Any reuse of existing systems is clearly an excellent priority.  The biggest challenge to overcome when utilizing existing systems stems from the legacy or heritage development processes that built those systems over a number of years.  In most cases the data models of these historically important systems are very different as the applications were developed and deployed over a long period by different developers using different tools and technologies.  The result tends to have been the implementation of silos of data that are not uniform.  In order to have systems of record that can ensure that the collection of ARRA mandated information and data formats there is a strong case to be made for technology that can cleanse, provide quality assurance, normalize data to a set of standards and then deliver that resulting data to a consuming application that relies on this level of consistency.  Tools need to work across a number of different types and styles of data repositories in order to collect the data and map the data to the desired forms.  When this is done to highest degree feasible then the business analytics tools that need to be used to perform reconciliation will have accurate input and the reports that are generated will reflect a complete view of the transactions involved in the dispersal and spending of ARRA funds.  Business analytics have to be accurate and that means that the source data input has to be absolutely consistent and accurate.  Please see the IBM Information Server tools.

A terminology can serve as the method for identifying needed improvements to existing systems and to quickly developing interfaces between those systems based on specified "meanings".

It seems the point made above on existing Federal, State and Local systems is one of the more important aspects to consider when designing a central system for collecting and reporting on payment and performance data - especially considering the data is coming from hundreds of unique sources, at all levels within the Recovery Act Fund disbursement/recipient chain.

Agencies at all levels have invested significantly into their existing systems and processes. Many of which have voiced that they will be stretched significantly thin in terms of financial and human resources in order to execute against the new reporting and tracking requirements. Understanding that all agencies are working hard to implement infrastructure that will enable the tracking of funds-flow and usage (along with performance) down to the sub-recipient level, it would help greatly if these agencies did not have to make costly investments into their existing systems or processes to be both compliant and accurate with their ARRA reports.

The central system in discussion should have a robust capability to collect and digest data from the hundreds of disparate systems/sources that make up the appropriation flow - Federal Bureau > Recipient > Sub-Recipient.  In other words, the system should map to existing systems and processes today and not require agencies to make big changes on their side.  Additionally, the system should have the capability to than apply rules-based processing to ensure the data is both compliant and accurate based on both OMB ARRA requirements and individual agency financial management requirements , while facilitating exception management where discrepancies might occur. In this scenario, the system is driving collaborative, real-time reconciliation that is supported by a consistent environment for data collection, matching and reporting. This would ensure all data reported into the system is reliable. By setting up a collaborative structure like this, we would have a central system for both collecting and reporting more precise information on Recovery Act funds, but also a critical and central source for analyzing the audit trail at a very detailed level.

Web-based COTS certainly exist today that provide the vehicles for collecting data from a multitude of reporting entities with virtually any format (comma delineated to XML) or BAU system that they choose (e.g. grants, accounting, ERP, etc.) without requiring costly systemic changes or investments into new technology. These options should be a consideration so that agencies can avoid re-inventing their systems today.