Remote Application Management
AirStack Edge™ is API software for deploying and managing RF AI application suites at the edge, which includes integration of multiple applications and models to execute larger RF intelligence tasks. The product also integrates with operations and orchestration platforms.
AirStack Edge powers intelligent workflows for RF AI. Edge enables remote control of the RF AI application suite on Deepwave’s edge platform, allowing for updates to RF intelligence development on the edge. Additionally, Edge integrates with operations platforms to enable the secure sharing of low-latency, low-bandwidth RF AI results, driving automated workflows with RF intelligence. Highlights include:
-
- An API-first design seamlessly integrates with enterprise systems — including operations, orchestration, and MLOps platforms — to enable automated workflows.
- Adapt to changing RF environments by remotely updating individual AI models, RF applications, or entire RF AI edge deployments.
- Extend enterprise AI model versioning and performance to the edge for high-performance AI insights.
- Built with enterprise-grade security, the Edge API product protects every connection with encryption and authentication —so enterprise data and workflows stay safe from end to end.
How it Works
AirStack Edge is based on NVIDIA Jetpack and includes support for the NVIDIA® Jetson Orin™ platform. Following is an overview of Edge in action.
First, Deepwave’s AIR-T edge platform receives and analyzes RF signals using RF AI applications running on the AIR-T’s integrated FPGA, CPU, and GPU hardware. AirStack Edge then provides FIPS-grade encryption, enabling the delivery of AI-generated data into automated workflows. Edge then provides orchestration capabilities to actively recollect new RF intelligence by running new models, recording signals, or broadcasting signals. This end-to-end process runs security in real time.
Noteworthy features include
- Enterprise Integration: REST API integrates with enterprise systems and platforms. Fully specified as an OpenAPI document.
- Integrated Inference Server: Provides direct connection of RF samples or demodulated data to the on-device inference server.
- Data Acquisition: Allows sampling of RF data to inform AI model development and updates.
- Model Validation: Provides a feedback loop for validating deployed models. Includes APIs that can ingest external data or training data to test AI model performance.
- Model Deployment: Ability to upload new AI models into data pipeline to deliver better AI model performance.
- Custom Models: Provides tools for deploying customized models. For example, injecting custom code into the AIR-T AI application workflow to handle pre- or post-processing of data, in addition to other models and applications simultaneously running on the AIR-T
- Custom Outputs: Ability to stream multiple classification results to enable downstream analysis
- Security: Includes mandatory encryption and authentication of all API calls. Uses FIPS 140-3 validated cryptography module to support FedRAMP and other regulated deployments.
Implementations and Outcomes
AirStack Edge is the key enabling software to scale RF intelligence solutions.
Radio Intelligence Agent
AirStack Edge provides the secure, integrated API essential for the RIA solution’s autonomous operation. It orchestrates NVIDIA AI model deployment, collects real-time RF data, and delivers FIPS-grade encrypted, AI-generated intelligence directly to the DataRobot Agent Workforce Platform.
RF Signal Characterization
AirStack Edge supports RF Signal Characterization by providing model deployment and model validation APIs. This capability allows developers to remotely upload and test new AI models, ensuring the deployed characterization models maintain superior accuracy against constantly evolving electromagnetic signal environments.
Integration Details
AirStack Edge complements enterprise operations and orchestration platforms. The product provides a robust, API-first software layer that developers can easily integrate into their existing MLOps and enterprise IT systems. This integration simplifies the deployment of RF AI application suites by providing a fully specified OpenAPI document.
Furthermore, Edge includes mandatory encryption and authentication of all API calls, utilizing a cryptography module designed to support FIPS 140-3 validation, ensuring end-to-end security and compliance for regulated deployments.
AirStack Core requires AirStack Core 2.2+.
Learn More
Frequently Asked Questions
What strategic advantage does AirStack Edge provide for scaling my RF Intelligence programs?
AirStack Edge delivers a critical competitive advantage by transforming RF AI from a development prototype into a scalable, centrally managed enterprise solution. It provides the secure, remote orchestration layer required to manage entire fleets of AIR-T devices operating at the edge. This capability ensures your organization can remotely deploy, update, and govern applications across all field assets, extending high-performance AI model versioning to the edge for consistent, reliable Electromagnetic Spectrum Monitoring.
How does the API-first design simplify integration with existing MLOps and enterprise systems?
AirStack Edge is built with an API-first design, fully specified as an OpenAPI document, which guarantees seamless integration with external MLOps platforms, enterprise systems, and orchestration tools. It provides APIs for model deployment, model validation, and data acquisition, allowing developers to manage the entire AI lifecycle at the edge. This approach abstracts away low-level hardware complexity, enabling engineers to focus on custom code and injecting post-processing logic into the AI application workflow.
What security features are included to ensure compliance for regulated deployments?
AirStack Edge is engineered for high-security environments. It includes mandatory encryption and authentication for all REST API calls, protecting every connection end-to-end. Furthermore, the underlying cryptography module is designed to support FIPS 140-3 validation, ensuring the product meets the stringent security requirements necessary for regulated deployments, supporting secure sharing of low-latency, low-bandwidth RF AI results within the security perimeter.
How does AirStack Edge enable automated workflows and adaptation to rapidly changing RF environments?
The software enables autonomous workflows by providing orchestration capabilities to actively recollect new RF intelligence. When the RF environment changes, Edge allows operators to remotely update individual AI models, RF applications, or entire edge deployments. It can actively run new models, record new signals via the Data Acquisition API, or broadcast signals, ensuring the deployed RF Intelligence application suite remains accurate and responsive to new signal threats.
Is AirStack Edge required to use AirStack Core, and what are the software dependencies?
AirStack Edge is a powerful, optional component that complements the foundational AirStack Core operating system, it is not strictly required for Core to run. AirStack Edge is required to manage RF AI application suites remotely across multiple deployed devices.