Database Systems: SQL, Indexing, Query Optimization, Transactions and Distributed Databases by CMU Database Group

About this course

Strong database knowledge is what turns basic CRUD into systems that stay fast, correct, and reliable under real-world load. This free online course helps you move beyond writing queries and into understanding how database engines actually work: how data is modeled, stored, found quickly with indexes, executed efficiently by the optimizer, and protected with transaction guarantees even when many users hit the system at once.

You will start with the relational model and the ideas that made it durable in industry, then connect those foundations to modern SQL used in production analytics and applications. From there, you will go under the hood to see why storage layout matters, why sequential access often beats random I/O, and how structures like log-structured merge trees and slotted pages shape performance characteristics. As you progress, you will learn how and why databases choose row or column formats, how memory and buffer management influence latency, and what tradeoffs appear when systems lean on the operating system versus owning their caching behavior.

Indexing and execution are treated as practical engineering topics: hash tables and tree indexes, what makes B+ trees effective for range scans, and how alternative structures such as inverted indexes or Bloom filters can accelerate specific workloads. You will also develop intuition for core operator algorithms like sorting and joins, then connect that to query execution models and parallel execution components used to scale out work across cores and machines. Along the way, exercises reinforce concepts so you can reason about why one plan is faster than another, not just memorize rules.

Correctness is as important as speed. The course explains concurrency control theory and brings it to life with two-phase locking, optimistic methods, and MVCC, showing how databases deliver isolation and consistent reads. You will then cover durability with write-ahead logging and recovery concepts, including how systems restore a consistent state after failures. Finally, you will explore distributed database architectures and replication patterns, learning the constraints and advantages of shared-disk versus shared-nothing designs and how transactional and analytical distributed systems make different tradeoffs.

By the end, you will be able to evaluate database behavior with an engineer’s mindset: predict bottlenecks, choose the right index strategy, understand what the optimizer is trying to do, and speak confidently about transactions, logging, and distributed designs in interviews and on the job.