Infosys’ blog on industry solutions, trends, business process transformation and global implementation in Oracle.

« Solutions for Linear Asset Management - A Data Model for Definiing the Linear Assets | Main | Power of Empowering - The new perspective of ERP implementations »

Oracle Exadata and DataWarehousing Impact - Part II

The Oracle's strategy to handle big data and business intelligence together, has got 3 key players in the winning combination namely Oracle Database Enterprise Edition 11g R2, Oracle Business Intelligence Enterprise Edition 11g and Oracle Exadata Database machine. Few additional players play an equally important role and those being OBIEE Applications for ERP/CRM, and Oracle Industry Business Intelligence/Analytics applications.

In continuation of the Part I of this blog series which can be referred here, lets continue the journey deep into Exadata world.

Let me begin with a YouTube video that might just open up eyes for the need of Exadata, please refer to the Video ( for better insights on how Information explosion is a reality.

  • From a functional standpoint the critical needs for a Data Warehousing platform are following:
    Faster response time to queries for effective analysis
  • Near real-time query analysis without impacting OLTP systems - Data availability, and Latency
  • Hide complexity of underlying data sources - Data Integration, Data Virtualization
  • Easier maintenance and monitoring - Integration with central administration toolsets


Diagram 1 - Key pillars for Effective Data Warehouse


From the technical standpoint following are broadly the Data Warehousing platform needs:
A) Query Performance
B) Fast I/O and data transfers between storage and database servers
C) Flexible Partitioning, various Indexing Options and Effecient Cache management
D) In-Database Analytical processing capabilities and features - OLAP, Data Mining, Statistics etc
E) Effecient Compression techniques which can enhance not only storage options, but faster data scans
F) Support Massively parallel processing which is scalable to handle larger data sets
G) Data virtualization and integration capabilities - supporting variety of data source options

Here's what Oracle Exadata has in the offering, to manage the above mentioned functional and technical requirements for effecient Data Warehousing and Reporting Platform:

  1. Query Offload processing in the Storage
  2. Smart scans to increase query performance and eliminating the need for indexing as DBA's had common practice in past
  3. Smart Flash Cache in database machine vastly increases for various workloads
  4. Oracle Database support for Analytics features within the Database (Predictive modeling and other data mining and advanced analytics, Geospatial latitude and longitude stored as geocodes)
  5. Flexible partitioning, Bitmap indexing, join indexing, materialized views and result cache
  6. OLAP, Statistics, Spatial, Data Mining, Real-time transactional ETL, Efficient point queries
  7. Data intensive processing directly in the storage
  8. Hybrid columnar compression (effecient compression increases the user data scan rates)


In today's blog we will cover Hybrid Columnar Compression technique. I specifically picked this as the first topic, as there's two fold large impact that Hybrid Columnar compression technique provides namely

  • Ability to retain more data in-memory for faster query results, as compression allows for more space in-memory.
  • Large compression allows for storage space savings
  • In standard database the typical format of data storage is as follows:


    Diagram 2 - Storing rows of columns sequentially in standard databases


    The compression techniques look for redundant data as one of the key criteria to compress the data. If the above diagram is to be transposed into columns, and then looked at rows as compression there's a better likelyhood of finding more redundant data in same columns that can be compressed more effectively. Effectively the data is grouped by columns & then compressed. The diagram below shows how the transposed database table leverages the Hybrid Columnar compression. This can enhance compression by a factor of 10x-50x, which gives you more flexiblity to bring larger amount of data into your Cache


    Diagram3.bmpDiagram 3 - Storing the columns of rows as Columnar model of compression


    There are two modes in which compression techniques help
    a) Query Mode - used for data warehouses, with key focus on optimization for speed. The order of compression can be 10x with column based compression. The data scans/smart scans improve proportionally with the compression orders
    b) Archival Mode - for archiving the infrequently used data and key focus being on reducing the space. This works with the view that your data can be directly archived onto the tapes instead of via the storage. Additionally with columnar compression, it allows for more data to be available for querying which is otherwise going to be archived and then retrieval takes it own sweet time before it's ready for query.

    Next blog I will be covering Smart Scans and Smart Flash Cache options that really power pack the Exadata.


    1. YouTube video on Oracle Exadata need - Oracle Exadata, are you ready?


    Neat explanation , curious to read Smart Scans and Smart Flash Cache ...

    when will these topics available

    Post a comment

    (If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)

    Please key in the two words you see in the box to validate your identity as an authentic user and reduce spam.

    Subscribe to this blog's feed

    Follow us on

    Blogger Profiles