ITPEC FE Morning April 2022 Question 15

A same-spec standby server keeps throughput after a failure — in hot standby one server works and an identical one waits, ready to take over the whole load.

In a hot standby HA cluster the secondary server runs and stays synchronised but does not process work. When the primary fails, the secondary takes over all its processing. Because it must carry the full load alone, it needs the same specifications, and then throughput after the failover is the same as before. The cost is that the standby machine does no useful work in normal operation.

In a load distribution cluster every server works, and a load balancer spreads requests over them.

Answer (a)

Why not others:

• •(b) — redistributing work to the remaining active servers, which then slow down, is how a load distribution cluster behaves after a failure
• •(c) — even use of every machine and scaling out by adding servers are advantages of load distribution, not of an idle standby
• •(d) — keeping data consistent across several servers that all process requests, through a shared database, is a need of load distribution; in hot standby only one server processes at a time

Key rule: Hot standby trades an idle, identical spare for unchanged throughput on failover; load distribution uses every server but loses capacity when one fails.

System Component · AI-written, may contain errors