CIO

9 key takeaways from SAP TechEd

SAP seeks to simplify app development on its cloud platform, while helping customers make the most of machine learning
Juergen Mueller (SAP)

Juergen Mueller (SAP)

SAP CTO Juergen Mueller outlined a series of improvements to the vendor's software platform at SAP TechEd in Las Vegas, the first of three big customer conferences the company will hold in as many months.

At the heart of many of these improvements is SAP’s Business Technology Platform, previously known as Digital Platform, a collection of data management and cloud technologies.

Together, the improvements provide a picture of how SAP plans to make it easier for businesses to develop applications linking their core SAP systems to one another and with data held on other platforms.

From a harmonised UI, to help with machine learning, the following nine announcements are likely to have the biggest impact on SAP-based enterprises.

Harmonising the UI with SAP Fiori 3

SAP is putting the emphasis on its ability to deliver end-to-end support for business processes such as design-to-operate, total workforce management, source-to-pay and lead-to-cash. Some of the tools with which it delivers that support were developed in-house, but others such as Concur, Gigya or Fieldglass were acquired, resulting in a hodgepodge of user interfaces and screen designs.

“For the design-to-operate process, you would experience the UIs of 14 different systems,” Mueller said.

That will change with the launch of SAP Fiori 3, the latest version of the company’s web-based interface design. SAP Fiori 3 will present users with a harmonised design and interaction model as the company rolls it out across its applications.

The UI is also available to enterprises developing their own apps on SAP, whether they’re targeting desktop or mobile devices.

Easing development with SAP Graph

On the back end, SAP is also in the process of adopting a single semantic data model across all its applications, with a unified API layer, SAP Graph, that is available to customers in a restricted beta test. The goal of SAP Graph is to simplify development that requires access to SAP data.

“It can be used by any application, running on SAP or non-SAP technology,” Mueller said.

As proof, he introduced Steve Wilson, vice president of product at Citrix, to show how Citrix uses SAP Graph to build micro-applications in Citrix Workspace using data held in an SAP database.

In response to a customer complaint at a recent hackathon about setting up access rights for Success Factors while developing an extension for SAP Cloud Platform, the company has simplified its Extension Factory and Cloud SDK. Now it should be as easy as pairing smartphone with a car using Bluetooth, Mueller said.

Bringing analytics to S/4HANA

SAP is also bringing its data analytics tools closer to its S/4HANA cloud platform, including a free SAP Analytics Cloud (SAC) tenant with each S/4HANA Cloud subscription. Users will still need a full SAC license if they want advanced capabilities such as planning or predictive analytics.

This means that enterprises can create “analytical stories” directly within S/4HANA using Fiori Launchpad, avoiding the need to switch between operational and analytical tools.

Leveraging machine learning

The company has also accelerated its adoption of machine learning in its applications.

In June 2018, Mueller said, the company promised to build 60 machine-learning cases into SAP applications within three years, but has progressed far faster than that. “By the end of this year, 2019, we will have more than 200 machine learning cases live,” he said.

Machine learning is now embedded in S/4HANA, C/4HANA (SAP’s customer experience suite), SuccessFactors, Ariba and Concur. “Soon it will be hard to find any SAP solution without machine learning embedded, and these 200 do not even include the ones that come with Qualtrics,” the customer feedback platform SAP acquired last in December 2018, Mueller said.

Introducing SAP Data Intelligence

Outside their use of SAP’s own offerings, the company’s customers have been slow to build their own machine learning applications, Mueller said.

To make things simpler, the company is releasing SAP Data Intelligence, an evolution of SAP Data Hub and SAP Machine Learning Foundation. This allows developers to build and train machine learning models from within a single tool, with access to a catalog of SAP and non-SAP data sources across the enterprise.

It tracks models through their full life-cycle from development to deployment, tracking who did what and when.

Data scientists should feel at home, Mueller said, because the system allows them to run Python scripts, work in their own Jupyter notebooks, and use their favourite machine learning frameworks, including TensorFlow.

Hybrid help

While Mueller had focused on cloud development, all these tools could be used by customers using SAP’s older ECC on-premises platform in a hybrid mode with the help of its Cloud Connector, he said.

Mixing and matching services from SAP Cloud Platform with those from hyperscalers such as Microsoft Azure is also possible, he said, and to help customers figure out how to do so safely, SAP has updated its reference architecture for building applications using SAP technology in others’ clouds.

“We assessed all hyperscaler services out there, and whitelisted the ones that you can use with neither the fear of being locked out of SAP innovations, nor the fear of being locked into a particular hyperscale offering,” Mueller said.

Improving Azure integration

One of the hyperscale architectures SAP supports is Microsoft’s Azure — inevitably, as both companies manage their own financials running SAP’s S/4HANA software on Azure infrastructure. With Microsoft’s head of cloud and AI, Scott Guthrie, Mueller discussed other companies doing the same, including Costco Wholesale, Coca-Cola and Daimler.

Microsoft and SAP are working to make it easier for enterprises to integrate the two companies’ cloud services, Guthrie says.

For example, he said, Microsoft has packaged its Azure AI Cognitive Services, which offers services such as speech and text recognition and face detection, in containers so that enterprises can integrate them with apps on their own networks without fear that data will leak out.

Building on the blockchain

Microsoft and SAP’s latest collaboration, though, involves blockchain technology, not AI: SAP’s blockchain integrated application portfolio can now connect to Azure’s blockchain services.

The idea is to help simplify the development of enterprise applications such as materials traceability, fraud protection or collaboration.

True pay-as-you-go

Cloud infrastructure is typically charged on a usage basis, but that hasn’t been the case for SAP Cloud Platform up to now. Admittedly, it does offer a consumption-based model in addition to its fixed-price subscription system, but that involves customers committing to buy “cloud credits” worth tens or hundreds of thousands of dollars up to a year in advance.

Next month, though, that’s changing, said Mueller.

“We are announcing the availability of a true pay-as-you-go model. This will first be available in North America, starting in October, and of course we will expand it in the next year. This means you have a real pay-as-you-go model for all solutions included in the SAP Cloud Platform Enterprise Agreement,” he said.

That agreement is also changing in scope to encompass new HANA, Analytics Cloud and related services as they become generally available, he said.

Application of the new usage-based pricing won’t be automatic. North American CIOs wanting to switch will need to contact their SAP account manager to make the change; those elsewhere in the world will have to wait.

Mueller promised more news at the SAP TechEd event next month in Barcelona.