Welcome to the Copy‑Paste Era: Rethinking Cloud‑Native Elasticity

Copy‑Paste as the Core of Cloud‑Native

Xu Hao explains that the fundamental capability of the cloud is the ability to copy an image and instantiate identical machines, which underlies elasticity.

From Scale‑Up to Scale‑Out

Traditional scaling replaced a single machine with a larger one (scale‑up). In the cloud era the preferred method is to add more identical instances (scale‑out), turning capacity problems into replication tasks.

Elasticity‑First Thinking

Identify “elastic boundaries” – groups of machines that should grow or shrink together. One practical way is to use business contexts (e.g., order processing during a sales event) as natural boundaries, leading to micro‑service decomposition.

Beyond Functional Boundaries

Even with unchanged functionality, different non‑functional requirements (SLA, NFR) can justify separate elastic groups. The speaker illustrates this with a courier‑service analogy and suggests grouping by lifecycle or performance needs.

Four Operational Modes

Coordination

Unification

Diversification

Replication – the cheapest operation in a cloud‑native world because the cloud’s sole capability is copying.

The talk concludes that embracing replication‑first thinking is essential to fully realize cloud‑native benefits.