Hadoop and Hana - the perfect interplay
Hana is clearly an important starting point for many of our customers. Although Hadoop also runs on low-cost servers, a significant increase in performance can be achieved through SAP's architecture.
And since time is perhaps more money than ever, we think the time has come for many of our customers to consider how they can get even more out of their existing infrastructure by combining Hortonworks and SAP.
From a purely technical point of view, one important advantage is that the data to be analyzed is stored in the server's main memory for analysis. The fact that no external I/O operations are required here and data access is based on a different pattern means that even complex analyses can be performed quickly.
Hana offers coordinated libraries for tasks within forecasting, planning, text processing, spatial analysis and business analytics.
A possible virtualized mapping of the systems makes the user independent of the hardware used, as long as sufficient performance is ensured.
Strengths of Hadoop
Hadoop scores high in Big Data analytics because of its ability to add missing information to datasets itself. Thus it can
z. e.g. add demographic data to customers' web logs before sending them for processing.
In addition, Hadoop excels at recognizing data patterns and performs intelligent evaluation of data clusters and the interconnection of different data types. The recognition of large but similar data sets is also important.
Hadoop also reliably detects data patterns in risk analysis, e.g., to detect anomalous operations in the banking sector, such as credit card operations, in a timely manner.
In areas where streamed data needs to be processed and analyzed directly, the interaction of Hadoop and Hana is a natural fit.
Useful fields of application
These include smart electricity meters, the analysis of data from vehicle sensors or from manufacturing plants. Here, very simple but mass data can be loaded directly into the working memory and analyzed with Hadoop.
The special architecture of Hadoop also favors the networking of several Hana instances. In addition, it also lends itself to the analysis of developments in social media. I
n contrast to streamed data, whose challenge is its sheer mass, the information gleaned from social media is less numerous, but more complex and unorganized. Here, too, processing with Hadoop with the help of Hana leads to a faster result.
For a company, this can mean that it recognizes emerging developments in the target group more quickly and can respond to them more agilely. In some industries, this is a significant competitive advantage, as it involves not just reacting to the target group, but real interaction, which in turn presents the company and its products in a better light.
The analysis of emerging consumer trends can therefore have a direct impact on the operating result. In combination with the in-memory DB Hana, Hadoop really comes into its own as a framework for scalable, distributed software.
It doesn't matter in which environment the solution is used - there are fields of application in every industry. Anyone who wants to carry out their Big Data analyses efficiently and quickly and include all source variables for a better result should take a closer look at the two-part solution.
