Which statement correctly reflects latency for simple indexed queries and caching guidance?

Which statement correctly reflects latency for simple indexed queries and caching guidance?

• A. Simple indexed lookups take sub-millisecond to a few milliseconds, so caching is never beneficial.
• B. All queries should be cached to reduce latency.
• C. Cache expensive queries and frequently read aggregations; do not add infrastructure for simple primary key lookups.
• D. Latency is irrelevant to caching decisions.

Answer: (C)

Caching should be targeted where it adds real value. Simple indexed lookups, like primary-key lookups, are already very fast because they hit compact index pathways. Adding a cache for these quickly becomes a waste of memory and adds the complexity of keeping data fresh, without giving you meaningful gains in most cases. The biggest wins come from caching operations that are expensive to run and are requested repeatedly—such as large, resource-intensive queries, joins, or frequently read aggregations. Those results can be served from cache quickly, substantially reducing latency for users. So the recommended approach is to cache the expensive queries and the frequently read aggregations, while not building cache infrastructure around small, simple primary-key lookups. Latency considerations should drive where caching is applied, focusing on patterns with high cost and high repetition rather than applying caching indiscriminately. Other options miss the mark because caching every query wastes resources and adds complexity, and claiming latency doesn’t matter for caching decisions ignores the fundamental goal of caching: to reduce latency where it’s most impactful.

Caching should be targeted where it adds real value. Simple indexed lookups, like primary-key lookups, are already very fast because they hit compact index pathways. Adding a cache for these quickly becomes a waste of memory and adds the complexity of keeping data fresh, without giving you meaningful gains in most cases. The biggest wins come from caching operations that are expensive to run and are requested repeatedly—such as large, resource-intensive queries, joins, or frequently read aggregations. Those results can be served from cache quickly, substantially reducing latency for users.

So the recommended approach is to cache the expensive queries and the frequently read aggregations, while not building cache infrastructure around small, simple primary-key lookups. Latency considerations should drive where caching is applied, focusing on patterns with high cost and high repetition rather than applying caching indiscriminately.

Other options miss the mark because caching every query wastes resources and adds complexity, and claiming latency doesn’t matter for caching decisions ignores the fundamental goal of caching: to reduce latency where it’s most impactful.