What performance improvement is typical when using an External Cache like Redis?

What performance improvement is typical when using an External Cache like Redis?

• A. Increases latency due to network overhead.
• B. No effect on latency.
• C. Reduces latency only for writes.
• D. Reduces latency from roughly 50ms to under 2ms.

Answer: (D)

Caching data in an external in-memory store speeds up access by serving frequently requested items from memory instead of hitting the slower primary database or disk. Redis stores data in memory and is optimized for very low latency, so a cache hit can drop latency from tens of milliseconds to just a few milliseconds or less. That’s why you often see large improvements, such as moving from roughly 50 ms to under 2 ms, when the requested data is already in the cache and the cache is sized and positioned well for the workload. There is a small network overhead to reach the cache, but it’s usually outweighed by the faster access compared with querying the primary data store. Writes can be more nuanced depending on the caching strategy (write-through, write-behind, etc.), but the primary performance win is the substantial reduction in read latency for cached data.

Caching data in an external in-memory store speeds up access by serving frequently requested items from memory instead of hitting the slower primary database or disk. Redis stores data in memory and is optimized for very low latency, so a cache hit can drop latency from tens of milliseconds to just a few milliseconds or less. That’s why you often see large improvements, such as moving from roughly 50 ms to under 2 ms, when the requested data is already in the cache and the cache is sized and positioned well for the workload. There is a small network overhead to reach the cache, but it’s usually outweighed by the faster access compared with querying the primary data store. Writes can be more nuanced depending on the caching strategy (write-through, write-behind, etc.), but the primary performance win is the substantial reduction in read latency for cached data.