Large Financial Services Institution – Classifying Customer Transactions

SQLstream's RAMMS provided an elegant and cost effective management platform for their existing applications built to classify customer transactions in real-time. There are many reasons why such companies wish to classify customers’ transactions in real-time: fraudulent or suspicious behavior, customers placed on a “watch list,” cross-selling, error detection, continuous reconciliation checking, performance monitoring, and so on.  The list of possibilities is long. 

In this case, the nature of the classification is covered by a strict non-disclosure agreement, so we are not at liberty to share further information. The customer had already built for themselves a solution to process and to classify the transaction streams of a large number of its end user customers, using home grown algorithms represented in Java.

Our customer had followed the progress of SQLstream and had seen us transform the RAMMS concepts into products with over a million lines of carefully crafted code over a period of several years. The customer had bought into the concept of RAM and RAMMS as a natural paradigm for combating the complexity of managing and processing dynamic data assets.

The customer particularly valued a number of key elements of the SQLstream solution set:

• Adopting ANSI and Industry de facto standards rather than creating our own proprietary approaches

• Building a RAMMS in the image of RDBMS with the same Enterprise scale and carrier grade quality, security, performance, architecture, driver, adapter and standards support

• Tight integration with Java – our customer was and is a heavy proponent and user of Java

• Our vision of declarative stream processing and our vision of letting the RAMMS worry about achieving the required level of performance, reliability and quality of service

• Our strong support for adoption of XMI standards for meta data management including Import and Export of meta data and enabling the use of external metadata management tools

• Our support for dynamic changes where changes can be made to a RAMMS instance without taking the system down or stopping existing executing queries (a repository for managing the processing of dynamic data rather than a fancy compiler and programming language)

• Our support for Eclipse

We demonstrated how our SQLstream RAMMS platform would allow the customer to migrate their existing investment in Java stream processing. It would allow them to model their Java modules as a set of reusable components as plug-ins within our architecture, while embracing the SQL:2003 UDX standards for User Defined Transforms.

Our central value proposition was our management and declarative processing capabilities.

Some immediate benefits:

• Single place to manage and administer all stream processing

• Easy to extend the existing processing with standards compliant SQL processing

• Easy to combine mixed SQL stream processing with Java processing

• Easy reuse of data, meta data, queries and views across many applications

• Replacing a spaghetti code architecture with a conceptually clean dynamic data repository

• Access to SQLstream's professional services, expertise and world class development team to help ensure not only immediate success but future success

• Mobile devices and machines

• Much lower total cost of ownership going forward

SQLstream's RAMMS provided a future-proofed architecture with a lower cost of ownership and low impedance into the rest of the organization, due to our strong support for standards for both SQL and metadata processing.

Database Limitations

Databases are excellent repositories of historical information, but have a lot of limitations in terms of being able to react in real-time or near real-time to changes in such data. There might be potentially thousands of queries to be run or other processing tasks that are required to be performed upon new or changed service data. Here are some of the activities that might need to be performed:

• Data aggregation (making data more meaningful and actionable)

• Data validation, cleansing and correction

• Application and Data Integration across the enterprise

• Business Activity Monitoring

• Security monitoring

• Service Level Monitoring

• Continuous ETL

• Application and Infrastructure Alerting and Monitoring

SQLstream's RAMMS provides a much more scalable and responsive SQL-based engine for managing, reusing, transforming and auctioning data than RDBMS.