Real-time reconciliation checking maintains continuous checks and balances over flowing data, performing continuous reasonableness and sanity checks.

For example, it might continuously check the total value or volume of transactions in a time window of five minutes for any given customer account. As an example, a rule for acceptability might specify that that value or volume be no more than one standard-deviation away from the running average over the last twenty-four hours.

It might also require continuous cross-checking of running totals of moneys transferred at regional, corporate, customer, and account levels. Alerts and alarms could be generated whenever money values failed to be properly and timely accounted for within configurable monetary ranges and time windows.

RAMMS's Direct Relevance

It is natural to express reconciliation checking, reporting, and error alerting as views in SQlstream's RAMMS for the following reasons:

• Required checking logic is naturally expressed in declarative SQL.

• The original transactions can constitute immediate message streams that RAMMS is built to process.

• The complexity of reconciliation constraints is simplified in very natural ways by using SQL views.

SQLstream provides easy-to-use facilities for chopping up data streams and regrouping or aggregating, along single or multiple dimensions.

The SQLstream solution allows for:

• Reconciliation checking to be performed in parallel as an overlay running alongside the regular, mainline processing that is being monitored

• Easy-to-generate alerts in real-time, piping the error stream into an alerting plug-in as soon as errors are detected

Such checks can be added to safeguard the correctness of the processing or business logic against software defects, logical errors or omissions, oversights and glitches.

RAMMS facilities address the core needs of reconciliation reporting:

• Continuous SQL filtering, grouping, aggregating, and multi-dimensional aggregation

• Built-in alerting mechanisms

• Error views – to create streams of erroneous records for reprocessing later

• Continuous cross-checking through using continuous joins and external database joins

• Parallel execution of processes and sub-flows, with automatic optimization

• Replaying and drilling-down by replaying with additional error-checking views enabled

• Allowing for arbitrary re-use of views

• Securing access to valuable aggregation data and error data

Beyond (or perhaps because of) the natural applicability of SQLstream's RAMMS facilities to reconciliation reporting's core needs, the SQLstream solution provides a previously-unavailable level of clarity, transparency, and responsiveness.