Tech Vs. Terrorism

The FBI stumbled badly in modernizing its IT to help fight terrorism. Here's
how the bureau plans to get on track.

By Larry Greenemeier,  InformationWeek
June 6, 2005
URL: 
http://www.informationweek.com/story/showArticle.jhtml?articleID=164300083

The Federal Bureau of Investigation's Virtual Case File system was supposed
to be the future of the FBI's crime-fighting and anti-terrorism operations,
a Web-based case-management application that bureau investigators would use
to search and analyze all information relating to criminal and
national-security cases. Instead, it turned into an expensive lesson in
outdated IT practices, in the depth and breadth of the dysfunction that
plagued the bureau's IT management, and in the challenges that lay ahead.

Initiated in June 2001, the Virtual Case File system suffered from having to
adapt to the FBI's dramatic post-9/11 mission change, which called upon the
bureau to focus on preventing terrorism as much as fighting more
conventional crimes. Yet the technology also fell victim to much more
workaday problems, including a shuffling of FBI CIOs and project managers,
ever-changing project requirements, and an insistence on building the system
from the ground up as a customized application. The result was a $170
million bust that the bureau tested briefly, then put on the shelf
indefinitely.

Rather than trying to fix the Virtual Case File software, the FBI over the
past couple of years has initiated an extensive overhaul of IT management.
The first big test of this makeover is Sentinel, an electronic
information-management system under development that's designed to help the
bureau leverage newer, standardized IT. Sentinel will test more than whether
the FBI can finally get the file-sharing ability envisioned for the
abandoned Virtual Case File project. It also will reveal whether the FBI's
efforts in recent years to create a better-run IT department and a modern IT
platform have worked. CIO Zalmai Azmi sees Sentinel--the FBI's first
software platform built on a Web-friendly, services-oriented IT
architecture--as nothing short of a way for the bureau to finally break free
from the shackles of its outmoded IT systems.

"Sentinel is my flagship moving forward and paving the path for me in
delivering capabilities," says Azmi, who has spent six years as an IT
executive in the Justice Department. "With every service I deliver with
Sentinel, I'll be looking at my legacy systems and retiring applications."

As the FBI develops each phase of Sentinel, it will replace corresponding
legacy systems, the most significant being its Automated Case Management
System. Other applications to be retired include the Criminal Informant
Management System, Bank Robbery Statistical Application, and Financial
Institution Fraud and Integrated Statistical Reporting Analysis Application.
One predicted benefit of Sentinel is its support for XML standards, which
can ease information sharing within the FBI and with other agencies.

Sentinel represents an ambitious first test for the FBI's new IT strategy.
The FBI expects by midsummer to issue a request for proposals for Sentinel,
which Azmi says will be comprised 80% to 90% of off-the-shelf software that
will, among other things, let FBI agents and analysts share case files,
search a variety of law-enforcement and intelligence databases, and automate
workflow. By contrast, the Virtual Case File system relied on a great deal
of custom-coded software, a reflection of the bureau's build-first culture
and belief that off-the-shelf software couldn't meet its needs. Although the
FBI hasn't disclosed when it will award a contract, Azmi says the first of
Sentinel's four phases will be due 12 months after the contract is signed.
The time line for the remaining phases stretches over the next four years.

The Sentinel project follows FBI Director Robert Mueller's recent moves to
dramatically increase the CIO's power and influence. In February, Mueller
gave Azmi overall control of the FBI's IT budget and began consolidating
several operations under the CIO's office. This means that where Azmi this
year controls $248.7 million of the bureau's IT spending, in the coming
fiscal year he'll control more than $500 million. The FBI's overall budget,
which has risen from $3.3 billion in fiscal 2001 to $5.1 billion in fiscal
2005, is proposed to be $5.7 billion in fiscal 2006. If this is the case,
Azmi's office would be responsible for about 9% of the bureau's spending.

Consolidating IT spending may seem far removed from the front-line concerns
of agents' basic inability to access and share information, but Azmi sees
his expanded financial responsibilities as a significant step. Soon after
being appointed acting CIO in December 2003, Azmi called for an inventory of
the bureau's IT assets and created a master list of applications, networks,
databases, and other key IT components. "We found that one of the reasons we
have the stovepipes was because different technology was being developed by
different agencies within the bureau," Azmi says. The "acting" designation
was dropped from his title in May 2004, and he's held the CIO post longer
than anyone else since 9/11. "Every division had its own
information-technology budget, network, brand of computers, and software."

To see why Sentinel is such an important--and difficult--project for the
bureau, it's necessary to understand why information technology is so
strategic to the FBI's changing mission and why and where it has stumbled in
recent IT projects. In the wake of Sept. 11, 2001, it became clear the FBI
needed to radically change how it works to prevent terrorism after nearly a
century focusing on investigating crime. Lag time in communication and data
sharing was no longer acceptable. "The FBI realized after 9/11 that it
wanted to share data in real time," says Jeff Vining, a Gartner analyst
covering homeland security and law enforcement and a former FBI lawyer.

The FBI needed better technology to fit its expanded responsibilities.
"After 9/11, the mission changed to prevention," Azmi says. "In preventative
mode, nothing's happened, so you have to do thorough analysis, and you have
a lot of what-if scenarios. For that you need really good technology."
State-of-the-art technology is often easier to envision than implement, and
the Virtual Case File system isn't the only time the FBI has had mixed
results. The bureau's Guardian incident-tracking system, designed to find
potential connections among local police reports and FBI counterterrorism
efforts, isn't always synchronized with state counterterrorism databases,
according to a March report by the National Academy of Public
Administration, a nonpartisan group chartered by Congress to help public
organizations improve their effectiveness. Earlier this year, the FBI said
it had effectively abandoned its controversial, custom-built Carnivore
Internet-surveillance technology, which was designed to read E-mails and
other online communications during investigations of suspected criminals,
terrorists, and spies. Guardian is on a list of programs being reviewed for
possible retooling, Azmi says. Carnivore didn't make the list.

Amid all this, FBI Director Mueller hasn't wavered on the importance of
technology in supporting the bureau's mission. Shortly after being appointed
in September 2001, Mueller put into high gear an existing program called
Trilogy to upgrade the bureau's aging IT infrastructure, which suffered from
years of neglect. The first two phases of Trilogy, completed by April 2004,
went smoothly: providing a high-speed, secure network that lets FBI
personnel worldwide share data, including audio, video, and image files, and
replacing outdated hardware with more than 30,000 new desktop computers,
nearly 4,000 printers, 1,600 scanners, 465 servers, and 1,400 routers. The
Virtual Case File system was the third and most ambitious and complicated
part of the plan, dependent on a custom software project that would make the
bravest corporate CIO blanch.

Mueller, during testimony Feb. 3 before a Senate Appropriations
subcommittee, praised Trilogy, which has an estimated price tag of $400
million to $600 million, for improving the FBI's ability to connect with
other intelligence and law-enforcement agencies. Still, he lamented the
problems with the Virtual Case File system, Trilogy's crucial third phase.
"If agents had the VCF capabilities we envisioned, they could directly input
information into their computers, receive electronic approvals, and, with
the push of a button, upload information into the database where it would be
immediately available to others who need access to it--agents, analysts,
other federal employees, and state and local officials," he testified.

How did it go wrong? The answer is complicated and involves mismanagement,
poor planning, and a lack of open IT standards. It also was a case of an
organization not keeping up with the pace of change in the IT industry, as
off-the-shelf software caught up to what the bureau was paying outsiders to
develop from scratch. Mueller summed it up this way: "The pace of technology
has overtaken the development of unique software applications for the FBI."

Commercial software wasn't an option when Science Applications International
Corp. began work on the Virtual Case File system, says Mark Hughes,
president of the system and network solutions group at SAIC, the vendor
hired to build the system. "The case-management system was very complex," he
says. On the one hand, the FBI had to ensure that only authorized personnel
had access to certain sensitive data. On the other, there was a demand that
data be readily shared among agents and with other law-enforcement agencies.
"There are also a lot of legal requirements governing how data is accessed
and used," says Frank Perry, chief systems engineer for SAIC's system and
network solutions group. "The data's admissibility in court was a big
consideration."

Put another way, Virtual Case File "was to replace a bulky system that was
processed nightly," Gartner's Vining says. "They wanted [case-file]
information in their agents' hands in real time. They didn't want to ever
again have an agent in Phoenix writing a memo that wasn't shared with the
right people," Vining says, referring to a memo an agent wrote in July
2001--which wasn't widely shared--warning that terrorists might be training
at flight schools. But work on Virtual Case File had begun even before the
terrorist attacks.

In June 2001, SAIC won the Virtual Case File system contract, which
initially laid out plans for the development of a standard Web interface to
legacy systems used by FBI agents. "Three months later, 9/11 happened,"
Hughes says. "Within six months of that original contract, we were told to
forget it and work on a new case-management system."

Sept. 11 was the most-important force influencing the system's development,
but not the only one. The February 2001 arrest of counterintelligence agent
Robert Philip Hanssen, who for decades had been selling national defense
information to Russia, prompted the Justice Department in March 2002 to
issue a report reviewing FBI security programs. The report found, among
other things, a "pervasive inattention to security, which has been at best a
low priority."

SAIC says the project was plagued by a lack of consensus among users
regarding what they needed. The company spent the early part of 2002
interviewing agents and analysts to find out what they were looking for in
the system. At the same time, however, the FBI was reengineering its
processes altogether. "They were trying to respond to the recommendations in
all of these reports," Hughes says. That didn't bode well for the Virtual
Case File system. "Neither we nor the FBI knew what the system's
requirements would be," he says.

Another problem was the FBI's management churn. "In the three years we had
this project, they had nine program managers and five CIOs," Hughes says.
"Every time someone new came in, they hadn't participated in the decision
making to that point."
SAIC delivered its original Virtual Case File system in December 2003,
though FBI executives believed even as it was being delivered it wouldn't
meet their needs. Bureau officials concluded the system couldn't do actual
case management and could only do certain basic functions like automate
workflow, especially routing documents between staff and supervisors.

Doubts about the system grew to the point that the FBI in January 2004
created a separate data-sharing system, called the Investigative Data
Warehouse. It provides more than 6,000 special agents, intelligence
analysts, and members of joint terrorism task forces with a single access
point to about 50 databases, including the FBI's primary legacy Automated
Case Support case-management database, its Violent Gang and Terrorist
Organization File database, and various news feeds that provide English
translations of major international news articles.

The system let agents for the first time use analytical tools across data
sources to create a more complete view of the bureau's treasure chest of
data. The FBI also imported more than 60% of the data from its legacy
systems to the data warehouse. But even this successful effort was colored
by the looming failure of Virtual Case File. The Investigative Data
Warehouse "was a risk-mitigation strategy knowing that our case-management
system, or VCF, was going to be late," Azmi says.

The final verdict on the Virtual Case File system? The bureau spent another
$2 million this year to have Aerospace Corp., a nonprofit contractor, assess
the system, and it concluded the FBI should cut its losses and start over
with commercial products. The FBI also had several hundred users at three
locations test the system's document-routing capabilities in a way that let
the bureau know just how much of its IT investment might be salvaged.
Mueller estimates that $104.5 million of the $170 million investment in the
Virtual Case File system is considered "a loss." The networking and hardware
equipment can be repurposed as part of Sentinel.

So how can Sentinel succeed where past efforts failed? One factor is that
Mueller appears to have found a CIO he can work with.

When Mueller took the reins, the FBI didn't have a CIO position. By November
2001, the position was created and filled by former IBM executive Bob Dies,
who retired in May 2002. Mark Tanner served as acting CIO over the next
three months, until the bureau gave its top IT job to Darwin John, former
managing director of information and communications systems worldwide for
the Church of Jesus Christ of Latter-day Saints in Salt Lake City. John
lasted less than a year. In May 2003, Wilson Lowery, a 30-year veteran of
IBM and former chairman of IBM Credit Corp. and general manager of IBM
Global Financing, took over as the FBI's CIO. Lowery originally joined the
bureau in June 2002 as Mueller's special assistant in charge of overseeing a
reengineering effort that looked at, among other things, how the FBI used
technology and how it collaborated with other agencies. By December 2003,
Lowery likewise moved on.

Finding himself once again without a top IT executive, Mueller looked to his
days as U.S. attorney in San Francisco when he'd had occasion to work with
Azmi, who at the time was CIO for the Executive Office for United States
Attorneys. The two former Marines had designed and created a reporting tool
called Alcatraz for Mueller's office. Azmi had pulled off a successful tech
turnaround at the U.S. Attorneys' Office, building a Trilogy-like
infrastructure project called the Justice Consolidated Office Network, as
well as a victim-notification system and enterprise case-management system.
Azmi's history with Mueller has inspired a new level of confidence in the
CIO position. "I am a CIO who has a chair at the table, discussing the
mission of this organization," Azmi says.

Azmi has been given the power to consolidate IT budgets, develop an IT
architecture, and create a system of checks and balances designed to keep
projects on schedule and on budget. The IT architecture identifies all of
the FBI's IT systems, applications, networks, and databases on a master
list. Azmi also organized an enterprise architecture unit that includes
chief enterprise architect Wayne Shiveley, who in April completed the first
version of the bureau's enterprise architecture--something mandated for all
federal agencies. All future IT projects must be consistent with Shiveley's
architecture plan.

Another approach the FBI has taken to improve IT management is its Life
Cycle Management Directive, which governs how projects are managed from
their inception. The goal is to centralize IT assessment, using seven
"gates," or review points, that serve as the mechanism for management
control and direction, decision making, coordination, and confirmation of
successful performance. "Do the money and the time match? Are we at our
performance goals? You have a measuring tool now," Azmi says. "Anytime it
deviates 10% or more, that's when we bring in project management." The
bureau is evaluating the health of all 479 of its IT projects in order to
provide data that will help FBI managers understand the costs, schedules,
and risks associated with each project.

Given all of the scrutiny following the Virtual Case File system, one legacy
will most certainly include shorter, better-managed contracts. "One of the
things we don't want to do is another giant contract of four years," Azmi
says. "It just doesn't make any sense."

The importance of the FBI's revitalized approach to IT management can't be
underestimated, one former FBI agent says. The past missions of the bureau
and the U.S. intelligence community didn't prepare it for a war on terrorism
that requires tracking individuals and small groups and making connections
among them. "The amount of data and our ability to deal with it has never
been tested like this," he says. Nevertheless, the FBI knows now that it's
in the information business. As such, the former agent says, "If you don't
make IT the core of your business processes, you're not going to succeed."

No one knows this more than Azmi, who has seen firsthand the cost of failed
IT projects, in terms of time and money. In some ways, the Virtual Case File
system was a victim of the bureau's haste to make up for lost time. "There's
a lot of preparation that's going on in Sentinel that didn't go into VCF
because we were: 9/11, let's get it done," he says, snapping his fingers
quickly three times. While the sense of urgency is still there, "the key now
is to do a proper transition from the legacy system to the new system."




You are a subscribed member of the infowarrior list. Visit 
www.infowarrior.org for list information or to unsubscribe. This message 
may be redistributed freely in its entirety. Any and all copyrights 
appearing in list messages are maintained by their respective owners.

Reply via email to