Faster Analytics, More Detailed Predictive Modeling for Large Working Datasets Create a Shared Memory Pool with RNAcache

Predictive Modeling & Analytics

Products

Solutions

Resources

Predictive Modeling & Analytics

Many firms run modeling and simulation on large historical data sets to develop and test analytic models. The results are used to make business investment decisions that ultimately drive top line revenue. There is a significant amount of data reuse and data is read-only. Duplication/replication is the current method to provide data to all nodes, which is time consuming and slows time-to-result.

Large Working Data Sets
These large datasets are often used to drive many different analytics applications in parallel to test and validate multiple scenarios. Moving large volumes of data from storage to working memory on multiple nodes chews up bandwidth and slows model wall time and time-to-value. Users often wait for retrieval of data from storage and models can take days to run. The reaction, in many cases, is to adopt a constrained analytics path using a subset of the data or partial models, trading time for thoroughness.

RNAcache improvements:

Monte Carlo Simulation for Capital Markets
Business Risk Modeling
Reservoir Simulation and Seismic Processing
Geo-spatial and Image Processing

RNAcache provides a shared cache that is accessible by all nodes with cache coherency. All data is held in a shared memory pool giving users rapid access to all files. RNAcache scales with ease, addressing future capacity concerns.

Business Benefits
RNAcache provides 20-100X improvement in application run time, simplifies the modeling environment and increases the capacity for data growth. The shared RNAcache offers unmatched, seamless acceleration of business critical applications without requiring application tuning, costly server upgrades or adding layers to your infrastructure. Looming capacity constraints are addressed, extending infrastructure life and adding system scale at extremely low cost.