Hana-Hadoop direct coupling

With the SAP Hana platform and in conjunction with Sybase IQ as well as Hadoop, Hana and S/4 Hana Big Data applications can be made a reality. And that's using SLES for SAP Applications for SAP Hana and SLES for Hadoop with Suse Manager as infrastructure management.

How do the uses of SAP Hana and Hadoop differ? How do the database technologies fit together? Hana is known to have its strengths in processing structured data of up to 100 terabytes at extreme speed.

Hadoop, on the other hand, was not primarily designed for OLTP transactions. Hadoop's advantage is clear: it can process unstructured, mostly text-based, and semi-structured information as efficiently as possible with low hardware storage costs per terabyte, using so-called Hadoop clusters with several hundred standard servers.

Integration from a single source

About two years ago, SAP announced that it would push collaboration on Intel Distribution for Apache Hadoop as well as Hortonworks Data Platform, a free Hadoop distribution, to integrate Hana and Hadoop and co-distribute Hana with Hadoop.

The goal of Hana-Hadoop integration here is above all: to intelligently link immensely high volumes of data or mass data and information distributed across diverse systems in such a way that they can be analyzed or evaluated efficiently and in real time.

SAP has implemented a Hana-Hadoop direct coupling including a Hana-Hadoop connector (also using BODS, Business Objects Data Services, and ETL) and also provides a reference architecture for this.

This means that structured, semi-structured and unstructured data can be combined and processed. This includes, for example, web logs, machine data, text data, CDRs (Call Data Records) or audio and video data. There is also an enterprise architecture based on Apache Hadoop, which was developed by SAP and the Big Data specialist MapR.

It allows all data to be stored and processed on a single, multi-tenant Hadoop cluster, processing real-time analytics with SAP Hana. Like SAP, MapR is a Suse partner.

Hadoop Cluster

Based on the above, SAP Hana can be coupled with Hadoop. It should be noted that Hadoop generally prefers a physical installation for reasons of IO behavior.

This is especially true because Big Data involves processing data in the petabyte or even exabyte range. When setting up Hadoop with Suse SLES for Hadoop, the management tool Suse Manager can be used to advantage.

Suse's close collaboration with SAP ensures that both Hana and the Hadoop cluster deliver maximum performance - and can be used in an optimized way when both technologies are combined with the Suse platform.

SAP users who rely on Big Data and cloud computing demonstrably benefit from this close integration.