Oracle advances its Advanced Analytics strategy



Oracle recently unveiled a new option to Oracle Database 11g called Oracle Advanced Analytics for complex statistical and predictive analytics. The solution, built largely on the open source “R” language, is aimed at “quants” and statisticians looking to do serious complex analysis against medium to large data sets – in an Oracle database of course. The new offering is a conscious effort to enhance Oracle’s analytics arsenal as well as sharpen its focus on Big Data analytics. It is also an attempt to put Oracle (through R) on the data scientists’ radar.

R will enhance Oracle’s existing data mining capabilities

Oracle Advanced Analytics contains two components: firstly, Oracle’s existing data mining technology with a set of graphical tools for building predictive models; secondly, a new set of advanced analytics capabilities built around Oracle R Enterprise, which involves integrating R’s statistical processing capabilities within the Oracle database and conversely taking advantage of the scalability of Oracle’s database platform.

Instead of using Oracle Data Mining between R and the Oracle database, Oracle has now plugged R directly into the Oracle Database engine. The Oracle R Enterprise allow the R user to access database objects transparently (rather than having to embed SQL in their applications).

The difference between these two approaches is notable. R itself does not provide Oracle with a differentiated advanced analytics product. However, this differentiated (and quasi-commercial) integration of R and the Oracle Database potentially does because it moves R from a PC-based analytics engine to one that uses the massive scalability of the Oracle Database engine.

Keeping the data scientist engaged

Oracle Advanced Analytics ensures that in the future, data scientists working in Oracle shops will have little reason to consider other providers. Ovum has seen many organizations working with an Oracle (or even an IBM) database look beyond their existing provider when advanced analytics is required. For the heavy lifting, statisticians/data scientists instead choose tools such as SAS, SPSS, and Matlab, which sit separate from the database. This invariably means additional time and expense in terms of integration costs, intermediate data staging areas, and training. Oracle Advanced Analytics attempts to change that. For existing R users, the ability to use the familiar R graphical client tooling while working on a massively parallel database is a key value add.

R is for Big Data customers now, and other Exa-products soon

At the moment, a key use case for R is enabling large-scale predictive analytics against Big Data sets. Oracle already provides an Oracle R Connector for Hadoop as part of its Oracle Big Data Connectors product. This effectively provides R users with direct access to Hadoop infrastructure and HDFS data from within the R development environment. The benefit for R users is that they do not have to learn a new language or be trained on a new interface to work with Hadoop/HDFS data stores. Nor do they necessarily have to have intricate knowledge or skills in Hadoop for production applications.

Additionally, Oracle is now re-distributing open source R on its Oracle Big Data Appliance, specifically targeting Hadoop customers building enterprise analytic applications. The integration allows R users to write R scripts that run seamlessly across R, Oracle Database 11g, and Hadoop environments.

The integration of Oracle Advanced Analytics is progressing, starting with an initial focus on Oracle OBIEE. This enables an R user, for example, to build a predictive model and visualization, and push it out in a parameterized form to a report or dashboard. End users are also able to bring up specific calculations generated by the R engine in the context of other OBIEE interfaces.

Oracle will allow users to run Oracle R Enterprise scripts for Big Data analysis against data stored in Oracle Exadata Database Machine, boasting impressive performance gains in testing. However, integration with the Oracle Exalytics In-Memory Machine is not yet on Oracle’s immediate roadmap for Oracle R Enterprise. For now, Oracle Exalytics users can only view Big Data analysis results from Oracle Exalytics (as per OBIEE).

Should SAS be worried?

R’s most significant drawback so far had been the lack of a strong association with an enterprise vendor. Tibco S+ barely scratched the surface of R’s potential, but Oracle clearly has the potential and muscle to take R mainstream. However, displacing or even competing with existing SAS deployments will probably be out of the question for the moment. SAS has a strong legacy and a thorough understanding of business painpoints, which in the past has helped it sell higher-value, industry-specific predictive solutions rather than a set of generic predictive tools. R is surely on the way to that goal, but even with Oracle’s significant help it will take a long time and several hundred deployments to get to that stage.

However, R and Oracle do represent a force that SAS should start to worry about. While there will always be more customers unwilling to bet on one database and analytics vendor, a few might actually consider it more convenient to go all Oracle, especially existing R users that also have an Oracle database, and those organizations building their analytic infrastructure from scratch.



All Rights Reserved.

No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the publisher, Ovum (an Informa business).

The facts of this report are believed to be correct at the time of publication but cannot be guaranteed. Please note that the findings, conclusions and recommendations that Ovum delivers will be based on information gathered in good faith from both primary and secondary sources, whose accuracy we are not always in a position to guarantee. As such Ovum can accept no liability whatever for actions taken based on any information that may subsequently prove to be incorrect.