Amortization of I/O Overhead¶

Fixed costs of I/O operations (seek time, protocol overhead, context switches) can be spread across multiple units of useful work, making larger operations proportionally more efficient.

Tension Addressed¶

Latency vs. Throughput

Recognizable As¶

Batch sizes, page sizes, block sizes, "vectored" I/O operations, any tuning parameter that groups operations

Canonical Sources¶

Patterson & Hennessy (storage section); database systems literature on buffer management

Transfers To¶

Network protocol design, database query optimization, batch processing systems, API design

This principle explains why databases read pages not individual records, why network protocols batch acknowledgments, and why compilers emit larger memory operations.