74 lines
5.4 KiB
JSON
74 lines
5.4 KiB
JSON
{
|
|
"title": "Target Detection Identifiers",
|
|
"type": "REPORTED",
|
|
"description": "A GCHQ presentation from March 2009 introducing the concept of Target Detection Identifiers — the \"fundamental atom of the Internet age\" — describing how cookies and other unique persistent identifiers extracted from bulk internet traffic at 2,500 per second are used to build the surveillance databases, with a live AutoTDI discovery table showing cookies extracted from Facebook, Yahoo, Google, Reuters, YouPorn, Bebo and other sites.",
|
|
"long_description": "This is a UK SECRET STRAP2 COMINT ORCON GCHQ Applied Research presentation dated March 2009. The document provides the foundational technical description of GCHQ's Target Detection Identifier framework — the conceptual and technical basis for the agency's entire internet identity tracking infrastructure.\n\nA TDI is defined as \"a definite indicator of presence, that is unique and persistent for a user/machine\" — described as the \"fundamental atom of the Internet age\" answering the questions Who, When, Where and What. The concept evolved from telephony signalling, where a phone number indicates a user is online, but applied to the much more complex world of internet traffic where identifiers must be extracted from raw packets rather than read from signalling fields. At time of writing 70 distinct TDI types had been discovered, extracted at 2,500 per second from a single 10Gbps link — equivalent to 200 million per day — with each TDI type costing approximately 250 analyst hours to develop and an automated discovery prototype under development.\n\nThe TDI sources listed include Google, Yahoo, Facebook, MSN Hotmail, YouTube, Reddit, WordPress, Amazon, Google Earth, CNN, BBC, Hi5, Maktoob, Bebo, Orkut, Paltalk, Cloob and vBulletin — covering the major English and Arabic language internet platforms of the period. The table of TDI types shows both user-level identifiers such as Yahoo-Y-Cookie and machine-level identifiers such as Yahoo-B-Cookie and MS-MUID-Cookie, reflecting the distinction between tracking an individual person and tracking a specific device.\nThe AutoTDI discovery report — a live sample from October 2008 — shows cookies being automatically extracted and ranked by frequency across multiple bearers, with Facebook c_user cookies, Yahoo B cookies, Google ID and TM cookies, Reuters tracking cookies, YouPorn session and analytics cookies, and Bebo username and email cookies all appearing in the top results. The YouPorn entries appear prominently — five separate YouPorn cookie types in the top 20 — which became the basis for the Intercept article's headline about GCHQ tracking porn site visits. The presence of Reuters and CNN cookies in the same table as YouPorn illustrates the indiscriminate nature of the bulk collection.\n\nThe TDI applications section names the full suite of bulk stores fed by TDI extraction: MUTANT BROTH, AUTO ASSOC, KARMA POLICE, INFINITE MONKEYS, SOCIAL ANIMAL, MEMORY HOLE, MARBLED GECKO and HRMAP. The MUTANT BROTH interface screenshot shows the database containing 18.41 billion rows from December 2007 to June 2008, searchable by identifier, IP address or password, with all queries logged for audit and requiring a MIRANDA number and JIC priority code as justification.\n\nThe MUTANT BROTH results screenshot shows the system returning a Hotmail address associated with a target whose Hi5 cookie was seen accessing the platform from Istanbul, Ankara, Chicago and Damascus across a two-day period in June 2008 — demonstrating the geolocation capability embedded in the TDI framework through IP geolocation enrichment.\n\nAdditional bulk event applications described include file transfer signature detection for proof-of-life videos, internet profile-based dead letter drop detection, Yahoo webcam image collection and airline reservation confirmation email interception.",
|
|
"tags": [
|
|
"Applied Research",
|
|
"TDI",
|
|
"AutoTDI",
|
|
"cookies",
|
|
"bulk collection",
|
|
"internet surveillance",
|
|
"Facebook",
|
|
"Yahoo",
|
|
"Google",
|
|
"YouPorn",
|
|
"Reuters",
|
|
"Bebo",
|
|
"Hotmail",
|
|
"Hi5",
|
|
"Maktoob",
|
|
"Orkut",
|
|
"PalTalk",
|
|
"Cloob",
|
|
"vBulletin",
|
|
"geolocation",
|
|
"IP addresses",
|
|
"pattern-of-life"
|
|
],
|
|
"document": {
|
|
"creator": "GCHQ",
|
|
"date_created": "2009-03",
|
|
"classification": [
|
|
"UK SECRET",
|
|
"STRAP 2",
|
|
"COMINT",
|
|
"ORCON"
|
|
],
|
|
"dissemination": [],
|
|
"banner": "UK SECRET STRAP2 COMINT ORCON",
|
|
"genre": "slides",
|
|
"programs": [
|
|
"MUTANTBROTH",
|
|
"AUTOASSOC",
|
|
"KARMAPOLICE",
|
|
"INFINITEMONKEYS",
|
|
"SOCIALANIMAL",
|
|
"MEMORYHOLE",
|
|
"MARBLEDGECKO",
|
|
"HRMAP",
|
|
"AUTOTDI"
|
|
],
|
|
"countries": [
|
|
"France"
|
|
],
|
|
"people": []
|
|
},
|
|
"publication": [
|
|
{
|
|
"publisher": "The Intercept",
|
|
"date_published": "2015-09-25",
|
|
"title": "Profiled: From Radio to Porn, British Spies Track Web Users' Online Identities",
|
|
"link": "https://theintercept.com/2015/09/25/gchq-radio-porn-spies-track-web-users-online-identities/",
|
|
"reporters": [
|
|
"Gallagher, Ryan"
|
|
],
|
|
"contributors": [],
|
|
"doc_source": "https://theintercept.com/document/2015/09/25/tdi-introduction/",
|
|
"doc_url": "https://s3.documentcloud.org/documents/2433314/tdi-introduction.pdf"
|
|
}
|
|
],
|
|
"pdf_sha256": "86718e5878a258de03726d7d783913b738afc33f3c2b99c1459f4ad8118d234c",
|
|
"pages": 18
|
|
}
|