Overview
Database researchers have been building data-access methods for decades, yet the same core question appears again and again: how do we minimize read cost, update cost, and memory or storage overhead at the same time?
As workloads diversify and hardware grows more heterogeneous and complex, small changes in environment can force major redesigns of access methods. What used to be a local tuning problem is now a central systems-design problem.
The RUM Conjecture frames this tension directly. It proposes that if we place strong bounds on any two of the three overheads, read (R), update (U), and memory (M), then the third overhead inherits a hard lower bound that cannot be pushed further down.
Why "RUM"?
RUM names the three overheads at the heart of access-method design: read, update, and memory. The conjecture captures the fact that improving any two of them inevitably puts pressure on the third.
Design Principle
Instead of viewing access methods as isolated structures, the RUM Conjecture offers a design lens for reasoning about tradeoffs across data systems.
