Replacing A Legacy System

ARC is a PCI-certified, B2B company providing financial services (including processing and settlement of about $80 billion in financial transactions annually) as well as data and analytical products and services for the travel industry. Its solutions are based on both legacy systems and a growing portfolio of SaaS and web-enabled, real-time products.

The Challenge

ARC decided that its legacy core processing system could no longer be easily extended and expanded to position the company for future success in the marketplace. At the same time, it was facing an urgent customer need to expedite financial reporting. In response, ARC launched an effort to replace its mainframe and COBOL-based core system with a newer J2EE/Oracle-based one.

The human aspect of the project was important as well. Since many team members lacked the appropriate skills to support the new platform, ARC faced the challenge of either turning over staff for new hires or launching major retraining effort. ARC chose the latter approach to offer employees the opportunity to transition while retaining often irreplaceable deep domain knowledge. The project objectives therefore also included developing team and individual skills in the new platform and related software engineering best practices.

Given the critical nature of the project and the risks associated with transitioning to a new platform and acquiring new skills, augmenting teams with external expertise was a key to increasing probability of success. Mike Russell, leader of ARC’s Product Engineering group, contacted Construx and arranged for advisory and training assistance in both project management and software engineering best practices.

How Construx Helped

Construx worked with ARC to build an integrated road map to remove obstacles and reduce risk. In the initial phase, Construx helped clarify the project scope, refine the architecture, develop next-stage comprehensive estimation models, and identify additional potential risks that might need to be mitigated. We also provided software engineering best practice training on estimation, facilitated Java training, and provided extensive onsite software engineering and technical coaching to team members.

With Construx’s support, ARC used a combination of estimation models to create a more accurate estimate. This revealed that the project could not provide the urgently needed expedited reporting functionality in a timely manner. ARC decided to charter a separate team to implement the urgent functionality in the legacy system, while keeping the project team focused on moving to the new technology.

Construx continued to provide support and assistance to the project team in numerous ways as the new technology project progressed. Major activities and deliverables included:

Architecture workshops to identify major technical questions or risk and create an action plan to answer them.
Identification of the software development lifecycle that would best meet the needs of the project.
Project training and management coaching as the team moved to a more agile software development lifecycle.
Requirements training and coaching as the team decomposed business requirements into more detailed functional and non-functional requirements.
Assistance in prototyping a vertical slice of the system to evaluate technical risk and answer requirements questions.
Increment retrospectives to identify lessons learned and mark incremental course corrections.

End Result

ARC’s early visibility into the issues with transition to a new technology/platform and the related project estimate allowed the team to achieve both the most urgent business goals and simultaneously begin updating the legacy system. Construx’s legacy lasts through today as more legacy systems are converted, the opportunity for existing team members to transition continues to pay loyalty dividends, and additional Construx training has been added to address specific needs over time. Construx’s vendor and technology agnostic best practice approaches provided an ideal addition that did not require vendor or technology changes to be effective, whether in the original project or follow-on ones that continue to leverage the approaches.