pageSize

PageSize is a term used in computing to denote the size of a page, the basic fixed unit of data transfer and memory allocation in paged systems. It appears in operating system memory management, database storage, and various APIs that process data in fixed-size chunks. The value is usually expressed in bytes and is a fixed property of the environment, though some software offers it as a tunable parameter.

In hardware and operating system contexts, page size determines how virtual memory is divided into pages. Common sizes include 4 KiB, 8 KiB, and 16 KiB, with larger “huge” or “super” pages such as 2 MiB or 1 GiB used to improve performance for memory-intensive workloads. Smaller page sizes reduce waste when memory holds many small objects but increase the number of pages and page table entries. Larger page sizes lessen page table overhead but can increase internal fragmentation and TLB misses, affecting memory access latency and throughput.

In databases and data storage, a page (often called a block) acts as the unit of data storage and transfer. PageSize choices, such as 4 KB, 8 KB, or 16 KB, influence how data is laid out on disk, caching efficiency, and I/O behavior. They also affect how index and table data are organized and how write-ahead logging operates. Some database systems require or allow configuring the pageSize when the database is created.

Choosing a pageSize involves trade-offs among memory usage, I/O efficiency, and architectural constraints. Default values usually reflect hardware and system design, but tuning may be appropriate for specific workloads. See also virtual memory, TLB, page fault, and database pages.