Scalability
The ability to handle growing load by adding resources, ideally with cost rising no faster than the load.
The word is used loosely, so pin it down: a system is scalable along a dimension if adding resources produces a proportional increase in capacity along that dimension. Scalable for reads and scalable for writes are different claims, and a system is routinely one and not the other.
Two shapes. Vertical — a bigger machine. Simple, immediate, no application change, and bounded by the largest machine available. Horizontal — more machines. Effectively unbounded, and it requires the work to be partitionable, which is the actual difficulty.
The limits are rarely hardware. They are shared mutable state, a single write primary, a lock, a hot partition, or a coordination step that every request must pass through. Amdahl's Law states it generally: the serial fraction of the work bounds the achievable speedup no matter how much parallelism you add.