Survey: IT hurdles hinder big data adoption

While the Defense Department has recently attempted to harness big data analytics techniques in areas like gauging the security risks posed by insider threats, overall adoption of big data technology by U.S. agencies has been slowed by inadequate IT infrastructure.

Nearly three-quarters of federal managers overseeing big data projects think they lack adequate computing, storage and networking infrastructure, according to a survey released by Unisys Corp. Hence, the study warned, federal agencies may find themselves in the position of collecting large data volumes while lacking the ability to analyze key data.

Unisys, which according to Defense Systems’ sister web site Washington Technology, ranked 39^th in 2014 with federal contracts worth more than $529 million, reported this week that 93 percent of U.S. agencies responding to its survey said they had launched big data projects (The survey did not break out projects by agency.) Eighty-seven percent said analytics had the ability to predict trends and quantify risks.

While 73 percent are concerned about the lack of adequate IT infrastructure to support big data projects, a slightly lower percentage of respondents added that one consequence would be the ability to collect but not analyze key data.

Along with infrastructure concerns, only 16 percent of federal respondents reported they have fully implemented big data initiatives. About 40 percent of federal agencies polled by Unisys said they have no current plans to launch big data projects. Others said they are either conducting pilot projects or investigating their big data options.

Uncertainty over IT infrastructure to support big data stems in part from the slow federal rollout of cloud computing initiatives under FedRAMP, the Federal Risk and Authorization Management Program. The CIA has moved quickly to deploy cloud infrastructure while other organizations such as DOD have struggled to launch cloud and other infrastructure initiatives that would expand IT support for big data projects, such as using behavioral analytics to identify insider threats.

Unisys argues that data analysis and sharing insights across organizations remains among the biggest challenges for federal agencies. As with the private sector, the government also is having trouble recruiting and retaining qualified data scientists. Unisys reported that 68 percent of survey respondents said their agencies are hiring more data analysts; half said their agencies are looking for a data analytics director.

Unisys and other federal contractors see an opening here since, according to the Unisys survey, 98 percent of respondents said they would maintain or increase their reliance on contractors and consultants to work on big data projects over the coming year.

"While government agencies have domain expertise, they also need skilled data scientists, tools and infrastructure to unlock the value of their data assets," the company argued in releasing the results of its big data survey.

The company's "Big Data-as-a-Service" approach combines a cloud platform with "custom analytics." It positions the service as helping federal agencies cut through "analysis paralysis" in deciding which technology to adopt.