Pennsylvania builds SOA for federated warrant search system
- By Rutrell Yasin
- Sep 12, 2008
Using service-oriented architecture and associated business process management tools, officials with the Pennsylvania Justice Network (JNET) successfully deployed a Federated Warrant Search engine that compiles warrant information from multiple systems.
JNET is the commonwealth's public safety and criminal justice information network. Established by executive order in 1998, the network was intended to streamline workflow through the integration and sharing of criminal justice information throughout the commonwealth.
JNET officials embarked on the SOA path about two years ago. An aging IT infrastructure and application silos hampered JNET's ability to give users a consolidated view of information. If law enforcement officers needed information on a suspect they would have to search one place to obtain a photo and a second place to find out if he was on parole.
'We found it was time to refresh our entire infrastructure and our approach to service delivery,' said David Naisby, executive director of JNET.
Naisby's team set out to enhance the JNET users' experiences through a new portal front-end, an enhanced security and provisioning model, and a SOA platform supported by Software AG's webMethods Enterprise Service Bus. JNET's secure Web portal provides access to more than 33,000 users throughout the commonwealth's 67 counties as well as federal and state agencies.
The first step was to assess existing business processes -- what was critical, what was not needed and what could be enhanced. The objective was to capture essential business needs, identify challenges, and implement solutions using SOA technology.
'Business requirements were very clear, application enhancements were well documented, and requests for these new services needed an avenue for rapid deployment,' Naisby said. 'We took a look at the requirements and the offerings of our SOA platform, and from there we hit the ground running.'
Using Software AG's modeling tools, Naisby's team was able to build out automated workflows and define business and technical requirements for the project sponsors, developers, and business analysts to approve. 'The modeling toolkits we received from Software AG were extremely helpful during the early stages of the software development lifecycle,' he said.
Working with SoftwareAG's professional services division also helped position JNET's technical staff to the point where they could tackle projects and use modeling tools to define what their architecture should look like.
The modeling software is really an extension to SoftwareAG's webMethods Enterprise Service Bus, said Tod Weber, vice president for the company's Federal, Aerospace and Defense Practice. The enterprise service bus combines application integration capabilities, high-speed messaging, support for business-to- business connectivity, legacy applications and event-driven technologies in an integration infrastructure.
Naisby noted that agencies are very protective of their data, so it was important to identify ways to give users the appearance that data is co-mingled, when in fact,it is individually protected by each governing agency in existing data silos. For example, data owned by the Pennsylvania State Police, the Department of Transportation, or the FBI is in effect owned by three separate agencies. Users do not want to search for data in each of these data repositories separately; they want Google-like search functionality.
'The enterprise service bus allowed us to tear down the walls that inherently separate data sets and search results. The bus allows us to take one piece of information, and using Web services, spawn subsequent requests to the correct data providers, retrieve the information, and finally assemble it for the user in one single interface,' Naisby said. The results are increased productivity, significant time savings, and increased accuracy of data by and between disparate record management systems.
Other Google-like functionality includes alerting users if words appear to be misspelled, offering search alternatives and recommending related information based on the initial query or results.
The ESB has also reduced search time dramatically.
Pennsylvania has four types of warrants issued by different authorities. A user of JNET can search the Administrative Office of Pennsylvania Courts' statewide judicial system for warrants issued from the bench. A second JNET search queries the Pennsylvania State Police for statewide criminal warrants. Using this data, the state police broker a third search for JNET and check the National Crime Information Center (NCIC) for national wants and warrants. Finally, a fourth warrant search queries the Pennsylvania Department of Public Welfare for outstanding child support and civil commitments.
Historically, this required separate searches at JNET, each taking up to 10 minutes, Naisby said. Now a user can execute all four transactions from one user interface, entering no less than two pieces of information, and get the results back in less than six seconds, he said.
While the infrastructure continues to mature, stakeholders, users, and business partners joining the network continue to increase. 'JNET must stay on the cutting edge of technology in order to ensure that we can meet the growing needs of current and future customers,' Naisby said. 'Scalability, reusability and cost-effectiveness remain at the forefront of JNET's business model, and SOA has served as a keystone for many of our recent successes,' he said.
Many people are responsible for the vision of JNET and its ongoing success, he added. Enterprisewide efforts and support from the governor's office have made JNET one of Pennsylvania's success stories.
Naisby pointed out, however, that SOA is only a part of the solution. 'We must not forget the importance of governance, policy, communication and the human element,' he said.