Cloud Storage Systems Questions and Answers

Answer: A

Correct answer: enabling cross region replication

Cross region replication stores a second copy and pays egress to put it there, so it increases cost - deliberately, in exchange for regional resilience. The other three all reduce it: lifecycle transitions move cold data to cheaper classes, deleting incomplete multipart uploads removes fragments that are billed and invisible, and compression reduces the bytes stored.

Answer: B

With few reads there are few retrievals to pay for, so the standing storage charge dominates. For a frequently read dataset the balance shifts to retrieval and requests. Few reads means few charges. The balance shifts entirely with access.

Answer: C

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

Both are true and the reason names the two charges that catch people out. Archive classes bill for retrieval, sometimes substantially, and impose a minimum storage duration so that an object deleted early is billed as though it had stayed. Data read or churned frequently can therefore cost more in an archive class than in a standard one, which is why lifecycle rules must be set from the actual access pattern.

Answer: D

The access pattern drives everything else, and monitoring closes the loop where the pattern turns out to differ from what was assumed. Choosing a class before understanding access is the usual error. Monitoring closes the loop at the end.

Answer: A

Repeated reads are served from the cache rather than from the origin, cutting both egress and latency. The origin remains authoritative and must still exist. Repeated reads never reach it. Egress and latency both fall.

Answer: B

Correct answer: 1 and 2 only

Statements 1 and 2 are correct: file storage gives a shared POSIX filesystem several instances can mount, and that capability costs several times object storage per gigabyte. Statement 3 is wrong - archives belong in object storage, and in its coldest classes at that. File storage is chosen when applications need filesystem semantics, not when data merely needs keeping.

Answer: C

They are local to the physical host and their contents are lost. They suit scratch and cache data where loss is acceptable, not anything that must survive. The volume is host local. Stopping the machine loses the contents.