROW reclaims fixed-width waste, PAGE adds dictionary compression for repetitive data, and neither touches high-entropy data. Plus the counterintuitive bits: compression that speeds up inserts, the warm-vs-cold read tradeoff, and what random-key inserts actually do to a compressed table.

FAQ

Questions about DBM

What is DBM?
DBM (Duomenų Bazių Magija, “Database Magic”) is a Microsoft SQL Server engineering studio. It helps teams make SQL Server fast, run migrations and version upgrades without surprises, and ship database changes through real CI/CD pipelines — with measurable before-and-after numbers.
Who runs DBM?
DBM is currently a one-person studio led by its founder, Dmitrij Kultasev, a SQL Server engineer working with databases since 2008. Every engagement is handled directly by him — there is no account-manager layer and no handoffs.
What services does DBM offer?
Three things, kept deliberately tight: (1) performance tuning and query optimization, (2) SQL Server migrations and version upgrades, and (3) database development and CI/CD — getting a database into source control as an SSDT/DACPAC project with tSQLt tests and automated deployments.
What does DBM not do?
DBM is SQL-Server-only. It does not take on MySQL, PostgreSQL, or other engines, application/front-end work outside the database, or open-ended “data strategy” engagements with no measurable outcome.
Which SQL Server versions does DBM work with?
SQL Server 2012 through 2022, on-premises and in the cloud (AWS RDS, Azure SQL).
Where is DBM based and who can it work with?
DBM is based in Lithuania and works remotely with teams worldwide. Working languages are English, Lithuanian, and Russian.
How does an engagement start?
With a free 30-minute consult. You describe the symptoms; DBM tells you upfront whether it can help and roughly what it would take. Most work then starts as a small, fixed-scope first phase with a clear deliverable, so you can judge fit before committing further.

Got a slow query, a stalled migration, or a deploy story you don't trust?

The first 30 minutes are free. Bring the symptoms — DBM will tell you whether it's worth a deeper look.

Book a free consult See what DBM does