Analysis: How to Take Big Data Advantage of Oracle Database 12c

by   |   June 27, 2013 10:09 am   |   0 Comments

Evan Quinn 200x200 Analysis: How to Take Big Data Advantage of Oracle Database 12c

Evan Quinn

As the fanfare around the long-awaited Oracle Database 12c (“12c”) launch quiets down, DBAs, data business analysts, data scientists, CIOs and application portfolio managers will want to sit down and carefully think through the implications, particularly if you are mainly an Oracle shop.  The multi-tenant and thus cloud nature of Oracle Database 12c will provide IT—and by extension the business—some fresh ways of using Oracle databases in a big data context.

Before examining what Oracle already offers in terms of BI and analytics, and then considering how to use Oracle Database 12c to your big data benefit, one important note:  We limit the focal point here on Oracle databases, versus say SAP HANA, IBM DB2, Microsoft SQL Server, or any of the many NoSQL (“Not Only SQL”) databases like MarkLogic, MongoDB, Cassandra or HBase.  The reason why is each database works somewhat differently in public, private, and hybrid cloud scenarios, and each works a little differently in various BI and analytics use cases.  So unfortunately, the lessons you may learn and apply to make full use of Oracle Database 12c in big data contexts often will not carry over to other databases used in a big data context.  In addition, the Oracle Database family, including the subsequent releases of Oracle 9i, Oracle Database 10g, Oracle Database 11g, and now Oracle Database 12c, house more structured data in enterprises than any other database in the world.

Related Stories

Oracle stresses all-in-one database approach.
Read the story »

The database resurgence fueled by big data.
Read the story »

Innovative relational databases create new analytics opportunities for SQL programmers.
Read the story »

Data warehousing wafts further into the cloud.
Read the story »

Advice for evaluating alternatives to your relational database.
Read the story »

An Overview of Oracle Big Data, BI and Analytics Solutions
Heretofore Oracle has taken big swings at big data with several of its Oracle Engineered Systems, also known as appliances, such as Exadata, Exalytics and the Big Data Appliance.  The appliances have met with a fair amount of recent success.  Oracle CEO Larry Ellison crowed about that success in Oracle’s fourth-quarter earnings call on June 20 when he stated, “All our Exa products, Exadata, Exalogic, Exalytics and the Big Data Appliance and the Oracle Database Appliance, all had their best ever quarters.”  Even though Oracle’s hardware numbers were down 11 percent year over year for the quarter, and new software license revenues were up only 1 percent, big data-related growth had to be one of Oracle’s bright spots.

Oracle’s software lineup remains the same it has for the past several years, albeit with updates to better deal with big data and Hadoop scenarios.  The core products include:

  • Oracle Data Warehouse (which in fact runs on the Oracle database)
  • Oracle Essbase used mainly for OLAP
  • Oracle Hyperion which is most well known for enterprise performance management BI/analytics
  • Oracle Endeca, which some use for customer discovery analytics purposes
  • A long list of role-oriented and industry-oriented analytic apps
  • Oracle Business Intelligence

 

OBIEE, or Oracle Business Intelligence Enterprise Edition, packages a fair amount of the above mentioned offerings.  In terms of key partnerships, Oracle has a close relationship with Cloudera, which supplies the Hadoop distribution delivered through Oracle, and related services.  Oracle has also continued to invest its own Oracle NoSQL Database, which offers the key-value store technique often used in advanced analytics.  The Oracle NoSQL Database is included along with open source R, one of the most popular statistical programming languages, in its Big Data Appliance.  And some will cite Oracle Times Ten database, an in-memory database established years before SAP HANA came to market, that is the database engine behind Exalytics, as participating in the big data race.

But the Oracle database remains the keystone in any Oracle shop, and it is an inescapable element in the big data endeavors of those shops – either as a primary data source, or as a data warehouse, or for analytics processing, or as a data mart or dataset repository.   Where does Oracle Database 12c fit into this big data Oracle technology picture?   There are two cases of low hanging fruit.

A Rapid Way to Make Oracle Data Warehouse More Rapid
One initial big data opportunity most organizations will discover for Oracle Database 12c involves using 12c to your advantage for Oracle Data Warehouse (ODW).   In particular, long-standing ODW implementations grow stale and slow unless they are closely tended to, perhaps because of too much unnecessary historical data, or perhaps using what once was an appropriate but is no longer a well-fitted infrastructure, or perhaps because it is handling too many complex queries.

Data analysts, data scientists and DBAs will have the opportunity to split out individual data marts from an ODW implementation, and allocate the data marts as pluggable databases in the Oracle Database 12c container.  In Oracle Database 12c parlance, the container database is what allows for multi-tenancy.  It handles the background and overhead processes, such as memory or storage management, for multiple databases that are plugged into a single container.  But each pluggable database contains its own metadata, actual data, and any embedded code like triggers, wholly contained.  Also, each pluggable database may receive its own priority, so the container knows which database(s) should receive more memory to speed along processing, for example.

Using this technique, organizations may finely tune data ingest and query resources of ODW for various BI and analytics purposes.  And in Oracle 12c, “spinning up” a pluggable database in a container literally takes seconds, and changing priorities are a simple, real-time task.  Thus, one data mart might need all the resources the container can muster for one day mid-month, literally having its day in the sun in terms of resource allocation using Oracle Database 12c, and then you could move the data mart to the back of the queue for the rest of the month.  Similarly, during intensive warehouse periods of data ingest, DBAs may pop up the priority of the effected databases with more temporary horsepower.  Similarly, this splitting of a larger ODW into data marts may service the requirements of multinationals as well, in terms of local rules and regulations about data privacy, but also to put the data required for BI or analytics closer to the user base.

Oracle Database 12c to the Rescue of Development and Test
When data analysts, data scientists, and DBAs develop dashboards or models, BI and analytics assets that will become permanent business assets, there is a development and testing process.  Figuring out the design of a durable final dataset may take months.   Ensuring that the data ingest and refresh is utterly accurate and performs as necessary takes many iterations.  And the analytics visualization decisions will cycle through plenty of testing and refinement, hopefully with a sample of the final business user(s) in tow to ensure they are comfortable with the visualizations when they go live.  In short, producing small, medium, or big data that helps the organization’s decision-makers takes plenty of development and testing.

It is literally easy to provision new databases for data marts and/or full warehouses with Oracle Database 12c; your DBAs may need to find a few other tasks to keep them busy during your analytics project.  Need to offer up the warehouse to multiple developers and/or testers on tight schedules?  Cloning is a snap.  Want to replicate a production warehouse to a test warehouse to ensure testing is based on the latest data?  Similarly, replication has taken a big step forward in the world of Oracle databases with Oracle 12c.  Want to run an operational data store that directly reflects a production database, but is not the production database?  Oracle Database 12c offers a full set of technologies to make that happen, but also the features to ensure the production database still receives the resources it requires above the operational data store, even inside the same container.   In short, expect to reduce the time you estimate for BI and analytics projects using Oracle-centric data stores.

Use Your Imagination
I suggest that if you are an in an Oracle oriented database shop, and depend on Oracle technologies to a fair degree for BI and analytics, that you take an introductory course on Oracle Database 12c, even if you are not a DBA; hopefully the DBA will not feel too threatened.  By making the one brilliant move of separating the multi-tenant container from the individual databases, Oracle Database 12c has given BI and analytics professionals an entirely new way to think about their implementations.  I could come up with a number of other potential uses cases for Oracle Database 12c in the context of big data, but you know your company, your data warehouse environment and the likely evolving needs of the business for BI and analytics.  With a little knowledge about the new found flexibility and function of Oracle Database 12c, you, and the DBA, have the opportunity to make considerable improvements for your decision-making solutions.

Evan Quinn is founder and principal analyst at Quinnsight Research, covering big data, business intelligence, analytics databases, integration and data-as-a-service. Reach him at ebquinn@quinnsight.com. Follow him on Twitter: @evanquinn.





2f4030c6 41e3 4507 9efb 8be1e3aad05d Analysis: How to Take Big Data Advantage of Oracle Database 12c



Tags: , , , ,

Post a Comment

Your email is never published nor shared. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>