What is it about?

Data alone is not information. Guerrilla techniques emphasizes the viewpoint that performance models are a vital complement to performance measurements, e.g., application monitoring, in order to fully understand what information those data are trying to convey, e.g., long-term scalability.

Featured Image

Why is it important?

The Guerrilla approach also provides validation that the measurements are correct, as well as defining a quantification framework to evaluate ROI, e.g., reducing cloud chargeback. Examples of applying this Guerrilla approach to modern computer systems, such as, cloud-based applications and GenAI pre-training LLMs, are presented.

Perspectives

Measurement + Models = Information

Neil Gunther
Performance Dynamics Research

Read the Original

This page is a summary of: Guerrilla Techniques for Robust Performance Engineering, May 2025, ACM (Association for Computing Machinery),
DOI: 10.1145/3680256.3721308.
You can read the full text:

Read

Resources

Contributors

The following have contributed to this page