distributedsystem/books/systemperformance
Core Idea
Systems performance studies the whole stack (application, OS, kernel, hardware) with two goals: improve price/performance and reduce latency outliers.
- The same skill applies to a slow laptop and to large-scale environments like Facebook’s data centers or the Netflix cloud.
- Adjacent work: benchmarking, capacity planning, bottleneck elimination, and scalability analysis, done early enough that scalability limiters are found while they are still cheap to fix.
- Note map: 01. System Performance - Introduction (with Flame Graphs) and 02. System Performance - Methodologies.
Systems performance is an important skill for all computer users, whether you’re trying to understand why your laptop is slow or optimizing the performance of a large-scale compute environment (for example, Facebook’s data centers or the Netflix cloud). Systems performance is the study of application, operating system, kernel, and hardware performance.
There are two general goals:
- Improving price/performance
- Reducing latency outliers
Other activities of systems performance include benchmarking to evaluate systems, capacity planning, bottleneck elimination, and scalability analysis – so that you discover scalability limiters early, in time to fix them.