Which performance metric is a measure of the time taken to complete an operation?

Latency, also known as response time, is the performance metric that directly measures the time it takes to complete a single I/O operation. It is the total delay from the initiation of a request by the host until the storage system completes the operation and sends a response. This metric is typically measured in milliseconds (ms) or microseconds (µs) and is a critical indicator of performance for applications that are sensitive to delays, such as transactional databases.

A. Bandwidth is a measure of the maximum theoretical data transfer rate of a network or connection, not the time for a single operation.
C. IOPS (Input/Output Operations Per Second) measures the quantity of read and write operations a storage device can perform in one second, not the time duration of a single operation.
D. Throughput measures the actual rate of data successfully transferred over a period, typically in megabytes per second (MB/s), reflecting data volume over time.

1. Dell EMC XtremIO X2 Storage Array: Introduction and Best Practices (White Paper), August 2019.

Section 3.1, Page 11: The document states, "XtremIO delivers consistent and predictable high performance with sub-millisecond latency." This directly associates latency with a time measurement (milliseconds) for operations, distinguishing it from IOPS and bandwidth, which are described as rates.

2. Dell PowerStore: Performance Primer (White Paper), April 2023.

Section 2.1, Page 6: Defines latency as "the time it takes to complete a single I/O request" and contrasts it with IOPS ("the number of I/O requests per second") and throughput ("the amount of data transferred per second"). This document explicitly confirms latency as the measure of time for an operation.

3. Patterson, D. A., & Hennessy, J. L. (2017). Computer Organization and Design RISC-V Edition: The Hardware Software Interface. Morgan Kaufmann.

Chapter 1, Section 1.6, "Performance", Page 35: This foundational computer architecture textbook defines response time (latency) as "the total time required for the computer to complete a task, including disk accesses, memory accesses, I/O activities, operating system overhead, CPU execution time, and so on." It clearly distinguishes this from throughput.

4. MIT OpenCourseWare, 6.004 Computation Structures, Spring 2017.

Lecture 19 Slides, Slide 5 ("Latency vs. Throughput"): The course material defines Latency as the "Time from start to finish of a task" and Throughput as the "Number of tasks completed per unit time." This academic source corroborates the distinction between a time-per-task measurement (latency) and a rate-of-tasks measurement (throughput/IOPS).