[fraserharris]

The leaked document reads: “Collection directly from the servers of these U.S. Service Providers: Microsoft, Yahoo, Google, Facebook, PalTalk, AOL, Skype, YouTube, Apple.” [1]

Its been reported that Facebook set up a secure server where the [NSA|FBI] could retrieve data that had been collected as legally required. [2]

A Google Chief Architect has claimed that no special systems have been built internally to support this. [3]

In summation, the likely setup is that someone (legal) vets the ordered request, oks the internal use of OAuth to spoof access to the data, & dumps it in a secure FTP server for the government to collect. To achieve "near real time", just look how they handle Google Product Feed's - every 30 minutes a "diff" of product data is updated.

[1] http://www.washingtonpost.com/investigations/us-intelligence...

[2] http://www.nytimes.com/2013/06/08/technology/tech-companies-...

[3] https://plus.google.com/u/0/+YonatanZunger/posts/huwQsphBron

edit: corrected Chief Architect title

[obstacle1]

Zunger is chief architect of Google+, not of Google as a whole. So he has deep knowledge of a single G product, not all G systems.