dremel(where facebook's presto comes from)
bigtable(where amazon's dyanamoDB comes from)
Most tech companies have google to thank for large scale data processing. Google could have not published anything, but they chose to publish these papers. There are other papers published that other companies haven't even started copying yet, as google is 5 years ahead of the industry in terms of large scale data processing. When these companies get there, they can check out the published papers, saving years of time.
Here's a random list:
Page Rank
Map/Reduce
Statistical Language Translation
Unsupervised image feature extraction[1]
Go
Dart
Large scale software defined networking.
The list could go on..
[1] http://arxiv.org/abs/1112.6209