swapinout

Swapinout refers to the paired operations of moving memory pages between RAM and secondary storage, a core concept in virtual memory systems. It covers both swap-out (evicting a page from RAM to swap space) and swap-in (bringing a page from swap back into RAM). The process maintains address translation consistency by updating page tables and, when needed, flushing or updating TLB entries.

Swap-out is triggered by memory pressure or explicit reclaim, where the kernel selects a page to evict. Dirty pages are written to the swap area, which can be a dedicated swap partition or a swap file, and the page is marked as swapped to free the physical frame. Clean pages may be swapped out with minimal writes. The result is that RAM becomes available for active data while the swapped-out page remains accessible on storage.

Swap-in occurs when a swapped-out page is referenced, typically via a page fault. The operating system reads the page from swap into RAM, updates the page tables to reflect its present status, and may evict another page to satisfy memory constraints. The latency of swap-in is higher than RAM access, reflecting the slower I/O path to disk or other storage.

In practice, swapinout is a feature of most modern operating systems, including Linux, Windows, and macOS. Linux supports swap partitions and swap files, with tunable parameters such as swappiness that influence how aggressively swapping occurs. Some systems supplement or replace traditional swap with memory compression (zram) or swap caching (zswap) to improve performance.

Performance considerations are central to swapinout. Excessive swapping can cause thrashing, where page faults dominate CPU time. Mitigation strategies include adding RAM, adjusting swappiness, and using alternative memory technologies to reduce reliance on swap. See also virtual memory, paging, page fault, and swap space.