Cloud Networking Questions and Answers

Answer: A

Correct answer: Both A and R are true, and R is the correct explanation of A

Both are true and the reason gives the mechanism. A DNS change only takes effect once the cached record expires, and resolvers and client libraries hold it for the time to live - some ignoring low values and holding longer. A load balancer removes an unhealthy target from rotation within seconds and needs no client to re-resolve anything, which is why DNS failover is used across regions and load balancers within them.

Answer: B

Correct answer: 1 and 2 only

Statements 1 and 2 are the direct trade: a shorter time to live propagates changes sooner but multiplies the queries, since every resolver must ask again more often. Statement 3 contradicts the first - failover speed *is* governed by the time to live, which is precisely why records intended for failover are given short values well before they are needed.

Answer: C

Correct answer: leaving the provider network outward

Providers charge for egress - data leaving their network - and generally not for ingress. The asymmetry shapes architecture: moving computation to the data is cheap, moving data to the computation is not, and it is a large part of what makes migrating away from a provider expensive once a substantial dataset is held there.

Answer: D

Correct answer: traffic between two instances in one subnet

Traffic within a single subnet stays inside one availability zone and is normally free. The other three all cross a boundary that is metered: out to the internet, between availability zones, and between regions, with the charge rising in that order. Knowing where those boundaries lie is most of what cloud network cost control amounts to.

Answer: A

Correct answer: True

Two components that exchange a great deal of traffic benefit twice from sharing a zone: the round trip is shorter, since the traffic stays within one data centre campus, and no cross-zone transfer charge applies. It is a real optimisation for chatty pairs such as an application server and its cache - and it is also a reduction in resilience, which the next item addresses.

Answer: B

Correct answer: Both A and R are true, and R is the correct explanation of A

Both are true and the reason states both sides of the trade. Cross-zone traffic costs money and adds latency, so co-location saves on both; but a zone is a failure domain, and everything within one fails together. The usual resolution is to keep the chatty pair together within each zone and run the whole stack in several zones, so the traffic stays local while the service survives losing a zone.

Answer: C

The address range is hardest to change later, so it comes first and should leave room. Security rules come last because they depend on what has been placed where. Growth room is worth planning for. The range is the hardest thing to change.

Answer: D

Correct answer: the address range chosen

Security group rules, load balancer types and DNS records can all be changed in minutes. The VPC address range cannot: it is fixed at creation, everything is numbered out of it, and changing it means rebuilding the network and re-addressing what runs in it. Choosing a range too small, or one that overlaps a range the organisation may later need to peer with, is a decision that has to be lived with.