This repository provides a curated list of essential academic papers for understanding database systems and building new data infrastructure. It targets engineers, researchers, and students seeking foundational knowledge in areas like relational databases, distributed systems, and data processing. The benefit is a structured, high-quality reading list that accelerates learning and comprehension of core database concepts.

How It Works

The project is a curated collection of academic papers, organized into thematic sections such as "Basics and Algorithms," "Classic System Design," and "Data-Parallel Computation." Each paper is accompanied by a brief description highlighting its significance, key concepts, and relevance to modern data systems. This approach provides a guided path through complex topics, emphasizing foundational research and influential systems.

Quick Start & Requirements

This repository is a reading list and does not require installation or execution. All linked papers are publicly accessible academic publications.

Highlighted Details

• Covers foundational papers like Codd's relational model, ARIES, and System R.
• Includes influential distributed systems papers such as GFS, Bigtable, Dynamo, and MapReduce.
• Features key works on columnar databases (C-Store, Dremel) and consensus algorithms (Paxos, Raft).
• Provides links to external reading lists from top universities for further exploration.

Maintenance & Community

The list is curated and maintained by Reynold Xin (@rxin). Contributions are welcomed via pull requests. The repository also subtly mentions opportunities at Databricks.

Licensing & Compatibility

The repository itself is not licensed for software distribution. The licensing of the linked academic papers varies by publisher and author, and users should adhere to the terms of each publication.

Limitations & Caveats

This is a curated list of papers, not a software project. The descriptions are brief, and understanding the papers requires significant technical background and effort. Some papers may be dated, though their foundational concepts remain relevant.