One of the most expensive, challenging, and time consuming information technology tasks is the integration of multiple disparate databases across a variety of unique platforms. A new San Diego company, Crossflo Systems, Inc. is introducing an innovative and compelling answer to this major issue and has already attracted significant interest by major federal, state, and local governmental agencies.

While database administration incorporates scalability into the database design, it is less likely that database integration is a significant component of the overall design process. One of the main reasons for this is that a set of industry accepted standards for database integration does not exist. Even if those standards were available today, there would still exist the challenge of integrating existing legacy databases, including off the shelf vendor products, with those that support the new standards. For federal, state, and local agencies supporting Homeland Security initiatives, this is a particularly daunting challenge.

While most people understand the difference between different operating systems, there is much less awareness about how databases can differ. Applications and/or databases that are designed to support a similar function (i.e., payroll, customer support, etc.) may have very different database structures. Moreover, the way that data is

captured may also be dramatically different (i.e., using codes to represent actual values, etc.). These differences result in databases that inherently cannot easily share data. Most people have no awareness into the difference because while applications leverage a database to store data, the users of the application have little or no exposure to the database itself.

"Integrating databases is traditionally a very labor intensive project, requiring months of work by highly trained database administrators and developers and can cost hundreds of thousands of dollars," explained Renney Senn, President and Chief Executive Officer of Crossflo Systems, Inc. "Following the terrorist attacks on the United States and the new requirements of information sharing between various governmental entities, the resolution of this technical problem is an urgent matter. Our technology makes the process much easier, quicker, and less expensive while using a drag-and-drop user interface and leveraging the capabilities of Web Services, XML, and Java behind the scenes. This enables Crossflo to get the most out of a customer’s existing infrastructure."

Crossflo Systems’ technology, called the DataExchange Data Sharing Platform, consists of two components: the DataExchange Agent and The DataExchange Server. The DataExchange Agent connects to databases while the DataExchange Server acts as a hub that connects all the agent-database pairs that need to be integrated. This "hub and spoke" architecture differs significantly from the traditional approach, which is typically a multiple point-to-point integration effort where every database has to be integrated to

every other database. By utilizing a central hub, each database only has to be integrated with the hub, vastly reducing the number of integration points.

For example, a local law enforcement agency may have information that it needs to share with multiple Federal law enforcement agencies, as well as other local "first responders." Each of these agencies has databases into which the information from the local agency should be inserted, however each is likely to have different databases, and the databases are likely organized in different ways. The integration of data from the local agency into these various databases would have to take into account all the differences between federal and local database structures and differences between the local database structures themselves. Local agencies also need access to information from each of the federal agencies, and from one another. The challenge of integrating all these databases is complex and requires a great deal of time and effort from individuals capable of understanding the natures of the differing database structures.

Under Crossflo Systems’ solution, there is a central point, the DataExchange Server, which acts as the clearing house for information sharing. The DataExchange Server functions as a central repository and contains not only the data to be shared but also data routing information and schemas describing the structures of the various databases. The schemas are specific types of eXtensible Markup Language (XML) files called XSDs (XML Schema Definition). The DataExchange Agent utilizes a schema to create files containing the data that needs to be shared. These data files are also written in XML. Web Services are used to pass information between the DataExchange Agent to the DataExchange Server and on to other DataExchange Agents connected to the

databases that need the data. XML is increasingly used as a standards-based means of transferring information between diverse systems and applications.

The DataExchange Server can support an unlimited number of schemas. Each schema represents a common "link" between disparate databases. For each database, a mapping is created that associates the unique structure of the database to the common structure of a schema; any other database can exchange data with any other database if they both have mappings describing how each relates to the common schema. The mappings are created once, and used every time data is move through the system.

Crossflo utilizes a "publish and subscribe" model as part of the hub and spoke architecture. Source databases (i.e., databases that contain information to be shared with other databases) "publish" data to the DataExchange Server via the DataExchange Agent. Target databases (i.e., databases that are the recipients of data from a source database) also use the DataExchange Agent to "subscribe" to the publications on the DataExchange Server. The publish and subscribe model allows any number of combinations of publications and subscriptions; when combined with the ability to support unlimited schemas, the Crossflo DataExchange data sharing platform provides an incredibly flexible approach to sharing data.

Referring back to our previous example, the local law enforcement agency would simply have to set up a utility that states how the data to be shared from its database relates to the common schema (which the DataExchange Agent can create), determine which agencies should receive all or specific parts of the data, and when and how often the other agencies should receive this information. The DataExchange Server manages data sharing with the other agencies’ databases by sending the information to the

appropriate agencies at the appropriate time. As the "owner" of the information to be shared, the local law enforcement agency would have the ability to determine, down to the data-element level, what specific information is published to whom and when.

In order to match the other agency’s target database(s) to the common schema, the DataExchange Agent is utilized again. Written in the Java programming language, the DataExchange Agent’s graphical user interface is used to map the local agency database to the schema. This allows a user to create a mapping, by a simple "drag and drop" of the local agency database’s structure against the common schema, greatly simplifying the process. The visual graphical user interface of the DataExchange Agent has revolutionized the process of integrating databases, taking a process that could easily take months and shortening the time period to hours or days," explained Senn. This mapping is performed once, and used every time data are inserted into the target database.

Once the mapping of the local agency database to the common schema is accomplished, the information to be shared is transferred from the DataExchange Server to the DataExchange Agent. The data to be shared can be published to the DataExchange Server and then sent to subscribing databases via Web services (XML, SOAP, WSDL) and hypertext transfer protocol (HTTP(S)). At the receiving end, the completed mapping permits the information from the DataExchange Server to be integrated into the receiving target database, to be used as if the target agency generated the data in-house.

"Our technology allows for users to perform several different data sharing functions. Among data-sharing initiatives, it can perform a uni-directional extract, transform and load, bidirectional transfers, synthesis of two or more databases into another database, or a broadcast of data to several or more subscribing databases; all at a

very low cost," stated Senn. "In addition, since our technology works at the database rather than the application layer, there is no employee retraining necessary."

Crossflo Systems was formed in January of 2003 and currently employs sixteen individuals. The company was founded by Bill Mohlenbrock, M.D., with the original idea of developing a better way to integrate data in the healthcare arena. Once the building blocks of the company’s technology were developed, however, Crossflo Systems determined that one of the best potential markets for the technology was in the field of Homeland Security and the Department of Defense.

"As you may know, there is a Presidential directive that requires various governmental agencies to share information, which our technology simplifies immensely," explained Senn. "We have been focusing on Homeland Security and ‘first-responders’ communities which include local police, medical, and fire agencies. These agencies have an enormous need for this type of service and very little time to accomplish their objectives." Senn was clear that Crossflo Systems is sensitive to the unique needs of local agencies, many of whom are on very tight budgets. By focusing on federal grant opportunities, Crossflo Systems can enable these agencies to meet the requirements of information sharing without taking money away from the agencies’ operational requirements. At least four local agencies are using Crossflo Systems’ technology.

While Crossflo Systems has been focusing on first-responders and the federal government, civilian agencies and the Department of Defense also have information sharing needs which the Crossflo Systems’ technology can address. In fact, one of the

first installations of the Crossflo Systems technology was at the Space and Naval Warfare Systems Command (SPAWAR), resulting from a conversation between a Crossflo Systems employee and a technician at SPAWAR. Now that Crossflo’s technology has been approved for deployment anywhere in the US Navy after an intense yearlong testing regimen at SPAWAR, additional opportunities are arising quickly.

Of course, nothing about Crossflo Systems’ technology limits it to sharing information between governmental agencies and it could easily be used by commercial entities with the need to integrate databases. "We feel our market can be measured in the hundreds of millions of dollars," Senn remarked. While the need for this technology is not bounded to any one sector or industry, presently the urgent need to share information among governmental agencies makes this Crossflo Systems’ initial target market.