<<

. 67
( 118 .)



>>

weeks earlier, on New Year's Day 2000, he had successfully dodged the
Y2K bullet. But for nearly a decade, pressured by ever-increasing
demand and overuse, NSA's brain had been heading for a stroke. The
first signs appeared in the early 1990s, when it became obvious that the
agency's massive system for processing, storing, and distributing Sigint”
codenamed Universe”had become technologically outdated. Universe
required 130 people to administer, took up 20,000 square feet of floor
space, and ate up enormous amounts in operations and maintenance
costs.
In an effort to replace Universe before it crashed, a new system was
developed in 1993 that used standard workstations, servers, and
supercomputers. Codenamed Normalizer, the new system took up 15,000


379
fewer square feet of floor space, saved $300,000 a year in costs, and cut
the number of people needed to operate it to just ten. But as the system
became smaller, the demands placed on it grew exponentially. Delivery
time for processed Sigint, for example, was shrunk from more than an
hour to only ten minutes. Given such pressures, an electronic aneurysm
was inevitable”and the most vulnerable time was millennium eve. In
addition to performing its normal hefty workload, the agency's computer
system would also have to figure out that switching from 99 to 00 meant
moving ahead to 2000, not behind to 1900.
In the weeks leading up to the new century, Hayden ordered that
contingency plans be developed "to maintain continuity of operations of
our critical intelligence mission" in case of a massive crash. His
predecessor, Lieutenant General Kenneth A. Minihan, had called the Y2K
problem "the El Nino of the digital age." In August 1998 he said, "As each
day passes we are coming closer to the date of one of the largest
technological and managerial challenges ever faced by our workforce."
As early as October 1996, the agency had set up the Millennium
Program Management Office, later named the Year 2000 Oversight Office.
NSA also began demanding that, as a prerequisite of doing business with
the agency, vendors state in writing that their products contained no Y2K
problems. Charged with coming up with a solution was Ronald Kemper,
NSA's chief information officer. Desperate for staff who could help repair
the millennium bug, the agency implemented an "Emergency 911"
operation to quickly find and recruit people with critical knowledge of
some of the older and more obscure computer languages. Incentives were
offered. General Minihan promised money bonuses and time off.
By 1998, however, agency officials were discovering that many
companies that early on claimed to be Y2K compliant were suddenly
retracting their words. "In some cases," said one NSA report, "the Agency
may not know there is a problem until something breaks." Said Minihan:
"Solving the Y2K problem is a tedious job and we are fighting a battle
against a deadline that will not move under any circumstances."
As the date approached, the systems governing the agency's
thousands of computers were assessed and stickers were placed on the
terminals. A green "Y2K OK" sticker indicated that the system would
pass over the threshold without problem; a yellow sticker marked
systems concerning which there was still some question, and a red
sticker warned, "Y2K NOT OK."
Less than a year before the critical date, NSA was still behind
schedule. Only 19 percent of the agency's computers were ready, and
repairs on nearly 60 percent were late. But as a result of a crash
program, computer programmers managed to bring 94 percent of the
computers into compliance by July 1999. The remaining 6 percent were



380
expected to be ready by the end of September. Ultimately, as in most of
the world, the millennium's arrival caused little or no disruption to NSA's
powerful computers and software; the agency continued to eavesdrop as
though nothing had happened. Until January 24.
Finally, after the agency spent thousands of staff hours and more
than $3 million on repairs, the system was patched together. After three
days, NSA awoke from its electronic coma, its memory still intact. "We
had the ability to store that which we collected over this three-and-a-
half-day period," said Hayden. "When we were able to go back and
process the information when that capability came back, it took eight to
twelve hours to process and analyze the information that we had
collected." During the outage, much of intercept traffic that would have
normally gone to NSA shifted instead to GCHQ. "We covered the whole
thing for them," said one GCHQ official, "to their acute embarrassment."
A year after the crash, with computer management now more
centralized, NSA's brain was again functioning normally, at about 12 to
15 percent of capacity. Nevertheless, Hayden concluded, "The network
outage was a wake-up call to our stakeholders and us that we can no
longer afford to defer the funding of a new infrastructure. And the
challenge doesn't stop there."


With his pudgy face, rimless glasses, and hairless dome, Hayden more
closely resembled a John Le Carr© spymaster than an Ian Fleming secret
agent. He also lacked the background of the stereotypical super-high-
tech spy chief. Shortly after his arrival on the eighth floor of OPS 2B, he
told his staff that arithmetic had never been his best subject. "I'll state
right up front," he admitted, "I am not a mathematician or a computer
scientist and I won't pretend to be one. I will be relying heavily on all of
you who are." To make the point, he added, "When I think about the
intellectual and mathematical brainpower that comes to work here every
day, I can recall the same intimidating feeling I experienced as a child
when Mrs. Murphy introduced me to the times tables in the second
grade."
Born on March 17, 1945, Hayden grew up in Pittsburgh. In college
and graduate school at Duquesne University he avoided hard math and
science courses and instead studied history. During the height of the
antiwar era, the late 1960s, Hayden excelled in ROTC, becoming a
Distinguished Graduate of the program. He entered the Air Force in
1969, shortly after finishing his master's degree in history, and was
assigned as a briefer at Strategic Air Command headquarters at Offutt
Air Base in Omaha, Nebraska. Two years later he was assigned to Guam
as chief of current intelligence for the headquarters of the 8th Air Force.
He spent the last half of the 1970s at various schools, mostly teaching



381
ROTC at St. Michael's College in rural Vermont.
In June 1980 Hayden, newly promoted to major, was sent to Osan Air
Base in South Korea as chief of intelligence for a tactical fighter wing.
Two years later it was back to the good life again, as a student and then,
in Sofia, Bulgaria, as air attach©. From there Hayden moved into a policy
job in the Pentagon and then over to the Bush White House, on the
National Security Council, until 1991. After an intelligence assignment at
U.S. European Command Headquarters in Germany, he took over the Air
Intelligence Agency and became director of the Joint Command and
Control Warfare Center at Kelly Air Force Base, Texas. There he became
heavily involved in the concept of information warfare. Finally he was
made deputy chief of staff for the United Nations Command in South
Korea, where he dealt with the issue of missing servicemen from the
Korean War.
Hayden was in Korea when he secretly received word of his new
assignment to NSA. Shortly afterward, on a Friday night, he went to the
base movie theater with his wife. Playing was a film he had not heard of,
Enemy of the State, in which Will Smith plays an average citizen
eavesdropped on by NSA, and Gene Hackman plays a retired NSA official
worried about the agency's enormous power.
"Other than the affront to truthfulness," said Hayden, "it was an
entertaining movie. And I will tell you, I walked out of there saying, you
know, that's not a good thing, to portray an agency so inaccurately. But
I'm not too uncomfortable with a society that makes its bogeymen
secrecy and power. That's really what the movie's about”it was about
the evils of secrecy and power. Then they tacked NSA on, that was the
offensive part. But making secrecy and power the bogeymen of political
culture, that's not a bad society."


When Hayden arrived at Crypto City, it was under siege. Congress
was lobbing mortar rounds. Morale was lower than a buried fiber optic
cable. Senior managers had become "warlords," locked in endless
internecine battles. "The term 'warlordism' has been going around for
years," said one NSA official. "It means that each deputy director acts like
a feudal warlord in a fiefdom. He will not give up anything for the greater
glory of the NSA mission. If we have something that requires taking some
of my stuff and putting in another deputy directorate, I won't do it. It's
like rearranging the deck chairs on the Titanic. The directors are, by and
large, afraid to tell the deputy directors to leave. They feel perhaps
unsure of themselves." The deputies, he added, "tend to stake areas out
like dogs marking their territory."
Another problem was that the Senior Policy Council, which advised
the director on major issues affecting NSA, comprised so many feuding


382
"warlords" that to reach agreement on anything was impossible. "I don't
know how anything gets done," said one member of the council. "There
are thirty-five of us in that room and we don't get anything done.
Anybody who was anybody was on the damn leadership team. It was
impossible for the director to get a consensus on anything."
Hayden also arrived to find the agency's financial system in shambles.
"The budget is one of his biggest problems," said an official in January
2000. "He doesn't know where it is, he can't account for it, that's what's
driving him nuts." Adding to the management problems were the
enormous technical challenges facing the agency at the dawn of the new
century.
Hayden candidly admitted that at stake was nothing less than the
survival of NSA. "As an agency, we now face our greatest technological
and analytic challenges”diverse and dynamic targets; nontraditional
enemies and allies; a global information technology explosion; digital
encryption; and others. Make no mistake, we are in a worldwide
competition for our future."
Hayden had been at NSA less than a year when the computer crash
came. It likely confirmed the worst predictions he had been hearing
about his agency's failing health. "The NSA used to have the best
computers in the world, bar none," said an official who had been briefed
on the crash. "Now they can't even keep them running. What does that
tell you? Do you know a modern company that goes off-line for four
days? They're struggling."
The agency that once blazed the trail in computer science, going
where the private sector feared to go or could not afford to go, was now
holding on to technology's tail for dear life. "Most of what they were
expert in is no longer relevant," said a former director. "Getting them to
embrace the new world has been traumatic. . . . All they're trying to do is
hang on and survive." Florida congressman Porter J. Goss, one of those
who did see the computer crash coming, was more blunt. "Believe me,"
he said, "it's patch, patch, patch out there. We no longer are capable of
doing what we used to do." Goss also said, "This should have come as a
surprise to no one. Indeed, the [House Intelligence] Committee has, for at
least three years, warned NSA and the intelligence community of
concerns in these areas."
"Signals intelligence is in a crisis," said House Intelligence Committee
staff director John Millis more than a year before NSA's crash. Like a
worried first mate seizing the ship's wheel from a captain who is headed
for the shoals, the committee began forcing change on NSA. Both Millis
and Goss had served for about a dozen years in the CIA's Operations
Directorate. Millis had also spent some time in the executive offices of
NSA. Now the two teamed up to reinvigorate NSA's sensitive Sigint



383
operations.
"We have been living in the glory days of Sigint over the last fifty
years, since World War II," said Millis. "Sigint has been and continues to
be the 'int' of choice of the policymaker and the military commander.
They spend about four or five times as much on it as they do on
clandestine collection, and the fact of the matter is, it's there quickly
when needed. It's always there. Or it has always been there. In the past,
technology has been the friend of NSA, but in the last four or five years
technology has moved from being the friend to being the enemy of
Sigint."
In the past, a major communications revolution”telephone, radio,
television, satellite, cable”might happen at most once every generation.
The predictable pace gave NSA time to find new ways to tap into each
medium, especially since many of the scientists behind the revolutions
also served on NSA's secret Scientific Advisory Board. Today, however,
technological revolutions”PCs, cell phones, the Internet, e-mail”take
place almost yearly and NSA's secret advisers no longer have a monopoly
on the technologies. "Increasingly," said Mike Mc-Connell, the NSA's
director from 1992 to early 1996, "we will have to deal with a much more
diverse electronic environment, cluttered not only with human
communications and sensor signals, but also with machines speaking to
other machines."
One major problem confronting NSA is a change in the use of
technologies throughout the world. Some of NSA's targets still use
traditional methods of communications”unencrypted faxes and phone
calls, transmitted over microwaves and satellites. As can be seen from
the intercepts surrounding Iran's attempt to acquire the C-802 missile,
NSA is still very capable of performing its mission on these technologies.
But other targets are switching to far more complex communications
systems”circuit encryption, fiber optics, digital cellular phones, and the
Internet. The problem is to spend the vast amounts of money, time, and
expertise needed to develop ways to penetrate the new systems, and yet
not to overlook the old ones.
"We've got to do both," said Hayden, sitting in his office. He had
walked into the middle of the problem when he joined the agency. "Do
more without giving up what you used to do. . . . Part of the world looks
like this now and is moving hell bent for leather in that direction, but in
this different part of the world it still looks like it did fifteen years ago.
And things of interest are happening to the United States in both
universes. . . . How do you do the new while the old is still important to
you”in a budget that doesn't allow you to create two Sigint systems, one
for the old, one for the new? You've captured the precise dilemma of this
agency."



384
Deputy Director Barbara McNamara outlined in stark numbers
another of NSA's key problems today: too much communication. "Forty
years ago there were five thousand stand-alone computers, no fax
machines, and not one cellular phone. ... In 1999 there were over 420
million computers, most of them networked. There were roughly 14
million fax machines and 468 million cell phones and those numbers
continue to grow. The telecommunications industry is investing a trillion
dollars to encircle the world in millions of miles of high bandwidth fiber-
optic cable." McNamara might have added that there were also 304
million people with Internet access in 2000”up 80 percent from just the
year before. And for the first time, less than half of those people live in
North America.
Not only is NSA spreading itself thin attempting to listen in on ever-
expanding modes of communications, the tasks assigned to it by the
White House, CIA, Pentagon, and other customers are also exploding. In
1995, the agency received about 1,500 "immediate" requests for
intelligence”known as ad hoc requirements. By the fall of 2000, the
number of such requests had already grown to 3,500”a 170 percent
increase. Analysts, said one senior NSA official, "brute force" their way

<<

. 67
( 118 .)



>>