I applied for a position at Amazon as an NDE. A recruiter reached out to me with details about the interview process and scheduled my interview.

The interview was a 1-hour phone call and it was quite difficult. There were no simple questions or definition-based questions; it was all troubleshooting.

1. Tell me about your current role, current responsibility, and daily activities.

2. Which part of the network do you handle? What kind of network is it? Can you elaborate?

3. What are the high-level steps for migrating devices from the 6500 series to the 9600 series?

4. Detail the migration plan. On the night of the migration, what exactly did you do to migrate, and what was the customer interruption?

5. Let's say old cables are connecting to a 6500 chassis, and traffic is flowing towards it. You want to migrate to a 9500 chassis. How would you route the traffic from the 6500 to the 9500? Would you move the cable from the 6500 directly to the 9500 chassis, or would you have a plan to offload the traffic from the 6500? How would you smoothly move the traffic from the 6500 to the 9500 with no or minimal impacts? How would you move the traffic from the 6500 to the 9500 with respect to configuration? How would you configure it?

6. What is the default behavior of a route reflector when receiving multiple copies of the same route from different sources/different routers? Let's say we have a route reflector, and two or more RR clients are sending the same route. Which route would the RR reflect and why? Answer: The RR would reflect the best route only to its other RR clients.

   Let's say if I want the RR to reflect the same route received from different routers, how would we achieve that? Load balancing... Answer: Command...

7. Let's say we have customer A (Customer Edge) in MPLS with two sites. Both are running BGP and are in the same autonomous system. Would this topology work?

   Answer: as-overwrite command.

   Let's say I have a CE that's multihomed with two different ISPs. How can we influence the traffic between the two ISPs?

   Answer: Local Preference or MED (MED is more practical).

   Which type of MED would we be using? There are two types of MED. Can you tell the difference between them? Answer: ________ MED is used for the same AS.

   When do we apply local preference configuration, and how do we do that?

   Let's say you have two CEs (CE1 and CE2) connecting to Verizon, and you have a LAN behind CE1 and CE2. How can you configure CE1 as the main router and CE2 as the secondary router and connect to the ISP? Answer: Using a redundancy protocol - HSRP.

8. Why do we need route damping in BGP?

9. What protocols are running in your current organization? LAN environment and WAN environment?

10. When is it okay for an OSPF neighbor to be stuck in a 2-way state? When is it normal behavior for an OSPF neighbor to be stuck in a 2-way state, and we don't want it to change?

11. Let's say we have full mutual redistribution between BGP and OSPF. What would be the risk? What is bad about this redistributed topology? Answer: High CPU usage on IOS due to many redistributed routes, routing table size could increase, and most importantly, this kind of topology can cause loops in our network.

12. In OSPF, let's say we have two ASBRs redistributing the same routes from another autonomous system. Both ASBRs are advertising the same routes from the other AS. How can we manipulate the behavior of the ASBR so that all the traffic goes to a single ASBR? Answer:

13. Let's say if you want an ASBR to only advertise E1 routes in the rest of the OSPF topology, but E2 should be blocked, how would we do that?

14. How will you take an OSPF router out of service? Let's say there's a router in OSPF which we don't want to use anymore and want to remove from the OSPF network. How would we do that smoothly without interrupting traffic? Answer: By using an option in OSPF known as overloop in OSPF - it will advertise an LSA with infinite loops.

15. How does OSPF handle loops in a multi-area network? Answer: Using Area 0, as all routes are traveling through Area 0.