MicroStrategy puts analytics in the hands of decision makers and leverages high performance features such as 64-bit processing, in-memory technology, and query optimization to deliver solutions quickly. Within the last few years two trends have created an opportunity to add additional performance solutions. First, the amount of data created, stored and analyzed has grown exponentially. Second, the complexity of the models involved in analyzing the data has increased as users become familiar with basic analytics and desire to run more complex models to yield deeper understanding.
One way to improve performance is to use native database functions and process the data in-database. While many MicroStrategy customers are familiar with the performance benefits of using native database functions, most databases do not have a complete library of functions for data mining, regression modeling, scoring, simulation and other types of more advanced analytics. Although it is possible to build models in individual databases, since many enterprises use multiple platforms, it would nearly impossible to have advanced functionality available across the enterprise and accessible by MicroStrategy unless a library of models could be deployed in all platforms. In addition, the models would need to be tested, certified, and easy to install.
Fuzzy Logix spent 12 years building exactly the type of solution needed and offers over 800 in-database models on eight database platforms (and counting). DB Lytix, installs in less than 30 minutes, contains a rich set of functionality, has never been beaten in performance, is highly scalable and is accessible via SQL statements. MicroStrategy can call the models with SQL or with deeper integration such as with apply function. Our in-database models are a vast enhancement of native functions.
The benefits have been proven by customers and include:
- Improving the speed of analytics by 5X to 100X (for example one of our customers experienced a 91% improvement in performance)
- Reducing data movement
- Running large complex models on big data (regression models with 11,500 variables on billions of rows)
- Reducing the processing burden of the Intelligent Server
- Allowing more complex models to be run by business users
In addition to the speed and performance benefits of using in-database analytics, the ability for business users to run analytic models on big data is also driving operational and organizational change. Traditionally, business users were dependent on model builders to run analytics. Once models are built, business users commonly request a variation of the model that only needs simple changes. Instead of innovating, model builders are resigned to running the same models over and over again with only slight variation because they own the analytic technology. Using in-database analytics and dynamic SQL, model builders can now build models and deploy them to the enterprise. The benefits are that model builders can focus on innovation and complex problem solving and business users can run analytics on demand. With in-database models from Fuzzy Logix, MicroStrategy users can leverage high-performance models to perform simple or complex analytics on big data .