whitebox

White box, or whitebox, is a term used to describe systems or methods that expose their internal structure, design, or data flows to observers. It is used in software testing, cryptography, artificial intelligence, and hardware, in contrast to opaque, black-box approaches.

In software testing, white-box testing (also called glass-box or clear-box testing) relies on knowledge of code, algorithms, and internal data structures. Test cases target control-flow and data-flow paths to improve coverage and reveal defects hidden from black-box tests. It requires access to source code and can be labor-intensive to maintain as software evolves.

In cryptography, white-box cryptography aims to protect keys and operations when the implementation runs in an environment where an attacker can observe or alter code and data. Techniques include obfuscation, masking, and transform-based schemes designed to resist key extraction. The field is active and debated; it is widely used for software protection and digital rights management, though it cannot guarantee absolute security.

In artificial intelligence, white-box models are interpretable, with transparent internal logic. They contrast with black-box models such as many deep neural networks. Examples include decision trees, linear models, and rule-based systems. White-box approaches support explainability, auditing, and regulatory compliance, though they may trade off some predictive performance on highly complex tasks.

In hardware procurement, white-box hardware refers to generic, unbranded components assembled into systems by cloud providers or integrators. White-box servers can lower costs and increase flexibility but may offer different support terms and warranties compared with branded equipment. Related terms include glass-box or clear-box, used as synonyms in some contexts.