Otsusteks

Otsusteks are modular, self-contained data units used in distributed artificial intelligence and edge computing. They bundle data, metadata, and small executable routines into a single transferable artifact that can be processed locally by edge devices or agents, reducing reliance on central servers and enabling offline operation. The concept is used in academic and industry experiments to improve latency, bandwidth efficiency, and interoperability.

A typical otsustek contains three parts: a header with a unique identifier, version, timestamp, and a digital signature; a payload carrying structured information such as sensor data, model parameters, or documents; and a sandboxed script or rules section that defines lightweight processing or decision logic. Sizes are intentionally constrained to enable fast transmission and safe execution.

Security and handling: otsusteks are usually transmitted over mesh or publish-subscribe networks and are validated on receipt through signature checks, version compatibility, and integrity verification. They are designed to be tamper-evident and to operate within sandboxed environments to prevent cross-contamination of systems. Caching and deduplication strategies are common to improve efficiency.

Applications and status: They target Internet of Things, autonomous systems, and research environments where low latency and offline capability are valued. While not standardized, several open formats and reference implementations exist, focusing on interoperability and secure execution. The concept remains evolving as researchers explore governance, versioning, and cross-platform tooling.