SAS announces cloud-native Viya 4
Cary, NC-based SAS is announcing today Viya 4, the next version of its cloud platform, to be released late this year. Viya 4 has been completely rearchitected and refactored for cloud deployment, manifesting as a collection of container-based micro services, orchestrated by Kubernetes.
This announcement quickly follows one yesterday, focused on SAS' partnership with Microsoft and its selection of Microsoft Azure as its "first among equals" cloud platform (to quote fellow Big on Data blogger Tony Baer).
Must read:
Primers
- Microsoft Azure becomes SAS's preferred cloud platform
- SAS is on the brink of generation change
- SAS wants to spread its footprint
- SAS embraces cloud, will its customers follow?
Continuously SAS
In a briefing with ZDNet, Bryan Harris, SAS' SVP of R&D Engineering, explained that the company will employ a continuous integration/continuous development (CI/CD) approach to its own release practices and will do so in a way that allows customers to deploy updated Viya container images in their own CI/CD pipelines.
To cater to differing customer profiles with respect to release adoption, SAS is establishing three release rings, each with its own cadence. Customers who are conservative in their adoption practices, especially those in heavily-regulated industries, can opt into the Viya "Long Term Stable" ring, which will deploy new releases every six months. Less risk-averse customers can join the "Stable" ring, which will see new releases on a monthly basis. Customers who prefer be on the bleeding edge can join the "Fast" ring, which will fully embrace the CI/CD approach, conveying new releases as soon as they are finalized.
Interactive, programmatic, automatic
The services in Viya 4 containers can be run interactively through graphical user interfaces (GUIs) or called programmatically via application programming interfaces (APIs). In addition, several automated services will be part of the platform, including automated data preparation, machine learning (AutoML) and model deployment.
Also: AutoML is democratizing and improving AI
As Tony Baer explained in his coverage of yesterday's Microsoft partnership announcement, Viya 4 will initially run, somewhat agnostically, on Azure Kubernetes Service, but the partnership anticipates integration of Viya with various first-party Azure services, perhaps including Synapse Analytics and Power BI, which have recently been revved and featured at Microsoft's (virtual) Build and Business Application Summit (MBAS) events, respectively.
Must read:
- Microsoft Build brings announcements for cloud data, analytics services, and intersection of the two
- Azure Synapse Analytics combines data warehouse, lake and pipelines
- .NET for Apache Spark brings enterprise coders and big data pros to the same table
Redmond, meet Cary
While Microsoft features its own Azure Machine Learning service, SAS has been in the analytics and machine learning game for a long time, and has great expertise in servicing such workloads at scale. SAS' Harris explained that the company has been working hard to ensure that Service Level Agreement (SLA)-caliber performance, comparable to what customers have become accustomed to with on-premises infrastructure, can successfully be had in the cloud. With the cloud's decoupled compute and reliance on cloud object storage (which is not known for its speed), this is no small task.
SAS and Microsoft both have multi-decade legacies, oriented towards enterprise customers and their mission-critical workloads. The cloud partnership between the two companies is therefore a natural one, and if the transformation SAS hopes to bring with Viya 4 is implemented successfully, the SAS-Microsoft tie-up may offer significant competition to Amazon Web Services, Google Cloud, and an array of pure play AI startups.