Have you heard about zero-copy cloning?

Zero-copy cloning is a technique used by operating systems to duplicate a process or a memory region without physically copying the underlying data. When a process is forked, the kernel marks the pages of the parent as read‑only and creates a copy‑on‑write (COW) mapping. Both parent and child share the same physical pages until one of them writes to a page, at which point the kernel copies that page to a new location. This avoids the overhead of copying large memory areas or disk blocks, making process creation and snapshotting much faster. Zero‑copy cloning is commonly used in container runtimes, virtual machine snapshots, and database replication. For example, Docker’s copy‑on‑write layers rely on the underlying filesystem’s COW support to efficiently create new images. In databases, zero‑copy cloning can be used to create read‑only replicas or to snapshot a database state for backup without blocking writes. The key benefit is reduced CPU and I/O usage, leading to lower latency and higher throughput in systems that need to create many lightweight copies of data.