Blog: SAS Viya
SAS Viya™ presents a step change in flexibility and compute technology of SAS® software.
Today, SAS platforms are relatively rigid. Scaling up or out requires manual intervention on some level be it modifying a hypervisor, introducing a new grid node or addressing contracts with SAS Institute. As business direction is increasingly driven by analytics, flexibility and automatic elasticity are required. SAS Viya addresses these challenges.
SAS Viya processes data in memory. By eliminating reads from data set or database the classic latency in SAS processing is overcome. This enables immediacy of analysis with big data and streamed data for real time analytics. With SAS Viya, even in-memory processing for the data step is expected. This allows SAS users to concentrate their efforts on the benefits of analytics to the business, rather than the business of producing analytics.
Implementing SAS platforms is a routine task for Amadeus because we are deploying SAS software for someone somewhere, every working day. However, if you are tasked with installing SAS at your company, understanding the intricacies of configuring the intelligence platform can be a challenge when required to do it infrequently. It is simpler with SAS Viya, because there is no intelligence platform as you know it. There is a new microservice architecture. It provides a radically simplified implementation for cloud and on-premise deployments. Amadeus expects to deliver a faster time to analytics with SAS Viya. Under the covers SAS Viya has two kinds of logical server: CAS servers conduct the computing (superseding today’s LASR server) and web servers host the microservices.
SAS Viya will be the most open and accessible SAS solution there has ever been. Since the 1990’s the SAS Integration Technologies product has provided the API to SAS software. SAS Viya moves integration to open REST API’s. These can be used to harness SAS analytics as a service or to provide bespoke SAS capabilities. Extending the analytical capability of SAS is rarely required, but users come to SAS from many different backgrounds. With SAS at the centre of analytics in your organisation you can encapsulate Python, R, Lua, Java and Groovy runtimes from within SAS Viya.
It is important that SAS users are assured SAS Viya does not mean existing solutions must be replaced nor radical changes will be forced on your business. Also, SAS Viya is not SAS 10, indeed a further maintenance release of SAS 9.4 is expected around the end of this year. We think of SAS Viya as a new version of the SAS Intelligence Platform onto which future SAS solutions will be deployed, taking advantage of the flexible architecture and in-memory processing. *Update: SAS Viya 3.1 was released in September 2016. This first, generally available release includes Visual Data Mining and Machine Learning solutions.
Amadeus Software are authorised SAS resellers, providers of Managed Services, Consultancy and Training for SAS software. Contact us to learn how SAS Viya provides opportunities to your business through analytics.