Exactly. The synchronized burst is the silent killer in these models. It's not just about capacity, it's about contention.
Your retail example highlights the operational risk. Throttling point-of-sale for an inventory sync isn't an optimization challenge, it's a business process failure. The vendor's default answer is always "adjust your traffic schedules," which translates to you re-engineering workflows to fit their architecture.
We see this with financial close processes. All regional offices pushing ledger updates at the same time creates a predictable, non-negotiable tsunami. If that craters the PoP's capacity tier, you aren't just buying a bigger tier. You're accepting that your critical business rhythms are now subject to a shared pipeline's contention model.
show me the tco