SAP is no stranger to the BI and Big Data world. Its home grown BI software, its acquisitions of Business Objects in 2007 and Sybase in 2010, and its own development of HANA and HANA Vora make it a true contender.
Also read: SAP HANA does Big Data...with ERP, CRM and BI savvy
Also read: SAP Introduces Spark-based HANA Vora
But HANA itself has always been a bit controversial. With a reputation for pricey-ness and an approach of putting everything in memory, it has seemed impractical, engendering skepticism around true adoption, beyond its being bundled into deals that were going to happen anyway.
Meanwhile, British SAP partner Centiq is today releasing the results from a survey it conducted on SAP HANA, and the results were surprisingly positive. Centiq says the survey canvassed 250 HANA users, across numerous industries and sectors, including Financial Services, Manufacturing, Retail and Distribution, and Technology and Telecoms in the UK.
Among the results: 92% of respondents said HANA reduced IT infrastructure costs; a further 87% stated it saved business costs. 72% of respondents whose HANA deployments are live said they implemented within a six month period.
Ask any industry veteran how many times they've heard "SAP" and "implemented in six months" in the same sentence and you're likely to hear boisterous laughter. Plus, the party conducting the survey is not exactly disinterested, and results can skew. But 250 customers across various industries seems like a tough customer cross-section to game. So what gives?
Hear them out
I'll admit to having been a HANA cynic for a while, but that changed for me when SAP S/4HANA -- a version of the company's ERP suite engineered to run on HANA -- was released in February, 2015.
Also read: SAP's S4/HANA master plan: The lingering questions
Also read: SAP takes ERP in-memory
SAP invited me to attend its big S/4HANA launch event at the New York Stock Exchange and even said yes to my audacious (and somewhat naive) request to interview SAP's co-founder, and once co-CEO, Hasso Plattner.
Plattner is a somewhat eccentric figure and travels with the personal detail that you might expect of a person of his wealth. But Plattner is also a lecturer in Enterprise Platforms and Integration Concepts at the Hasso Plattner Institute for software systems engineering, based at the University of Potsdam.
So our conversation was pedagogical, with Plattner explaining HANA and the reasoning behind it to me, drawing diagrams and enthusiastically teaching. Plattner was patient and generous with his time, providing me, for the first time, with a perfectly distilled explanation of why HANA was so important.
ERP, and especially SAP, are intensely busy, chatty relational database applications, because ERP databases have a ridiculous number of tables, and there's no such thing as a simple update. As enterprises that use SAP grow, that chattiness only increases. And because of that, conventional relational databases can be real bottlenecks. As it turns out, HANA was designed to take on this very problem.
Can HANA handle big data, at least to a certain scale? Sure it can. But where it really shines is in two places: intense transactional/operational workloads and low latency/operational analytics. HANA's in-memory design makes the transactional work less contentious and makes in-place analytics practical.
With this application of HANA in mind, the survey results make sense. In fact, another of the survey's findings was that HANA's appeal is no longer only analytics: 74% of organisations deployed HANA for its high levels of data compression. A further 72% rolled out the in-memory database to save on IT infrastructure costs. After all, even if HANA itself is expensive, customers using it may not have to throw as much hardware at SAP to get it to scale or as much storage to make everything fit.
Suddenly it all makes sense
In this light, yesterday's announcement from SAP that it's releasing SAP BW/4HANA, a HANA-only version of its BW (Business Warehouse) product almost looks like a no-brainer
And if SAP really is going to acquire Altiscale, as VentureBeat reported last week, that would make sense too, because a cloud Big Data company can probably help HANA cloud and its in-memory-database-driven enterprise mission be more successful.
By extension, SAP's data and BI acquisitions over the years now look like a great strategy, even if it took the better part of 10 years to prove it out.
There are only a small number of companies (maybe as few as four) with the resources, enterprise focus, and timeless core business model to pull this off. Keep this in mind the next time "disruption" comes along -- sometimes it takes history and gravitas to put it to work contextually and successfully.