Understanding QPS, TPS, Concurrency, Throughput, and Related Performance Metrics

1. QPS

QPS (Queries Per Second) measures how many queries a server can respond to each second, reflecting the maximum throughput capability of a specific query server.

2. TPS

TPS (Transactions Per Second) counts the number of complete transactions per second, where a transaction includes a client request, server processing, and response.

3. Difference between QPS and TPS

TPS counts each complete request‑response cycle, while QPS can count multiple server requests generated by a single page view.

Example: a page request may generate three server queries, resulting in one TPS but three QPS.

4. Concurrency

Concurrency (or concurrency level) indicates the number of requests a system can handle simultaneously, reflecting its load capacity.

5. Throughput

Throughput is the amount of work a system processes per unit of time; both TPS and QPS are common quantitative indicators of throughput.

System Throughput Factors

The system’s throughput is closely related to CPU consumption per request, external interfaces, I/O, etc.

Higher CPU cost or slower I/O reduces throughput, while lower cost increases it.

Key Parameters

QPS/TPS, concurrency, and response time are the primary parameters.

Formula: QPS (or TPS) = concurrency / average response time.

Relationship

When either concurrency or response time reaches its limit, overall throughput cannot increase further and may even decline due to overload.

6. PV

PV (Page View) counts each page load or refresh as a separate view.

7. UV

UV (Unique Visitor) counts distinct users visiting within a day, usually derived by deduplicating logs.

8. DAU

DAU (Daily Active Users) measures the number of unique users who interact with a service in a single day.

9. MAU

MAU (Monthly Active Users) measures the number of unique users over a month.

10. System Throughput Evaluation

When designing a system, consider CPU, I/O, external service latency, and estimate performance accordingly.

Typical evaluation involves correlating peak TPS with daily PV to estimate daily traffic capacity.

Methods include identifying the system’s maximum TPS and daily PV, then using stress testing or experience to predict overall capacity.

11. Basic Concepts and Formulas for Software Performance Testing

Performance testing focuses on response time, resource utilization, scalability, maximum concurrent users, potential bottlenecks, and 24/7 availability.

From a developer’s perspective, assess architecture, database design, code efficiency, memory usage, thread synchronization, and resource contention.