SAS for Real-Time Applications

Overview

The SAS language is well suited for expressing program logic; people have been using it for over 30 years. There are many applications for building predictive models and then deploying these models to provide scores "On demand" as integrated components in a larger system. The paper describes and explores several patterns for using SAS in a real-time or near-real-time context, and describes the performance characteristics of each pattern. 1. Stored Processes; 2. Web Services; 3. Raw Sockets; 4. Message Queues.