09dACID

Database Management

Mick McQuaid
mcq@utexas.edu

University of Texas at Austin

13 May 2026

Computing Systems

  • One way to categorize most contemporary computing systems is that they are either data-intensive or compute-intensive
  • We are only interested in data-intensive systems in this course—by the way, a good resource on data-intensive systems is Kleppmann (2017)

Data-intensive systems

  • Data-intensive systems can be categorized in many ways, among them OLAP vs OLTP
  • OLAP stands for OnLine Analytic Processing Systems and are usually denormalized data warehouses
  • OLTP stands for OnLine Transaction Processing Systems and are what we are concerned with in this course
  • OLTP systems mostly rely on relational databases
  • These relational databases must obey four properties to work correctly
  • These four properties are usually known by the acronym ACID

The ACID Principles, 1 of 2

According to Gemini, the four ACID principles are

  • Atomicity (“All or Nothing”): Each transaction is treated as a single unit, which either succeeds completely or fails completely. If any part of the transaction fails, the entire transaction fails, and the database state remains unchanged
  • Consistency (Data Validity): Ensures that a transaction brings the database from one valid state to another, maintaining all predefined rules, constraints, and triggers

The ACID Principles, 2 of 2

  • Isolation (Independent Execution): Concurrent transactions do not interfere with each other. Even if multiple transactions occur simultaneously, they are isolated, so each behaves as if it were the only one in the system
  • Durability (Permanent Changes): Once a transaction is committed, it remains committed, even in the event of a system crash, power failure, or error. The changes are permanently stored, usually via write-ahead logging

Why ACID?

  • The database must maintain data integrity, even in the face of problems like concurrent users, power outages, hardware failures, and more
  • The ACID properties ensure data integrity and are the goal of any relational database management system
  • We haven’t been concerned with them in your projects for this class because you are running single user projects without concern for concurrency
  • We assume that you back up your data and you know the state of the data if there is a power failure or hardware failure or the like
  • In a multi-user setting, we can’t make these assumptions

BASE

  • Believe it or not, NoSQL databases coined the term BASE to be the opposite of ACID for their databases
  • BA = basically available
  • S = soft state
  • E = eventually consistent
  • BASE emphasizes Availability over Consistency

CAP

  • A further acronym you will encounter in more advanced databases is CAP
  • The CAP theorem says that a distributed database system can guarantee at most two of the three letters in CAP, which stand for consistency, availability, and partition tolerance
  • Note that distributed database systems are kind of the opposite of what we’ve been talking about in this course—they occur when the database is divided among many machines
  • Partition tolerance means the way the data is partitioned among machines to account for when communication between the machines fails

END

Colophon

This slideshow was produced using quarto

Fonts are Roboto, Roboto Light, and Victor Mono Nerd Font

Kleppmann, Martin. 2017. Designing Data-Intensive Applications. O’Reilly.