A site to site VPN over the internet is quick to stand up, encrypted and cheap, but bandwidth and latency depend on the public internet. A dedicated connection gives consistent throughput and lower latency at higher cost and a lead time of weeks. Many organizations run the dedicated link with VPN as automatic backup. Whichever you choose, plan non overlapping address ranges, routing, and DNS resolution in both directions before you start.
Why interviewers ask this
Hybrid connectivity is where planning failures become expensive, so the interviewer wants to know you think past the tunnel itself. They listen for address planning, since overlapping ranges are the classic disaster, plus routing propagation, DNS in both directions, and redundancy. Mentioning a hub such as a transit gateway shows you have dealt with more than one VPC needing the same connection.
How to structure your answer
- Present the two connection options with their trade offs.
- Recommend the redundancy pattern most teams end up with.
- Stress address planning and routing before anything else.
- Cover DNS resolution in both directions and how you scale to many VPCs.
Example answer
There are two real options. A site to site VPN goes over the internet, takes an afternoon, and is fine for modest, tolerant traffic, but you inherit whatever the internet is doing that day. A dedicated circuit gives predictable bandwidth and latency, which you need for chatty protocols or bulk transfer, but it costs more and provisioning takes weeks, so it is never the answer to an urgent problem. The common landing point is the dedicated link as primary with a VPN configured as automatic failover, which is cheap insurance for a circuit that will eventually be cut by somebody with a digger. Before any of that I want the addressing sorted, because if the on premises range overlaps the VPC range you are into network address translation workarounds forever. Routing has to propagate both ways and be filtered so I am not advertising more than intended. DNS is the part that gets forgotten: cloud workloads need to resolve internal names and the reverse, so I set up forwarding rules in both directions. Once more than a couple of VPCs need this, it all goes through a transit hub rather than a mesh of peers.
Walking into this interview soon? GhostPilot listens to your live call, spots the question the moment it is asked, and puts a structured answer on your screen in real time. Try it on your next mock, or grab a $29 Session Pass, no subscription, for the real thing.
See how it worksFollow-up questions to expect
- What do you do when the address ranges already overlap?
- How would you test the failover from the dedicated link to the VPN?
- How does a transit hub change your routing and cost model?
Related cloud engineer questions
Your interviewer will ask their own version of this. Paste your actual job description into the free Question Predictor and get the 20 questions that role is most likely to ask, with what each one is really probing.
Predict my questions