NSA to crack codes with big data

Part of the Pentagon's $250 million investment in finding significance within the gazillions of disparate data produced worldwide will go toward decoding foreign secrets, according to White House officials.

Under a governmentwide big data enterprise launched Thursday, the Defense Department will continue with a program that lets companies demonstrate to the National Security Agency new commercial tools that could aid military code makers and code breakers. The attempt to draw meaningful trends from a constantly increasing abyss of electronic information is the crux of the big data problem, technologists say, and adversaries are taking advantage of that digital maze to ferry secrets.

The NSA-Central Security Service Commercial Solutions Center program will be "hosting vendor capabilities presentations that showcase new commercial technology developments that meet the strategic needs of NSA-CSS and the national security community," noted the White House big data agenda. CSS employs the government's top cryptology specialists.

Other computer security programs under Defense's big data initiative include a quest to uncover cyberspies.

At the Defense Advanced Research Projects Agency, the funding will help support the Cyber-Insider Threat, or CINDER, a program that identifies network activity that is consistent with the hallmarks of cyberespionage.

CINDER will run simulations of adversary missions over Defense networks to expose hidden operations, according to White House officials. The goal is to develop ways of more quickly detecting advanced persistent threats that linger silently in U.S. networks to carry out strikes -- such as data extraction or manipulation.

"Such threats take many forms but include variants like trojaned code, backdoors in embedded systems, worms and logic bombs, all of which could prove detrimental to the warfighter if not discovered," the CINDER website states.

Another big data-fueled project at DARPA -- Insight -- fills in for the human brain to find relationships among morasses of intelligence from myriad sources. The automated thinking uses information from sensors and prediction algorithms, among other things, to support commander decision-making, according to the program's site.

In February, DARPA melded data from multiple intelligence sources to provide 135 terabytes of information -- about 270 years' worth of nonstop MP3s -- to users across government, industry and academia.

NSA also will wield big data to fight cyberassaults, but the agency intends to delegate some of the heavy lifting to volunteers. "'Vigilant Net: A Competition to Foster and Test Cyber Defense Situational Awareness at Scale' will explore the feasibility of conducting an online contest for developing data visualizations in the defense of massive computer networks," the White House's agenda states.