It Just Keeps Getting Better – SAP HANA as a Multi-Model Processing Engine
In my last blog post, “Back to Basics – SAP HANA and the Virtual Data Model,” I mentioned that the SAP HANA intelligent data platform continues to see more and more innovation with each passing year including the upcoming release of HANA 2.0 SPS04. In this post, I will explore some of the innovation that’s been delivered in recent years, with a focus on the more advanced capabilities of SAP HANA that go beyond the in-memory columnar structure and the value of the virtual data model.
While the blazing fast performance of SAP HANA is compelling, and the virtual data model provides a unique differentiator to provide value as a combined transaction and analytic platform, today’s businesses demand more to give a platform the title of intelligent data platform. You might have noticed that those of us from SAP work hard not to refer to SAP HANA as a database; there’s good reason for that. Of course, at its core SAP HANA has all the database management capabilities you expect in a database. However, with SAP HANA we strive to go well beyond core database management and have introduced three additional categories of innovation inside the same platform: 1.) application development, 2.) advanced analytical processing, and 3.) data integration & quality. While I have a deep appreciation for the application development and data integration & quality capabilities, my focus here is on the advanced analytical processing capabilities. In fact, I’ll broaden that one topic to the term multi-model processing as the capabilities go beyond analytic purposes.
I mentioned above that we deliver additional innovation inside the SAP HANA platform. The word “inside” might be the most important word in that sentence. If we take an honest look at a modern IT ecosystem, we can quickly observe that there is a proliferation of purpose-built technologies to solve diverse use cases. One can find all types of databases, and depending on the data structure (relational, graph, key-value, document, etc.), we can also find advisors who will make a case for each type. Some pundits will talk about a ‘Polyglot Persistence’ while others might suggest a ‘Multi-Model Database.’ It will surprise no one to hear that, in the end, SAP believes the suite always wins and that a single integrated suite of solutions is optimal. Ultimately, this means that rather than build individual technologies for each specific purpose, we should bring all those capabilities into the SAP HANA platform. This doesn’t just mean different types of data persistence in the same platform, but rather a fully integrated set of capabilities allowing for all different types of processing of that data regardless of format/type. To explain all of this, I’m going to return to an analogy like I did in my last post…
When I went away to college in the Fall of 1997, we were in the midst of a major revolution in communications technology. While I arrived with a brand new @iastate.edu email address, we still had a traditional land-line phone in our dorm room. I would call my mom once a week using a calling card, and I went to the computer lab every couple of days to email using a Unix command-line email client called Pine. By the time I graduated and started my career as an SAP Basis administrator, I had a brick of a cell phone and had moved on to a graphical email client. In addition, I was given a standard issue one-way pager, and if I was lucky enough to be on-call, I got a second two-way pager so I could see all the alerts from Remedy.
Beyond the communications technologies mentioned above, my bags were loaded up with other digital and analog devices. It’s interesting to look back on those days and consider all the different products we needed and how they have evolved today – including my now famous collection of mix tapes. To be honest, we were hauling around an awful lot of stuff that has now been crammed into a smart phone for us. Consider the following quick list of transformations:
I look back on those college days in the late 90s fondly, but it’s sure nice not to have to haul around so much stuff to accomplish what are now seemingly simple tasks. In a way, I’m glad every moment back then wasn’t digitally recorded (once again, I admit no wrong doing!), and we were able to live so disconnected. But like most people, I’m now addicted to the connection and can’t imagine a world where I can’t pull out my smart phone any time I want to connect with a colleague, capture a moment with the kids, or record my run.
Once again, that was a fun little diversion into the past for me, but there is a reason I brought it up in the context of SAP HANA. The point here is that a smart phone isn’t just a phone – what makes it unique is that it combines many different capabilities in one physical device. This is possible because the smart phone has a variety of technical features beyond a traditional phone including a data connection, a camera, a GPS receiver, an accelerometer, and much more. In the same way, the SAP HANA intelligent data platform isn’t just a database, and the multi-model processing capabilities I’ll discuss next are a key part of what make it unique.
So, I’ve made the claim that SAP HANA is an intelligent data platform because it offers capabilities beyond a typical database. Let’s examine that claim and review a subset of those multi-model processing capabilities at a high level:
- Predictive Analytics & Machine Learning – by embedding predictive and machine learning algorithms and connectors (i.e. R and TensorFlow) directly in SAP HANA, business users can get a new class of insights. By having these capabilities on the same platform with transactions, applications can enable users and machines to take action before or as events happen – not after they occur.
- Spatial Data Processing – for decades we’ve been focused on answering the questions of what, why, how and when. But in today’s world, we also need to answer the question of where. SAP HANA goes beyond just storing spatial data and enables spatial processing so that businesses can analyze their data in a whole new way and offer location-aware business applications.
- Graph Processing – graphs are a powerful way of modeling linked data and processing it to determine the effects of relationships. SAP HANA provides a mechanism to define a graph workspace on top of existing data, along with algorithms to perform a variety of common calculations. Graph processing enables uses cases like financial services fraud detection, manufacturing bill-of-material analysis, and utility outage tracing among many others.
- Text Analysis/Mining & Search – the combined capabilities of text analysis and mining plus search enable the discovery of insights from unstructured documents and text fields using built-in capabilities of SAP HANA. With these capabilities, businesses can identify and correlate information that was previously buried in complexity. They are also able to complement this with consumer-grade search capabilities that make the new insights easily accessible.
- Business Function Library – the SAP HANA Business Function Library (BFL) delivers various pre-built parameter-driven functions in the financial area. By implementing and linking the functions at the database kernel level, application developers are subsequently given the flexibility to build compound business solutions that have incredible performance.
- Document Store – for the ultimate flexibility in data structure for application developers, SAP HANA offers a JSON document store that allows data to be stored without a pre-defined structure. These collections of data can be joined with any other data type with full interoperability.
- Blockchain Adapter – blockchain is undoubtedly a hot topic whether you believe in its potential for business value or remain skeptical. The SAP HANA blockchain adapter connects SAP HANA to a blockchain network via SAP Cloud Platform blockchain services. This connection enables a consistent programming model across data that resides in SAP HANA and data that is distributed in a blockchain – allowing use cases that read from and write to blockchain via SQL.
In conclusion, while there are plenty of other tools in the market to address each of the above capabilities individually, I would argue no competing solution puts them all into a single platform to enable an intelligent enterprise like SAP HANA. Just like the smart phone demonstrates its value via the solutions built using its combined technical capabilities, the proof for SAP HANA is shown by delivering unique business outcomes leveraging its multi-model processing capabilities. This is true whether these outcomes are delivered with packaged applications like machine learning in SAP S/4HANA or are the result of customer/partner innovation like our 2019 Innovation Award Winners. As I said from the start, SAP HANA just keeps getting better, and we’re excited to continue the innovation journey with the intelligent data platform and ultimately help our customers run at their best.