X
Business

Microsoft Azure becomes SAS’s preferred cloud platform

SAS goes all in on Azure as both conclude a strategic partnership covering go to market and product development. SAS analytics capabilities could be integrated across the Azure portfolio to provide an additional option for model lifecycle management.
Written by Tony Baer (dbInsight), Contributor
enterra-solutions-ml-cloud.png
Credit: Enterra Solutions

SAS and Microsoft are jointly announcing today a strategic partnership making Azure the preferred cloud platform for SAS's analytics portfolio. Although nothing is firm at this point, there is potential to integrate SAS with many Azure services. While SAS will support other clouds going forward, Azure will be first among equals.

For SAS, this is a coming out party of sorts as they transition towards going cloud-first. For Microsoft, it's another notch in their belt for positioning Azure as a more enterprise-focused cloud, coming on the heels of partnerships announced over the past year with SAP and Oracle.

The SAS/Microsoft Azure partnership extends to go-to-market and joint technology development. Existing customers will be able to get SAS from the Azure marketplace; the same goes for channel partners, who can resell from the marketplace. Additionally, both SAS and Microsoft field teams will have incentives to sell SAS on Azure. One can imagine that the commercial models for SAS offerings on Azure could look like Microsoft's Redis Labs partnership to co-market Redis as a premium alternative to existing Azure caching services.

For SAS, the deal spans the current portfolio although the highlight, native integration with various Azure services, will be reserved for the forthcoming cloud-native release of Viya 4. Watch this space. We discussed SAS's plans for transitioning Viya a few months back, and Big on Data bro Andrew Brust will have more to say about Viya 4 very shortly.

All of the announcements of SAS services on Azure are planned for GA in Q4 this year, and as noted below, there will be some additional options planned for the cloud-native Viya 4. Specifically, this encompasses SAS 9 (this is the classic SAS Analytics portfolio) and Viya 3.5 (the current version), and the forthcoming SAS Viya 4. At launch, the Viya 4 Azure service will include all of the capabilities from Viya 3.5 except Visual Investigator, which will come later. The big difference with the Viya 4 offering will be the degree of integration with other Azure services, and later on.

The wish list for Viya 4 on Azure

On both sides, potential targets are ridiculously fat. For Microsoft, there's the Azure Synapse Analytics data warehouse and data lake platform; the open source-oriented data science platform, Azure Machine Learning; the low-code/no-code Power Platform (including Power BI); along with front office applications including Dynamics 365 and Microsoft 365. On the SAS side, there's not exactly a shortage of assets either, as there is model lifecycle management to decisioning (decision management), forensic analysis, vertical industry models, and model lifecycle management.

The press release states that both "will explore opportunities" to integrate SAS analytics to the Azure portfolio. Rome won't get built in a day. At launch, Viya 4 will run atop Azure Kubernetes Service (AKS); use Azure Active directory for role-based access and authentication; and use SAS's PostgreSQL data server in place of Azure PostgreSQL. Neither party has provided any timelines for what will come after initial launch.

But here's one item topping our wish list. SAS Intelligent Decisioning, which is essentially a rules management offering for managing how and why decisions are made, would be a natural fit as an embedded decision support capability for front office suites such as Microsoft 365 and Dynamics 365 as well as with Power Automate on the Power Platform.

Under consideration for the next waves after this year's launch is integration between SAS Methods and Azure Machine Learning, and in turn, integrating SAS Methods to run natively in Azure SQL Database, Azure Data Factory, Azure Synapse, and other data platforms. An interesting possibility is adding SAS analytics to the Power Platform AI Builder, or packaging SAS dashboards as Power BI templates (and/or vice versa). And the list goes on with potential goodies like developing native SAS plug-ins to Microsoft 365 that would hit our wish list.

Partnership paths

For Microsoft, this falls into a pattern of partnerships with established enterprise technology providers. Last fall, Microsoft and SAP launched joint go-to-market for the Project Embrace program announced a few months earlier encompassing services, reference architectures, and roadmaps for migrating to SAP S/4HANA, its next-generation ERP suite, to Azure. In effect, SAP anointed Azure as the preferred cloud for S/4HANA customers. Also last year, Microsoft and Oracle announced plans to provide high-speed interconnects between the Oracle Public Cloud and Azure (links are now live in US East and West, the UK, Amsterdam, and Seoul). The guiding notion was the large enterprise customer base that uses Microsoft 365 backed with the Oracle database.

For SAS, the Azure partnership is designed to address the question we posed a few months back on whether SAS customers will follow its new embrace of the cloud. When long-time customers face platform change, all cards go on the table – staying with the incumbent is not always the automatic first choice.

While SAS is well-entrenched with quants and decision support specialists across large enterprises, open source alternatives have grown popular with data science grads entering the workforce. Over the past few years, SAS has been making peace with open source, but when it comes to generational upgrade, why should large enterprises stick with SAS as they embrace new AutoML and other data science services that are proliferating in the cloud? Plugging analytics, model lifecycle management, and perhaps industry solutions into the front office services that are already in use across the enterprise could be the place where SAS offers its answer.

Editorial standards