"My expectation is that we would not end up with SBInet all along the border. Already that doesn't look like the wise thing to do," said Mark Borkowski, executive director of the DHS Secure Border Initiative. Borkowski spoke June 17 before a joint hearing of two House Homeland Security subcommittees.

SBInet is a multi-billion dollar program to blanket U.S. borders with a chain of radars, cameras and heat and motion detectors, allowing border patrol agents working from a common operational picture to make targeted responses to incursions. The effort has cost $1.5 billion so far; the prime contractor is Boeing. DHS Secretary Janet Napolitano halted new work on the project in March, pending the outcome of a project review she ordered in January. So far, SBInet has been readied for deployment along a 53 mile stretch of Arizona border.

Borkowski's remarks came just weeks after Alan Bersin, Customs and Border Protection commissioner, told a Senate panel that SBInet is "not practicable" in the near term.

It would cost $8 billion to deploy SBInet's first planned technology rollout along the southwest U.S. border and require until 2016 or 2017 to complete, Borkowski said. But it's likely that DHS will instead to deploy alternate technologies to border regions, alternate technologies such as surveillance from a UAV, a blimp or from camera towers, he said. The project review ordered by Napolitano seeks to answer whether SBInet is viable and "if it is, [is it] the right way to spend money," Borkowski said.

SBInet ran into trouble "months after it started and it was spiraling downhill after that," said Randy Hite, Government Accountability Office director of information technology architecture and system issues, while testifying at the hearing.

The GAO released publically on June 17 a report criticizing SBInet program management. In particular, the report states that DHS has shrunk the capabilities of the first iteration of SBInet, called "block 1." For example, block 1's geographical footprint fell from about 655 miles to about 387 miles. The success rate for when SBInet technology correctly identifies an item of interest in the border region--distinguishing between, say, a human being and a cow--is now down to 49 percent, the GAO report states. Also acceptable downtime has gone up from 18.25 days per year to 54.75 days per year, excluding downtime for planned maintenance.

The acceptable range of daytime cameras also fell from 10 kilometers to 5, the GAO report states. Borkowski said during the hearing that the camera range has improved. The range fell "at the time when we had a very ambitious schedule," he said.

"We compromised on the performance of some of the hardware in order to get it on time, in order to meet the then-anticipated schedule. Now, in hindsight, that proved not to be wise, because we both failed to meet that schedule, and ended up with cameras that were less than optimal," Borkowski said.

For more:
- go to the House Homeland Security webpage on the hearing, complete with video and prepared statements
- read the GAO report, "DHS Needs to Reconsider Its Proposed Investment in Key Technology Program," GAO 10-340 (.pdf)