PRE-Y2K PROBLEMS UNDERCUT OPERATION OF U.S. SATELLITE IMAGERY NETWORK

January 13, 2000

By

When a Year 2000 computer glitch impaired a critical satellite-based intelligence system on New Year's Eve, U.S. intelligence and defense officials had already been grappling for several days with a variety of technical problems not related to the date change that further undercut their ability to process and distribute the high-quality imagery products the system is designed to provide, Inside the Pentagon has learned.

On Dec. 27, difficulties with the primary system for disseminating imagery to national security "customers" resulted in lengthy "outages" that affected the U.S. Central and Pacific commands, which are responsible for operations in Southwest Asia and on the Korean peninsula, respectively. Three days later, while still working to assess the damage caused by the delivery troubles, officials scrambled to restore access to a Lacrosse imaging radar (IR) satellite, according to defense sources and internal DOD e-mail traffic.

Due to the highly classified nature of U.S. reconnaissance satellite operations, Deputy Defense Secretary John Hamre made no mention of these earlier problems when he told reporters about the Y2K anomaly at two separate briefings, the first on Jan. 1, the second on Jan. 4. At the Jan. 4 briefing, Hamre said the Y2K problems had been fully repaired and the intelligence system was operating at full capacity. In fact, Hamre insisted the episode could even be considered a "huge success story" because manual backup procedures ensured that "high-priority" intelligence requirements continued to be met despite the millennium-induced failures.

However, what Hamre also didn't say was that the National Reconnaissance Office, which builds and operates U.S. reconnaissance satellites, had never before attempted "sustained manual operations" to handle the feed from more than one satellite "for any prolonged period of time," according to sources. Because the Y2K glitch affected the ground stations that process the data from the three Advanced Keyhole electro-optical satellites and two Lacrosse IR birds that make up the imagery collection constellation, U.S. officials were unsure how long their ability to monitor events in global hot spots like North Korea, Iraq and the Balkans might be diminished.

The timing of the problems was especially critical. The defense and intelligence communities, concerned terrorist groups or rogue states might strike U.S. targets during the century roll-over period, were on heightened alert in late December. A Dec. 21 "Worldwide Caution" published by the State Department noted that the "U.S. government believes that terrorists may be planning to conduct attacks including against official and non-official Americans, in and around the New Year period." On top of that, DOD officials had long been concerned the military's Y2K repairs could expose its warfighting systems to electronic attacks.

Aside from the operational issues, the year-end complications with the imagery network generated a good deal of finger pointing. Lt. Gen. James King, the director of National Imagery and Mapping Agency, was reportedly furious over the imagery dissemination problems, which sources say stemmed from an unstable asynchronous transfer mode (ATM) switch at Ft. Belvoir, VA. The switch is operated by the Defense Information Systems Agency and has created problems in the past. King is demanding that DISA's leadership explain why the switch has not been repaired properly, these sources say.

The imagery distribution troubles come as Congress is voicing concern over the state of the intelligence community's "tasking, processing, exploitation, and dissemination" activities. Specifically, lawmakers believe too much money is being spent on sophisticated intelligence collection systems and too little funding is allotted for the TPED activities that are necessary to ensure the collected data is useful. Processing and exploitation, the synthesizing of raw data into a form that can be assessed by an intelligence analyst, are especially important processes. In November, the House of Representatives voted to restrict the acquisition of a next generation of imagery satellites until the administration develops a fully funded TPED plan (Inside the Air Force, Nov. 12, 1999, p1).

The congressional TPED concerns are not lost on the administration. The Pentagon's office of the National Security Space Architect recently completed a report that listed a series of current and prospective problems in the TPED arena (ITAF, Nov. 26, 1999, p1). For example, the report said NIMA, which was formed in 1996, "inherited a disparate suite of hardware and software from its predecessor organizations." Further, the report noted that "current estimates of the communications bandwidth within and between storage facilities indicate that the network will be insufficient to support the predicted imagery transfers in a timely manner."

Imagery experts agree the New Year's Eve Y2K troubles alone were likely more embarrassing than catastrophic. However, when the millennium-induced problems are coupled with the technical challenges that developed in the days prior to Jan. 1, the result suggests the nation's imagery processing and distribution system is in need of improvements sooner rather than later. And, although senior defense officials ultimately sought to cast the event as a positive story, information obtained by Inside the Pentagon reveals NRO was sailing in uncharted waters. Indeed, an hour before Hamre briefed the press on Jan. 1, intelligence officials were still unable to say when the imagery system would be fully restored to normal operations.

Imagery expert Dino Brugioni told Inside the Pentagon that all U.S. intelligence systems have backup procedures that are to be used in the event of a mechanical or computer-related failure. However, Brugioni, who spent more than 40 years with the CIA's National Photographic Interpretation Center, said he does not remember ever having to use a backup system to process imagery information.

According to U.S. intelligence officials, the Year 2000-related problems with the imagery network were caused by the failure of a Y2K "patch," a term that refers to repairs made to fill holes in the functionality of software. Patches are generally temporary solutions and smooth the transition between software applications that are supposed to be integrated.

The failure of the Y2K patch was first reported Jan. 12 by United Press International.

Last week, Hamre said the Y2K problem occurred because defense officials were unable to test the repairs as much as they would have liked. "Because we are operational all the time in this world, you don't have a backup capability that you can just test independently," he explained Jan. 4. "And so you have to test it in segments. And we tested things in segments, and they proved out to work in segments. When we put it all together, it turned out that we had an anomaly."

That anomaly directly affected the ground-based units that transform satellite data into usable images and produced a two-hour "blind" period. The imagery satellites themselves were unaffected by the Y2K problems. As emergency measures were put in place to perform by hand functions that take high-speed processors seconds to do, intelligence officials scrambled to assess the damage. Early on Jan. 1, NRO's Defense Communications Electronics Evaluation Testing Activity (DCEETA) at Ft. Belvoir was capable of capturing, on average, no more than 70 percent of the planned coverage by the imagery satellites.

DCEETA, known within intelligence circles as Area 58, is a highly secure facility, and defense officials do not discuss operations at the complex. According to the Washington, DC-based Federation of American Scientists, however, Area 58 is responsible for the tasking and primary processing of national imagery acquired through overhead systems, like the Keyhole and Lacrosse satellites. It's at Area 58 where NRO and its sister organization, NIMA, manage the daily operations of the imagery network.

While NRO operates the satellites, NIMA is responsible for determining what targets need to be covered. NIMA also "exploits" the images, a reference to the process through which analysts determine the intelligence value of a particular photo. NIMA also ensures delivery of the images to military commanders as well as senior civilian officials. NIMA was formed more than three years ago in response to the expanding requirements for imagery intelligence. A variety of federal entities were merged to create NIMA, including Brugioni's former employer, the NPIC.

While New Year's weekend was a long one for senior officials at NRO and NIMA, the final week of December was no more restful. On Monday, Dec. 27, a problem that defense officials have referred to as "chronic" surfaced again, crippling the operation of NIMA's Defense Dissemination System, the agency's primary platform for shipping "near original quality digital imagery to customers in near real time," according to information posted on the Federation of American Scientists' web site by John Pike, director of the group's Space Policy Project.

NIMA spokeswoman Jennifer Lafley declined to comment on DDS, noting that the system is classified.

According to sources, DDS routes imagery through DISA's ATM switch at Ft. Belvoir. From the ATM switch, the intelligence products are loaded into a "pipe," like the Defense Satellite Communication System," through which they are delivered to the various warfighting commands. However, the ATM switch has a record of poor performance, and the Dec. 27 problems produced outages of 12 hours or more for Pacific Command and Central Command, according to an internal Pentagon e-mail message.

At the same time, NIMA was also having trouble with its National Exploitation System. NES, also a classified system, supports a "collaborative community exploitation environment, providing products to consumers in a timely manner," according to Pike. Current plans call for NES to migrate in 2001 to the Exploitation Support System, which will be the intelligence community's common imagery exploitation unit.

The NES and DDS problems were not Y2K-related. While declining to comment on either system, Lafley did note that only two NIMA systems were not Year 2000 compliant, and neither system has any "impact or control on the employment of weapons or weapons systems or weapons platforms."

On Dec. 30, sources say access to a Lacrosse IR satellite was cut off. While the loss was temporary -- and also unrelated to Y2K -- the problem added to the list of concerns for NIMA and NRO.

NRO spokesman Art Haubold declined to comment on any of the week-long imagery difficulties, noting only that the Y2K problems "inhibited our ability to process the data that came from the satellites." Haubold added that "this was an NRO problem to fix."

Ironically, Hamre would never have mentioned the imagery network in public had it not been for the Y2K difficulties. Indeed, Hamre, citing the classified nature of the system and the sensitive missions it performs, provided few details. During two press briefings, he declined to give the name of the system, what type of intelligence it collects, or the exact nature of the problems experienced by the ground processing stations.

"It's extraordinary that they acknowledged [the Y2K problems] at all," said Pike of the FAS. -- Richard Lardner