diskio

Disk input/output (diskIO) refers to the set of operations that move data between a computer system and its storage devices, such as hard disk drives, solid-state drives, or NVMe devices. I/O happens in terms of blocks and is managed by the operating system, the storage controller, and the device itself. The term is often used when discussing performance, latency, and throughput.

An I/O request originates when a program performs a read or write. The operating system converts that request into block I/O operations and queues them for the storage subsystem. An I/O scheduler determines the order, buffering and caching may hide latency, and the device driver communicates with the storage hardware. Data may pass through caches in memory and on the device, which can influence observed performance.

Key performance metrics include IOPS (input/output operations per second), bandwidth (MB/s or GB/s), and latency (time from issue to completion, typically microseconds to milliseconds). Hard disk drives are limited by mechanical factors such as seek time and rotational latency; solid-state drives and NVMe devices deliver much lower latency and higher parallelism, affected by queue depth and wear-leveling.

Storage systems employ I/O scheduling algorithms to optimize responsiveness under concurrent workloads. Caching, RAID configurations, and write policies (write-back versus write-through) influence perceived performance and data safety. Common measurement and tuning tools include iostat, iotop, fio, and similar utilities, used to diagnose bottlenecks and guide optimization.

Disk I/O is a critical factor in databases, virtual machines, file servers, and high-performance computing. Modern systems increasingly rely on fast solid-state storage and NVMe interfaces, often over PCIe or fabric-based connections, to improve concurrency and reduce latency. Effective tuning aims to balance throughput, latency, and data integrity.