{ "title": "Statistics Configuration Read Me", "type": "REPORTED", "description": "XKEYSCORE administrator readme describing the configuration of statistics collection across the XKEYSCORE distributed infrastructure, including how collection nodes report operational metrics back to a central statistics server via MAILORDER, and defining the hierarchical terminology of the XKEYSCORE deployment architecture.", "long_description": "This is a technical administrator readme for XKEYSCORE's statistics configuration system, classified SECRET//REL TO USA, AUS, CAN, GBR, NZL. While primarily a routine systems administration document, it provides useful insight into the operational architecture and monitoring infrastructure of XKEYSCORE's global deployment.\n\nThe document describes how each XKEYSCORE node generates statistics files that are forwarded to a central statistics server called stats_central using MAILORDER — the same session forwarding mechanism used to distribute intercepted intelligence — or other site-specific file transfer methods. This reveals that NSA was collecting operational metrics from its global network of collection sites in near real time, enabling central visibility into processing volumes, system health and collection performance across all field deployments. Statistics are identified by SIGAD (Signals Intelligence Activity Designator) with optional cluster-specific suffixes, allowing granular tracking of individual collection systems within larger multi-cluster sites. The default trigraph used for routing statistics MAILORDER files is XKJ.\n\nThe glossary section at the end is particularly useful for understanding XKEYSCORE's deployment architecture. The document defines five levels of infrastructure hierarchy: Slave machines perform raw processing and receive configuration from their Master; Masters distribute configuration across their cluster and report to an Overlord in complex systems; Overlords control multiple clusters at a single site; Sites correspond to individual SIGADs — the unique designators assigned to each collection location; and Systems consist of one or more clusters with an optional Overlord. The distinction between front-end clusters performing raw packet collection and back-end clusters performing protocol processing reveals that large XKEYSCORE deployments separated the high-speed ingestion function from the more computationally intensive parsing and analysis work, a common architecture for handling high-volume network traffic.", "tags": [ "statistics configuration", "SIGAD", "cluster", "distributed infrastructure", "collection sites", "field sites", "operational metrics", "system monitoring", "administrator guide", "system administrator" ], "document": { "creator": "NSA", "date_created": "", "classification": [ "SECRET" ], "dissemination": [ "REL TO USA, AUS, CAN, GBR, NZL" ], "banner": "SECRET//REL TO USA, AUS, CAN, GBR, NZL", "genre": "manual", "programs": [ "XKEYSCORE", "MAILORDER" ], "countries": [], "people": [] }, "publication": [ { "publisher": "The Intercept", "date_published": "2015-07-01", "title": "XKEYSCORE: NSA’s Google for the World’s Private Communications", "link": "https://theintercept.com/2015/07/01/nsas-google-worlds-private-communications/", "reporters": [ "Marquis-Boire, Morgan", "Greenwald, Glenn", "Lee, Micah" ], "contributors": [], "doc_source": "https://theintercept.com/document/2015/07/01/stats-config-readme/", "doc_url": "https://s3.documentcloud.org/documents/2116167/stats-config-readme.pdf" } ], "pdf_sha256": "a9e803458488ca3990830a3eef89de4f1439bd12829eca107f73324cc8ee8113", "pages": 4 }