SAP data and analytics: in with the new, don't squander the old

SAP uses its Enterprise software heritage, buying power and guile to offer a highly-competitive data and analytics platform. Part of SAP's analytics approach is respecting the power of classic technology, even as it continually modernizes its platform.

khan-kazmaier.png

SAP's Irfan Khan and Gerrit Kazmaier

Credit: SAP

When you think of SAP, you probably think of Enterprise Resource Planning (ERP) software. Or maybe, if you're focused on industry news, you think of all the software companies and platforms SAP has acquired over the course of this decade, like SuccessFactors, Ariba, Concur or, more recently, Qualtrics. Do you think of SAP as a database and analytics company, though? Perhaps not, but you almost certainly should.

Also read: SAP HANA does Big Data...with ERP, CRM and BI savvy

That idea has been in my own consciousness since the early says of SAP HANA, but it solidified last week when I met with Irfan Khan, President, Platform and Technologies for SAP Global Customer Operations (GCO) and Gerrit Kazmaier, EVP, SAP HANA and Analytics. Khan and Kazmaier were in New York to meet with several of SAP's financial services customers. While they were here, they sat down and illuminated for me how much of the data innovation at SAP has its roots in the major acquisitions the company made in the 2000's. While Mr. Khan himself came to SAP as an executive from one of the acquired companies and may therefore have some bias, I found the explanation very sound.

From agnostic to interested party

In the early 2000s, the 1990s and years prior, SAP was purely an Enterprise software company. The software was open and flexible, and could run on various commercial database platforms, including Oracle and Microsoft SQL Server. And while it worked hard at developing its own "Business Warehouse" (variously branded as SAP BW and NetWeaver) SAP similarly accommodated a range of third-party BI platforms.

By the end of the last decade, SAP's approach to database platforms changed immensely, as the company acquired Enterprise BI player BusinessObjects in 2007 and Enterprise database contender Sybase in 2010. With these two acquisitions, SAP shifted from an assembly-required approach for implementing its software to one of providing a turnkey solution to customers that wanted it.

Technology acquired, technology leveraged

Mr. Khan himself came from Sybase, where he was Senior Vice President & CTO. That shows a leadership continuity from the Sybase days to now, but the continuity shows up in the technology itself, as well. Keep in mind that when SAP acquired Sybase, it got not only the classic Sybase operational RDBMS, but also Sybase IQ, one of the industry's first columnnar data warehouse platforms. And while, yes, these products are still available as SAP Adaptive Server Enterprise (ASE) and SAP IQ, they aren't just legacy products on the price list.

Kazmaier made it very clear how this all comes together today, in SAP HANA Cloud Services. Specifically, he told me that SAP's Data Warehouse Cloud platform combine's the Sybase/SAP IQ engine technology, HANA's memory acceleration and the structures pioneered in SAP BW. Kazmaier also explained that the semantic model of SAC (SAP Analytics Cloud) is highly influenced by BusinessObjects Universes and their paradigm.

Professor Plattner

If the fidelity to technological heritage pleases you, you might be interested in an SAP-related anecdote of mine, from back in 2015. At the launch event for SAP S/4HANA at the New York Stock Exchange, I had the privilege of meeting with SAP co-founder and Chairman Hasso Plattner who, rather incredibly, sat down and described to me the major motivations behind HANA. 

Plattner patiently explained that because of ERP database schemas' extreme normalization, the number of tables involved in any one transaction can be huge. Given that fact, and the realities of disk access physics, the amount of time required to commit an ERP transaction safely with a disk-based database can be immense. That pain point was the driving force behind HANA: by creating an in-memory database, the latency involved in disk seeks could be eliminated and ERP transactions could be far faster.

Operational analytics

Another motivation behind, and hallmark of, HANA was to create a database that could accommodate both transactional/operational (OLTP) and analytical (OLAP) workloads. By doing so, businesses could eliminate the delay and effort involved in extracting data from the transactional database, transforming it, and loading it into an OLAP cube, all before any analysis can take place.

Interestingly, what made both of these breakthroughs possible was the columnar technology that came out of Sybase IQ. Columnar storage allows for high degrees of compression, which in turns allows gobs of data to fit in-memory. In addition, columnar databases work extremely well for analytical workloads since the strategy of storing a single column's values all together makes them much easier to aggregate than is possible in a conventional row store database.

More deals, more stuff

Of course, Sybase technology isn't the only cornerstone of the SAP data and analytics suite, nor was it the only important acquisition. For example, the SAP Leonardo platform leverages both Apache Hadoop and Apache Spark and in 2016, SAP acquired AltiScale in order to gain the expertise necessary to implement Hadoop in the cloud.

Also read: SAP confirms Altiscale acquisition
Also read: SAP Introduces Spark-based HANA Vora
Also read: SAP unveils its Data Hub

Leonardo includes data visualization too, expertise for which SAP got with the BusinessObjects acquisition as well as the buyout of Roambi, a specialist in mobile BI, also in 2016. Then there's AI and machine learning, for which relevant acquisitions include KXEN in 2013 and Recast.AI in 2018. 

Also read: AI applied: How SAP and MapR are adding AI to their platforms

Partnerships are important too, right up to the modern day. At its SAPPHIRE NOW conference in Orlando in May, SAP announced an expanded partnership with Apple that integrates that company's CORE ML on-device machine learning technology with the SAP Cloud Platform SDK for iOS. The partnership also entails expansion of SAP app offerings on the Mac.

Then and now

Once upon a time, SAP was mostly about about business software. It was text heavy, ran on PCs and required database and analytics technology from other companies. Today SAP has its own database, optimized for Enterprise software workloads; its software runs not just on PCs, but in the cloud, on iPhone, iPad, and even the Mac. Analytics is embedded into the platform, in which open source technology is highly leveraged throughout.

Also read: SAP ups the ante for its Analytics Cloud and Cloud Platform

One thing that doesn't seem to change at SAP is the context and motivation for analytics provided by the company's expertise in a growing range of Enterprise software domains. SAP is serious and competitive around data in its own right. But because it's laser-focused on applications, implementations and use cases, SAP's approach is often more customer-centric than its analytics competitors'. Messrs. Khan and Kazmaier are articulate messengers for SAP around that mission.