Intelligence software that the U.S. would rely on in a war with North Korea froze up repeatedly during a joint military exercise in South Korea in August, hampering the ability of U.S. and South Korean commanders to watch the movements of simulated enemy forces, a senior intelligence official said.
The Distributed Common Ground System-Army (DCGS-A) software is designed to link intelligence analysts to processed communications intercepts, imagery and radar collections stored in massive databases. When American intelligence analysts tried to use the software to track simulated North Korean troop movements, the screens on their DCGS-A workstations sometimes went black, forcing them to reboot the software, the senior intelligence official said.
Analysts could not always feed the latest enemy positions into the Command Post of the Future, the large computer displays that U.S. commanders would rely on to view troop positions and orchestrate defenses with their South Korean counterparts.
"What happened is the volume of information essentially crashed the software," the senior intelligence official said. "We learned to manually do [data retrieval] in chunks of information so DCGS would not crash."
The problem was discovered during the 10-day Ulchi Freedom Guardian exercise, a computer-generated North Korean attack in which tens of thousands of American and South Korean troops were mobilized in and around Seoul. The Pentagon billed the exercise as a "command post exercise" that would improve coordination of U.S. and South Korean forces.
"Initial analysis indicates that the use of legacy hardware was likely the primary cause of the system reliability issues," said a spokesman for the DCGS-A office in an email. "Personnel running current DCGS-A hardware during the same exercise in Yongin reported no major interruptions, issues, or outages. The issues identified during this exercise are currently being evaluated/corrected as needed."
U.S. intelligence officials have lately expressed concern that the wars in Iraq and Afghanistan have honed their ability to untangle insurgent networks and track people, but at the expense of practicing the more traditional military intelligence role of tracking forces during high-intensity conflicts involving artillery, tanks and fast-moving troop formations. This year's Freedom Guardian exercise offered a chance to show that DCGS-A, which is used by analysts and troops in Afghanistan, could perform well in a conventional war.
Software engineers will need to explore whether the greater volume of data stored in the conventional warfare database caused DCGS-A to lock up, the official said.
In a related problem, the DCGS-A system took 2 to 2½ minutes to nominate targets for bombing, a process that should take seconds.
Despite the problems, the senior intelligence official said the exercise should not be viewed as an indictment of the multibillion-dollar DCGS-A initiative.
"I'm going to make DCGS-A work," the official said.
All told, the DCGS-A system spent 10 out of 96 hours of planned operations locked up or being rebooted, the official said.
The Distributed Common Ground System-Army (DCGS-A) software is designed to link intelligence analysts to processed communications intercepts, imagery and radar collections stored in massive databases. When American intelligence analysts tried to use the software to track simulated North Korean troop movements, the screens on their DCGS-A workstations sometimes went black, forcing them to reboot the software, the senior intelligence official said.
Analysts could not always feed the latest enemy positions into the Command Post of the Future, the large computer displays that U.S. commanders would rely on to view troop positions and orchestrate defenses with their South Korean counterparts.
"What happened is the volume of information essentially crashed the software," the senior intelligence official said. "We learned to manually do [data retrieval] in chunks of information so DCGS would not crash."
The problem was discovered during the 10-day Ulchi Freedom Guardian exercise, a computer-generated North Korean attack in which tens of thousands of American and South Korean troops were mobilized in and around Seoul. The Pentagon billed the exercise as a "command post exercise" that would improve coordination of U.S. and South Korean forces.
"Initial analysis indicates that the use of legacy hardware was likely the primary cause of the system reliability issues," said a spokesman for the DCGS-A office in an email. "Personnel running current DCGS-A hardware during the same exercise in Yongin reported no major interruptions, issues, or outages. The issues identified during this exercise are currently being evaluated/corrected as needed."
U.S. intelligence officials have lately expressed concern that the wars in Iraq and Afghanistan have honed their ability to untangle insurgent networks and track people, but at the expense of practicing the more traditional military intelligence role of tracking forces during high-intensity conflicts involving artillery, tanks and fast-moving troop formations. This year's Freedom Guardian exercise offered a chance to show that DCGS-A, which is used by analysts and troops in Afghanistan, could perform well in a conventional war.
Software engineers will need to explore whether the greater volume of data stored in the conventional warfare database caused DCGS-A to lock up, the official said.
In a related problem, the DCGS-A system took 2 to 2½ minutes to nominate targets for bombing, a process that should take seconds.
Despite the problems, the senior intelligence official said the exercise should not be viewed as an indictment of the multibillion-dollar DCGS-A initiative.
"I'm going to make DCGS-A work," the official said.
All told, the DCGS-A system spent 10 out of 96 hours of planned operations locked up or being rebooted, the official said.