A subnet hosts 100 VMs, and each VM may open up to 500 concurrent outbound connections through Cloud NAT. Which factor determines how many NAT IP addresses the gateway needs?
Show answer and explanation
Answer: C
Cloud NAT assigns translation ports to each VM from its NAT IP addresses. The minimum ports per VM sets the baseline each instance reserves, and dynamic port allocation can raise a VM's ports toward the configured maximum as demand grows. Each NAT IP provides roughly 64,512 usable ports, so the total reserved ports across all VMs, divided by that figure, gives a baseline address count. Adding headroom for bursts is prudent. The answer depends on these port settings rather than on a fixed ratio of VMs to addresses.
Why the other options are wrong
- A. NAT address needs are not tied to a fixed number of VMs per address. Port allocation and the usable ports per IP determine the count, so a per-instance ratio is the wrong model.
- B. Cloud NAT can use several external addresses, and one address is not always enough. With 100 VMs and high concurrency, the port demand can exceed what a single IP provides.
- D. Firewall rules decide which traffic is permitted into and out of instances. They do not allocate translation ports or addresses, so they have no effect on how many NAT IPs are needed.
Exam tip. Size Cloud NAT from ports: VMs times minimum ports, divided by about 64,512 per IP.