SAP HANA: Powering Next-Generation Real-Time Analytics

eWEEK’s Q&A with SAP executive board member and technology lead Vishal Sikka gets to the nitty-gritty of SAP’s HANA in-memory database.

SAP HANA is SAP AG's implementation of in-memory database technology. HANA DB takes advantage of the low cost of main memory (RAM), data processing abilities of multi-core processors and the fast data access of solid-state drives relative to traditional hard drives to deliver better performance of analytical and transactional applications. It offers a multi-engine query-processing environment, which allows it to support both relational data—with both row- and column-oriented physical representations in a hybrid engine—as well as graph and text processing for semi-structured and unstructured data management within the same system. eWEEK Senor Editor Darryl K. Taft sat down with Vishal Sikka, SAP executive board member and technology lead, for a detailed discussion on SAP HANA.

Where did the idea for SAP HANA come from?

SAP's co-founder Hasso Plattner asked a fundamental question: If we were to design SAP R/3 from scratch, given the incredible advances in silicon and software technology, what would it look like?

Hasso had been building real-time systems since starting SAP, enabling companies to work in entirely new ways and creating a whole new industry. The three-tier client server architecture pioneered by SAP helped companies do business at scale—across customer orders, product shipping, and HR and financial processes—and brought in an era of great productivity gains. This was "real time" in 1992, bringing multiple departments and regions onto one system.

But because of the underlying architecture, these systems had to pump data into different databases to make sense of it. The limitations in the underlying architecture were aggravated by a general and massive explosion of data, brought on by social, mobile, Web and machine data entering the enterprise. Information was only accessible to the experts; it was always immediately outdated; the process of collecting and making sense of data was too slow. User expectations had also changed. They wanted direct access to the data; they wanted information in real-time; they wanted it to be a simple and beautiful experience to look at and interact with data.

It was clear that the current architecture would not serve the next 20 years of enterprise software. The divide between transactions and analytics was too big. We needed to figure out a way to bring real-time transactions and real-time analytics together on one database.

The answer was SAP HANA—a modern data platform that runs massive, parallel multi-core processing on one copy of data that is all in main memory. This is a radically simplified architecture that kills batch programs, captures data, and powers interactive and instant analytics on transactional data. It informs every decision in real time, allowing live collaboration and supporting the 'always on' mobile lifestyle, instead of after-the-fact reporting. This is an architectural advancement that delivers unprecedented speed of data analysis, solving some of the world's greatest problems at magnitudes-of-order improvement across industries, from health to security, science, finance, engineering, design and beyond.

What was your role at SAP in conceiving and driving HANA?

In 2009, Hasso and I met and discussed his concern that innovation had stalled at SAP. He challenged me to lead an intellectual renewal at SAP, renewal that would transform SAP internally as a company and would transform the entire software industry. It became very clear that the intellectual renewal had to be based on products that would galvanize everything we do.

I saw clearly that this had to be a fundamental breakthrough. It could not be incremental improvement on existing products. It had to be truly revolutionary and an innovation which would become the core of everything we do going forward. The hardware market was changing significantly, and we wanted to provide a modern data platform that puts OLTP and OLAP together, doing massively parallel multi-core processing on one copy of data that is all in main memory, but how do we create this real-time platform? We knew we wanted to combine transaction and analytics, but how do we timelessly evolve all of our older applications to this new platform? How do we attract new kinds of developers, such as people from startups and people working to build Web- or mobile-based applications? How do we renew the old but also bring entirely new applications—our own and our partners'—all on this common platform? How could we bring breakthrough new innovation to our customers without disruption, without compromise?