{ "title": "RT10 Overview", "type": "REPORTED", "description": "This RT10 overview document is a high-level briefing on the implementation of a real-time SIGINT architecture designed to modernize intelligence operations in the Iraq theater, specifically Baghdad. The project aims to replace legacy, \"manual-intensive\" processes with a \"SIGINT Brain\"—a distributed database and automated analytic framework capable of processing over 50 million GSM metadata events per day.", "long_description": "The RT10 initiative was born out of a critical need to handle the increasing \"velocity and volume\" of converged network data. Before RT10, analysts faced substantial lags, non-integrated tools, and a lack of comprehensive theater knowledge. The transition is described as moving from the static \"Atlantic Monthly\" style of analysis to the high-speed \"CNN scrolling bar\" model.\n\nKey technical pillars of RT10 include:\n\n* VoiceRT: A voice-processing engine designed to perform phonetic indexing on 1 million voice cuts per day and match them against 1,000 specific voice prints. This was intended to optimize linguist time by filtering for high-value targets automatically.\n* Advanced Sensors & Checkpoints: The implementation of \"active cell phone interrogation\" and RF illumination at military checkpoints (such as Route Irish) to fingerprint vehicles and identify threats like IED triggers or artillery shells.\n* Automation: The \"One-Click\" philosophy, allowing analysts to instantly generate relationship-chaining diagrams, frequent calling lists, and geospatial trends directly from alerts.\n* Connectivity: A robust local network utilizing Wi-Max (35 Mbps), Wireless Mesh, and Trojan Spirit to connect analytic nodes at the \"Perfume Palace,\" JIOC-I, and NSA-G.\n\nThe implementation followed a \"Spin\" methodology—90-day iterative cycles to expand capabilities. Spin 1 (August 2006) focused on initial installation and JUGGERNAUT data integration, while later spins incorporated FORNSAT (Foreign Satellite) data and next-generation analytical tools. The architecture specifically targeted various points in the GSM network, from tactical \"G-Box\" air-interface collection to national-level targeting at the Abis and A-Link interfaces using systems like OBELISK.", "tags": [ "GSM", "metadata collection", "phonetic indexing", "real-time alerting", "geospatial analysis", "geotemporal analysis", "tactical intelligence", "target development", "theater intelligence", "surveillance architecture" ], "document": { "creator": "NSA", "date_created": "2006-06", "classification": [ "TOP SECRET", "COMINT" ], "dissemination": [], "banner": "TOPSECRET//COMINT//20291123", "genre": "slides", "programs": [ "RT10", "OBELISK", "WISPYKNIT", "VICTORYUNIFORM", "JUGGERNAUT", "COBRAFOCUS", "MATTERHORN", "EXPIATION", "VoiceRT", "GREENDRAGON", "TAREX", "KEYCARD" ], "countries": [ "Iraq" ], "people": [] }, "publication": [ { "publisher": "The Intercept", "date_published": "2015-05-05", "title": "How the NSA Converts Spoken Words Into Searchable Text", "link": "https://theintercept.com/2015/05/05/nsa-speech-recognition-snowden-searchable-text/", "reporters": [ "Froomkin, Dan" ], "contributors": [ "Fishman, Andrew" ], "doc_source": "https://theintercept.com/document/2015/05/05/rt10-overview/", "doc_url": "https://s3.documentcloud.org/documents/2072024/rt10-overview.pdf" } ], "pdf_sha256": "d965e7b2be55a01c3d6fe6fda0531ab6f9467a7766adf905f717eb9d0aa8caaf", "pages": 15 }