SQLstream's RAMMS provides an elegant and high-performance solution for Rating, the activity that calculates actual service charges by applying pricing plan rules to service detail (usage) records. Software that performs the rating function is known as a Rating Engine.

Modern Rating Engine Practices and Requirements

Some modern rating engines, such as Portal Software's Solution42 engine, heavily utilize SQL along with high-performance RDBMS databases and proprietary processing modules. Such modules are readily and naturally expressible as plug-ins within SQLstream's RAMMS.

The heart of any rating engine is a classification and aggregation process to satisfy the following requirements:

1. Match each service detail record against rating plans, such as 700 minutes per month nationwide, plan XYZ
2. Classify each service detail record by usage type, such as “off-peak weekend call”)
3. Aggregate the records into tiers of usage, such as “first 700 minutes flat rate, thereafter $0.25 per minute"
4. Allow for particular per-plan discounts linked to combinations of service usage, such as “once the user has ordered and downloaded three games within a given time period, we will give them the fourth game for free”)

Rating Operations in Relational Asynchronous Messaging (RAM)

Rating operations are a natural application for RAM, since SQL's grouping, filtering and routing capabilities can, through SQLstream's RAMMS, be done in real-time with superb throughput. For traditional RDBMS, the transaction load would prohibit doing the operations in real-time.

SQLstream's plug-in extensibility enables easy expression of the most complex query plans. Long-term aggregation windows allow for easy calculation of long-term discounts, such as for aggregated monthly usage.

The benefits of SQLstream's RAMMS-based Rating Solution go beyond high-performance, since RAMMS enables integration of the rating process across many integrated services and applications. For example, the same relational queries that perform rating of customer records can also directly feed revenue forecasting queries, or simultaneously feed fraud detection systems, and so on.