What is SAS Scoring Accelerator?
SAS Scoring Accelerator is an in-database model scoring solution that deploys SAS analytical models directly to supported database environments. By executing model scoring where the data resides, it reduces data movement, latency and the complexity of deploying models into production.
SAS Scoring Accelerator benefits
Accelerate in-database model scoring
Automate data-scoring processes for faster insights. Publish analytical models into database-specific functions that are deployed and executed directly within the database environment.
Streamline model deployment
Faster deployment frees your analytics team to focus on new projects. By reducing the need to manually revalidate code for your models, you'll incur fewer labor costs.
Reduce scoring latency and complexity
There's no need to move data between SAS and the database for scoring purposes. This reduces the cost, complexity and latency of the scoring process. And it streamlines your analytic deployment.
Reduce data movement and improve governance
Reducing data movement and replication ensures data integrity, which is crucial for data governance. By scoring data inside the database, you can execute programs so that detail data is analyzed without crossing the database boundary.
Key features
Eliminate inconsistencies caused by excessive data movement and latency from conventional processes for scoring data.
SAS format library
This XML file is deployed once to the database system.
SAS Enterprise Miner Score Export node
Lets you export the model into files that are ready to deploy to the database for scoring.
Register SAS/STAT linear models to SAS Model Manager
Supports in-database scoring for SAS/STAT models.
Publishing client
Pushes the scoring function into the database. Once inside, a SQL expression is used to execute the score.
Integrated environment for tracking & monitoring model performance over time
Further streamlines the registration and validation of vendor-defined functions (VDFs) inside the database.
Available for multiple data sources, including Hadoop & Spark
See the Features List for a complete list of supported databases.
SAS Viya models
Select models created by submitting a CAS action or a SAS procedure, or by using SAS Model Studio, can be deployed in-database.
