# Edge Computing **Entity class:** Distributed-computing architecture ## Definition **Edge computing** places computation, storage, inference, or decision support near the sensors and users generating data. In contested or disconnected operations, local processing reduces dependence on distant cloud services, lowers latency, limits backhaul demand, and can preserve function during communications disruption. ## Relationships - **field venues:** [[wiki/Experimental Demonstration Gateway Event|Experimental Demonstration Gateway Event]], [[wiki/USSOCOM Technical Experimentation Events|USSOCOM Technical Experimentation Events]], and [[wiki/SOF Week|SOF Week]]. - **constraints:** [[wiki/SWaP-C|SWaP-C]], hardware trust, bandwidth, power, and model size. - **security:** [[wiki/Hardware Security|Hardware Security]]. - **collection:** [[collections/Terrorism, Counterterrorism, and the Intelligence Environment|Terrorism, Counterterrorism, and the Intelligence Environment]]. ## Sources / Provenance - [NISTIR 8320 — hardware-enabled security for cloud and edge computing](https://www.nist.gov/publications/hardware-enabled-security-enabling-layered-approach-platform-security-cloud-and-edge) ## Simple Reminders, Quotations, and Thoughts > "Reduced Operator Workload: Cognitive Decision Assessment and Aiding technologies to automate or reduce crew workloads. This should include data fusion to process, filter, and decipher available data to meet crews need." > **— USSOCOM**, *accessed September 23, 2026, USSOCOM Capability Areas of Interest* [[reminders/Information/Automation Should Reduce Crew Workload Through Data Fusion by USSOCOM|Automation Should Reduce Crew Workload Through Data Fusion by USSOCOM]] > "Secure Mesh, Self-forming Mobile Ad-hoc Networks: Secure, robust accredited devices that allow for the establishment of secure self-forming, mobile ad-hoc networks interoperable with Joint SOF and Joint Services for dismounted SOF mobility platforms including unmanned systems and sensors." > **— USSOCOM**, *accessed September 23, 2026, USSOCOM Capability Areas of Interest* [[reminders/Information/Secure Mesh Networks Should Self-Form Across SOF Systems by USSOCOM|Secure Mesh Networks Should Self-Form Across SOF Systems by USSOCOM]] > "Decrease in latency of ground systems to less than one millisecond from external activity to viewing by Operator. Decrease in latency of ground systems to less than one millisecond from Operator activity to external action." > **— USSOCOM**, *accessed September 23, 2026, USSOCOM Capability Areas of Interest* [[reminders/Information/Ground Systems Should React in Less Than One Millisecond by USSOCOM|Ground Systems Should React in Less Than One Millisecond by USSOCOM]] > "Advanced Data Management: SOF requires technologies that provide automatic ingestion, metadata tagging, indexing, storage, synchronization, fusion, deduplication, mining, and dissemination of data collected by widely dispersed SOF resources. Data repositories require the ability to run AI/ML algorithms against the data sets to reduce analyst workload and rapidly deliver answers to warfighter problem sets." > **— USSOCOM**, *accessed September 23, 2026, USSOCOM Capability Areas of Interest* [[reminders/Information/Automated Data Management Should Rapidly Answer Warfighter Problems by USSOCOM|Automated Data Management Should Rapidly Answer Warfighter Problems by USSOCOM]] > "And to add to that, the latency of our models – our time lags in running inference specifically – has improved 80-percent in the year since NGA Maven became a program of record." > **— Frank Whitworth**, *May 19, 2025, 2025 GEOINT Symposium Keynote* [[reminders/Information/Model Inference Latency Improved Eighty Percent by Frank Whitworth|Model Inference Latency Improved Eighty Percent by Frank Whitworth]] > "Accelerating innovation in artificial intelligence hardware to make decisions at the edge faster" > **— DARPA**, *accessed September 23, 2026, Electronics Resurgence Initiative — What Is ERI?* [[reminders/Information/AI Hardware Should Make Decisions at the Edge Faster by DARPA|AI Hardware Should Make Decisions at the Edge Faster by DARPA]] > "NGA is also making strides with generative AI, developing geospatial agents to enable Large Language Models (LLMs) to seamlessly access geospatial data and aid in the creation of GEOINT — freeing up analysts for GEOINT exploitation requiring deeper analytical insight." > **— Frank Whitworth**, *May 14, 2025, 2025 HASC Testimony on National Security Space Programs* [[reminders/Information/Geospatial Agents Can Give Language Models Direct Access to GEOINT by Frank Whitworth|Geospatial Agents Can Give Language Models Direct Access to GEOINT by Frank Whitworth]]