FBI showcases Trilogy, information sharing

Senior FBI officials today pointed to the recently completed Trilogy network and their progress in creating a data warehouse as the leading projects in their IT overhaul.

Wilson Lowery, executive assistant director for administration, said the March 28 deployment of the first phase of the Trilogy project at 591 sites included fielding 22,000 workstations, 2,612 switches and routers, 622 Ethernet LANs and 291 servers at FBI locations worldwide.

The Trilogy network relies on a four-tier infrastructure with redundant connections and an asynchronous transfer mode frame relay backbone.

To bolster security in the network, Lowery said, 'We can see if anyone in the organization is going into files they are not allowed access to.'

The Trilogy network later this year will host the Virtual Case File system, a browser-based application that will let FBI agents store, inspect and correlate data about criminal cases and national security investigations. The bureau plans to deploy the VCF on Dec. 13.

Bureau officials last year scrapped the technical approach they previously had taken to creating the VCF, which involved building graphical user interfaces on top of the FBI's existing IBM 3270 mainframe-based, green screen Automated Case File System.

A team of FBI agents reviewed the GUI approach to overhauling the system last spring and concluded that it amounted to 'putting lipstick on a pig,' an official said. Last summer, bureau systems specialists responded to FBI director Robert Mueller III's directive not to automate obsolete business processes by holding meetings with dozens of subject area experts brought in from across the agency to plan a better system.

As a result of that process, they designed VCF to be based on an Oracle 9i database and combine several FBI forms into a unified VCF file.

The VCF will replace not only ACS but also individual databases that agents have built on their own to track cases. 'The problem is we haven't had an IT structure; we have been inefficient in the use of resources,' a senior FBI official said. 'The VCF will be a model for the replacement of other stovepipe systems.'

The FBI's decades-old personnel system is a key candidate for overhaul, Lowery said.

Lowery said he was surprised to learn how backward FBI systems were when he joined the bureau last year. 'We were starting much further back than I envisioned.' he said. In many cases, the original documentation for FBI systems is no longer available, so changing them requires reverse engineering.

As it creates the VCF, the FBI is moving to a collaborative environment in which documents, images and other types of information will be stored once and hyperlinked. As a result, the bureau will avoid problems such as the loss of files that hobbled the prosecution in the trial of the 1995 bombing of the federal building in Oklahoma City.

The first release of VCF will replace the ACS and an intelligence system called Intel Plus, a senior FBI official said during a press briefing today at FBI headquarters.

The next release will incorporate the bureau's database of telephone intercept records, its Integrated Intelligence Information Application and its Communications Assistance for Law Enforcement wiretapping application. The bureau will release additional versions every 60 days after Dec. 13, the senior official said.

'We have from 45 to 170 applications out there [to replace],' the senior FBI official said. 'We are starting with only five.'

Lowery described how the agency's new information warehouse, built on an Oracle 9i database with a capacity of nine terabytes of data, will be expanded beyond its current 23 million document size as information from Afghanistan and Iraq flows into it. Officials said they expect to expand it to 50T by July and 100T by December.

Officials plan to partition the TID warehouse when it grows larger than 10T because the analytical tools the agency uses with it become slower when the system grows larger, Lowery said.

The FBI itself has gathered about 95 percent of the information in the TID warehouse, while the rest has flowed in from other sources, including the news media.

Lowery and other FBI officials shied away from using the term 'data mining' to describe how agents and other officials would exploit the data in the warehouse, preferring instead to use the term 'advanced analytical tools.' They said their data warehouse would not be related to the Defense Advanced Research Projects Agency's controversial Total Information Awareness data mining project.

The core of the warehouse's terrorism resources will be a system called the Terrorism and Intelligence Data Information Sharing Data Mart.

The bureau plans to connect the data mart to state and local databases, telephone records, Pentagon databases, Drug Enforcement Agency and Bureau of Alcohol, Tobacco and Firearm databases, the State Department's visa database and other agencies. Some of the connections will retrieve information only and will not disseminate it to other agencies, and some of the bureau's TID data will stay in the bureau because it includes privileged grand jury records and other data that can't be shared, an FBI official said.

Lowery said that connecting state and local databases to TID is 'some way out,' because of problems in maintaining access control among the systems.

Information from the TID will flow upward to homeland security analysts and intelligence systems as well as to the Terrorist Threat Integration Center, officials said. However, CIA analysts will not have direct access to the VCF and TID systems, officials said. The FBI will provide some terminals to be in the TTIC and other secure locations.

The TID data mart is part of the Intelligence Community System for Information Sharing, a framework of systems and data in the intelligence world that the intelligence community CIO Alan Wade oversees.

The TID received its certification as a secret system last October and its certification as a top secret/sensitive compartmented information system in December. About 300 officials now use TID, officials said.

The bureau has adopted commercial data analysis tools to exploit the content of TID. Some of the tools include the Chiliad Business Intelligence Suite from Chiliad of Amherst, Mass., which the bureau used to build its Athena Search System; ClearResearch from ClearForest Corp. of New York; RetrievalWare from Convera of Vienna, Va., and Infoworkspace by Ezenia Inc. of Burlington Mass.

Changes and additions to the Trilogy system have increased its cost from an initial level of $450 million to $596 million, not all of which Congress has provided, Lowery said. The bureau plans to slow or stop some other projects to complete the Trilogy and VCF systems, he said. The ICSIS provided the $50 million cost of creating the TID warehouse, officials said.

Some officials in the FBI's 66 Joint Terrorism Task Forces will have access to TID data, officials said, based on their security clearances.

When the projects are complete, one official said, 'We will finally know what we know.'

Also today, FBI director Mueller buttressed the agency's intelligence capabilities by unveiling the new post of executive assistant director for intelligence. Maureen A. Baginski, formerly signal intelligence director at the National Security Agency, will hold the job. Mueller also announced a new Office of Intelligence, to be led by Steven C. McCraw, a 20-year FBI veteran.