Some Numbers Most Nsa's Information Collection

(Updated: July 16, 2017)

Today it's precisely 1 twelvemonth agone the Snowden-leaks started. Among the many highly classified documents which were disclosed during the past times twelvemonth are diverse charts that supply us amongst actual numbers most the total of information the National Security Agency (NSA) is collecting.

Here nosotros volition accept a hold off at those numbers in addition to come across what nosotros tin acquire from them past times comparison diverse sources in addition to from breaking them downwardly into NSA-divisions, countries in addition to collection programs. As however only fragmented parts have got been published, this overview cannot supply completeness or total accuracy (estimates are shown every bit circular numbers).

Numbers related to:
- BOUNDLESSINFORMANT
- NSA volumes in addition to limits
- GCHQ metadata collection
- NSA collection past times country
- NSA collection past times division
- SSO Collection programs
- Shared past times 2nd political party partner agencies
- Shared past times tertiary political party partner agencies

BOUNDLESSINFORMANT

The most detailed numbers most NSA's information collection are from the BOUNDLESSINFORMANT tool, which is used past times NSA officials to catch the metadata volumes collected from specific countries or past times specific programs.

Influenza A virus subtype H5N1 worldwide overview is provided past times a oestrus map which was published past times The Guardian on June 11, 2013. It displays the figures over a 30-day stream ending inward March 2013:

NSA worldwide total:

Internet records (DNI):
Telephony records (DNR):

221.919.881.317

97.111.188.358
124.808.692.959

This total of 221 billion telephony in addition to cyberspace records a calendar month equals 2,6 trillion a twelvemonth in addition to 7,3 billion a day. However, the actual issue of what NSA collects worldwide mightiness travel higher - come across the update below.

The BOUNDLESSINFORMANT worldwide overview for March 2013
(click to enlarge)

NSA volumes in addition to limits

The BOUNDLESSINFORMANT tool seems to travel really accurate, but there's closed to other nautical chart that gives dissimilar numbers. It's from a 2012 presentation for the SIGINT Development conference of the Five Eyes community in addition to shows the volumes in addition to limits of NSA metadata collection. The nautical chart was published past times The Washington Post on Dec 4, 2013 in addition to 1 time again inward Greenwald's mass 'No Place To Hide' on May 13, 2014.

Chart showing the volumes in addition to limits of NSA metadata collection
betwixt Jan in addition to June 2012
Redactions past times Greenwald or the press, explanations added past times the author
(click to enlarge)

This nautical chart shows the numbers of:
- telephony metadata which are received past times FASCIA, which is NSA's top dog ingest processor for telephony metadata;
- cyberspace metadata that are transferred to MARINA, which is a huge NSA database that tin shop cyberspace metadata for upwards to a year;
- cyberspace metadata that had to travel deleted because at that spot was patently non plenty storage space.

Except for the deleted metadata, the charts shows ca. 10,4 billion cyberspace metadata (DNI) a day, which makes 312 billion a calendar month or 3,7 trillion a year. There are ca. 4,5 billion telephony metadata (DNR) a day, which makes 135 billion a calendar month or 1,6 trillion a year. If nosotros compare these numbers amongst those from BOUNDLESSINFORMANT, nosotros come across a large difference:

Internet metadata (DNI):
Telephony metadata (DNR):

Volumes in addition to Limits
(a month, 1st one-half 2012)

312.000.000.000
135.000.000.000

BOUNDLESSINFORMANT
(a month, 1st one-half 2013)

97.111.188.358
124.808.692.959

There's a divergence of eleven billion telephony metadata betwixt both charts, but an fifty-fifty bigger gap exists betwixt the cyberspace metadata: the Volumes in addition to Limits nautical chart shows 215 billion to a greater extent than than BOUNDLESSINFORMANT. This discrepancy wasn't noticed inward the press reportings, nor inward Greenwald's book, thus at the 2d there's no clear explanation for this.

Update:
Influenza A virus subtype H5N1 possible explanation for the discrepancies betwixt these numbers tin travel constitute inward a FAQ document for the BOUNDLESSINFORMANT tool, which says the numbers shown inward the "map view" are lower than inward the so-called "org view" of the tool because for the latter, also records are counted that doesn't incorporate the Earth identifiers which are needed to travel counted inward the "map view".
This would also explicate the far bigger divergence betwixt the numbers of cyberspace metadata, because for cyberspace communications it is oft much to a greater extent than hard to attribute them to a exceptional Earth than for telephone conversations (which ever incorporate Earth in addition to part codes). This agency the Volumes in addition to Limits slide provides the to a greater extent than realistic numbers.

Telephony metadata

After existence processed past times FASCIA, the telephony metadata acquire to Hemisphere project, closed to 4 billion telephone metadata records are collected every twenty-four lx minutes stream from whatever carrier that uses AT&T switches inward reply to grand jury subpoenas inward counter-narcotics investigations.

Update #2:
During a parliamentary hearing inward Germany, an official of BND explained that 1 jail cellphone telephone creates betwixt 100 in addition to 200 metadata in addition to job organisation records a day. For 4.5 billion jail cellphone telephone users worldwide that would equal at to the lowest degree 450 billion metadata each day.

Update #3:
Influenza A virus subtype H5N1 2017 tourism written report from the Netherlands provided numbers showing that inward Jan 2013, Dutch mobile telephone users generated 255 1 yard m metadata a twenty-four lx minutes stream or 7,65 billion a month. The written report also confirms that for Dutch users, mobile phones create most 100 "transactions" a day.

GCHQ metadata collection

Even to a greater extent than metadata seem to travel collected past times NSA's British partner agency GCHQ, which according to this slide from 2011 collects 50 billion metadata per day. This makes 1,5 trillion a calendar month in addition to an astonishing eighteen trillion (18.000.000.000.000) a year!

This (partial) slide was published inward Greenwald's mass No Place To Hide, but without whatever farther explanation, thus nosotros don't know whether GCHQ is able to genuinely shop everything or has to delete large amounts, similar NSA. From the slide itself it seems that the issue of 50 billion refers to cyberspace metadata alone, which would brand this issue fifty-fifty to a greater extent than remarkable.

According to a written report past times The Guardian, GCHQ also collects 600 1 yard m telephony metadata a day, which makes eighteen billion a calendar month - a pocket-size issue compared to the cyberspace metadata this agency receives:

Internet metadata per month:
Telephony metadata per month:

BOUNDLESS
INFORMANT

97 bln.
124 bln.

Volumes
in addition to Limits

312 bln.
135 bln.

GCHQ

1500 bln.
eighteen bln.

For indexing in addition to searching the content of cyberspace communications, GCHQ uses the TEMPORA system, which is capable of processing the traffic from 46 fiber-optic cables of 10 gigabits per second. This makes that 21 petabytes of information stream past times these systems every day.

NSA collection past times country

The top dog BOUNDLESSINFORMANT interface amongst the oestrus map also lists the names of the countries which supply the highest numbers of data. These tin travel sorted inward iii dissimilar ways: Aggregate, DNI (internet) in addition to DNR (telephony), each resulting inward a slightly dissimilar top-5. The next aggregated totals (so both DNI in addition to DNR) are known:

NSA worldwide total:

Pakistan:
Afghanistan:
Iran:
Jordan:
India:
Saudi Arabia:
Iraq:
Egypt:
...
United States:
...
Brazil:

221.919.881.317 (100%)

27.275.944.618 (12%)
24.293.973.693 (11%)
15.834.475.801   (7%)
14.374.155.469   (6%)
12.616.915.557   (5%)
11.367.867.117   (5%)
10.487.011.026   (4%)
9.064.623.040   (4%)
...
3.095.553.478
...
2.300.000.000

These numbers signal from which countries NSA gathers most data, but the exact pregnant of the numbers has however non been clarified. We exercise know that BOUNDLESSINFORMANT counts metadata records, but what these records precisely are (for example: how many records are created past times 1 telephone call?), in addition to how they are attributed to a specific Earth is non clear.

Communications past times Definition have got 2 ends: the originating in addition to the receiving end. When both ends are inward the same country, it's slowly to attribute it to that exceptional country. But when the originating in addition to the receiving ends are inward a dissimilar country, how is such a communication registered? Maybe for both countries, although that would brand many of them appear inward these numbers twice.

United States

Edward Snowden saw the oestrus map amongst the 3 billion attributed to the U.S. every bit a proof that NSA was conducting domestic surveillance, although the oestrus map itself cannot supply sufficient bear witness for that. The 3 billion could really good relate to unusual communications which are simply transiting the US or to the American halt of for instance telephone calls where the other halt is a unusual suspect. Somewhat to a greater extent than information could have got been provided past times the bar charts for the US, but these haven't been published.

The issue of 3.095.553.478 for the U.S. is the aggregated total. The issue of cyberspace records (DNI) for the US is 2.892.343.446, which leaves simply 203.210.032 telephony records (DNR) or 0,065% of the aggregated total. In a tabular array this looks similar this:

U.S. total:

Internet records (DNI):
Telephony records (DNR):

3.095.553.478 per month

2.892.343.446 per month
203.190.032 per month

This tiny portion for telephone metadata is rather unusual given the fact that NSA is collecting all American telephone records, but does non thus amongst cyberspace metadata. This seems to signal that these domestic telephone records are non counted past times BOUNDLESSINFORMANT in addition to that the cyberspace records are from communications amongst at to the lowest degree 1 halt foreign.

NSA collection past times division

With a BOUNDLESSINFORMANT nautical chart most the NSA's Special Source Operations (SSO) segmentation published inward Greenwald's book, nosotros tin also compare the issue of information collected past times this segmentation amongst the total issue of NSA information collection. We come across that SSO, which is responsible for tapping the world's top dog fiber optic cables, accounts for 72% of all data:

NSA worldwide total:

Special Source Operations (SSO):
Other NSA divisions:

221.919.881.317 (100%)

160.168.000.000 (72%)
61.751.000.000 (28%)

This leaves the remaining 28% of the information to travel collected past times NSA's other top dog divisions: Global Access Operations (GAO), which operates mobile collection platforms similar satellites, planes, drones in addition to ships, in addition to Tailored Access Operations (TAO), which collects information past times hacking into unusual estimator networks. The remaining 28% could also embrace information collected past times the articulation NSA/CIA Special Collection Service (SCS) units in addition to past times tertiary Party partner agencies.

BOUNDLESSINFORMANT nautical chart most the SSO division
(click to enlarge)

SSO Collection programs

From the BOUNDLESSINFORMANT nautical chart most Special Source Operations nosotros tin come across how the total issue of information collected past times this segmentation breaks downwardly into the five biggest collection programs. From other charts nosotros also know the numbers collected past times closed to other programs, in addition to these are added hither too:

SSO worldwide total:

DANCINGSOASIS (US-3171):
SPINNERET (US-3180, purpose of RAMPART-A):
MOONLIGHTPATH (US-3145, purpose of RAMPART-A):
INCENSER (DS-300, purpose of WINDSTOP):
AZUREPHOENIX (US-3127, purpose of RAMPART-A):
...
FAIRVIEW (US-990):
...
PRISM programme are excluded. However, closed to other source (pdf) says that nether PRISM, to a greater extent than than 227 1 yard m "internet communications" are collected annually, which is ca. nineteen 1 yard m a month, but it is non known whether these "internet communications" are the same form of records every bit presented past times BOUNDLESSINFORMANT.

Processing in addition to storing

Metadata from a issue of large in addition to of import SSO collection programs are processed past times a organisation codenamed SHELLTRUMPET. As tin travel read inward the document below, this organisation processed almost 500 billion metadata records inward 2012, which gives an average of 41,6 billion a month, but past times the halt of 2012 SHELLTRUMPET was already processing 2 billion telephone phone exceptional records a day, which would brand lx billion a month:

MUSCULAR contributes lx gigabyte of information to the PINWALE database for cyberspace content every day, which is 1,8 terabyte a month. As BOUNDLESSINFORMANT counts 181 1 yard m records for MUSCULAR, this would hateful that 1 1 yard m cyberspace metadata records stand upwards for almost 10 gigabyte of (content) data.

This correlation tin travel used to brand a really petroleum gauge of the total total of cyberspace information collected past times NSA. The worldwide total of 97 billion cyberspace records a calendar month would in addition to then equal closed to 961 terabyte of information each calendar month or 11,5 petabyte a twelvemonth (some numbers to compare are here; the novel NSA information oculus inward Bluffdale, Utah tin shop an estimated 12 exabytes, which is 12.000 petabytes).

Shared past times 2nd political party partner agencies

The really closed working human relationship betwixt NSA in addition to the 2party partner agencies from the Five Eyes community leads to a regular telephone commutation of data, of which the most productive facilities tin travel seen inward a BOUNDLESSINFORMANT chart that was published past times Der Spiegel:

DS-300 (INCENSER):
...
DS-800:
DS-204A:
UKC-302A:
UKC-215:
...
DS-200B (MUSCULAR):

14.100.359.119
...
4.412.803.504
1.691.419.171
1.245.109.650
937.317.036
...
181.280.466

The SIGAD codes starting amongst DS announce closed to form of articulation collection program, those starting amongst UKC stand upwards for civilian operated facilities of the British signals intelligence agency GCHQ.

Shared past times tertiary political party partner agencies

NSA also gets information provided past times tertiary Party partner agencies. These are counted past times the BOUNDLESSINFORMANT tool too, every bit nosotros know from charts most a issue of European countries:

Federal Republic of Federal Republic of Germany (US-987LA):
? (US-985HA)
Federal Republic of Federal Republic of Germany (US-987LB):
Poland (US-916A):
French Republic (US-985D):
Espana (US-987S):
Italy (US-987A3005):
Kingdom of Norway (US-987F):
Kingdom of Denmark (?):
Kingdom of the Netherlands (US-985Y):

471.258.864
181.115.922
81.786.967
71.819.443
70.271.990
60.506.610
45.893.570
33.186.042
23.000.000
1.831.506

The total issue of information received from these nine countries is slightly to a greater extent than than 1 billion a month, which is simply a tiny 0,0045% of NSA's overall collection every bit counted past times the BOUNDLESSINFORMANT tool.

Initially, Glenn Greenwald reported inward diverse European newspapers that these numbers represented the telephone calls of European citizens intercepted past times NSA. But gradually it came out that his interpretation was wrong.

The charts genuinely demonstrate numbers of metadata that were collected from unusual communications past times European armed forces intelligence agencies inward back upwards of armed forces operations abroad. These information were after shared amongst partner agencies, most probable through the SIGDASYS organisation of the SIGINT Seniors Europe (SSEUR) group, which is led past times NSA.

> See also: NSA's unusual partnerships