From Zero to Clarity: SAS Compute & CAS for Low-Code Users
In this post, part of my From Zero to Clarity series for low-code users, we’ll break down two core processing engines in SAS Viya: SAS Compute and SAS Cloud Analytics Services (CAS). You’ll learn what they are, why both exist, and when to use each one so you can confidently make the most of your SAS Viya environment.
What Are SAS Compute and SAS Cloud Analytic Services?
First, let’s define each engine.
Select any image to see a larger version.
Mobile users: To view the images, select the "Full" version at the bottom of the page.
The SAS Compute server is the default engine and executes traditional SAS code. Data is processed row-by-row on a single machine, typically in a single-threaded or lightly multi-threaded fashion.
In contrast the SAS Cloud Analytic Services server, or CAS, is a distributed, in-memory, massively parallel processing engine. CAS is designed for large data volumes, advanced analytics, artificial intelligence, and machine learning. It processes data in parallel across multiple nodes simultaneously, making it ideal for workloads that require heavy computation or repeated access to large datasets.
CAS also allows data to be shared in-memory across multiple sessions, users and even different SAS Viya applications, avoiding repeated reads from disk and enabling collaborative, real-time analysis at scale.
Why Do We Have Both?
So, why does SAS Viya offer both SAS Compute and CAS servers?
The SAS Compute server exists because many workloads simply don’t require distributed processing. It starts quickly, has minimal overhead, and works well for routine tasks, data preparation, and smaller data sets. For many users, the SAS Compute server delivers fast results without the need for any complex infrastructure.
CAS, on the other hand, is built for scalability. When you’re working with big data — typically 50 gigabytes or more — or running complex models, simulations, or machine learning pipelines, CAS provides the power to handle these resource-intensive jobs efficiently.
Having both engines allows the SAS Viya platform to handle a wide range of processing needs — from simple to highly complex — and ensures that users always have the right tool available. Best of all, SAS Viya manages these engines behind the scenes, so you don’t have to worry about switching manually.
When Should You Use Each?
Let’s talk about when you should use each engine.
Use the SAS Compute server when:
- You’re working with smaller or moderate-sized data.
- Your programs run quickly and don’t require heavy computation.
- You’re generating one-off reports, simple queries, or routine data preparation tasks.
Use CAS when:
- You’re working with large data volumes.
- Your code requires multiple passes over the same data.
- You’re running advanced analytics, machine learning, or complex iterative processes.
- You need to leverage cloud-native storage.
- Multiple users, applications or processes need to work from the same in-memory data without duplicating or reloading it.
Summary: The Hybrid Power of the SAS Viya Platform
In summary, the SAS Compute and CAS servers each serve a purpose. SAS Compute gives you speed and simplicity for small to mid-sized tasks, while CAS gives you scalability and power for large, complex analytics.
The real power of the SAS Viya platform is that you don’t have to choose manually. SAS Viya manages the handoff seamlessly — running your code on the most appropriate engine based on your data and/or what you submit. This hybrid approach gives you flexibility, performance, and simplicity, no matter how your workload evolves.
Closing
Understanding the what, why, and when of the SAS Compute and CAS servers helps you design better processes, optimize performance, and get the most from the SAS Viya platform. Review this excellent post Simplifying SAS Viya Part 1: Choose Your Server for a deeper, more technical comparison — including how these servers differ from the traditional SAS 9 server.
Which server will you use next?
Find more articles from SAS Global Enablement and Learning here.
