Wednesday Jun 02, 2010

Partition Wise Joins

[Read More]

Friday Apr 23, 2010

Oracle Communications Data Model

[Read More]

Thursday Apr 22, 2010

Collaborate10 – THEconference

[Read More]

Friday Apr 09, 2010

Auto DOP and Concurrency

[Read More]

Tuesday Mar 23, 2010

Big Data’s Killer App…

[Read More]

Thursday Feb 25, 2010

Auto DOP and Parallel Statement Queuing

[Read More]

Wednesday Feb 10, 2010

Data Model or Knowledge Model

[Read More]

Monday Jan 18, 2010

New Data Mining Blog

[Read More]

Monday Jan 11, 2010

Integrating Hadoop Data with Oracle Parallel Processing

[Read More]

Monday Nov 30, 2009

TechCast: Data Warehouse Best Practices – Dec 2nd

[Read More]

Monday Nov 09, 2009

Next Generation Data Warehouse Platforms

[Read More]

Wednesday Oct 21, 2009

Edition Based Redefinition

[Read More]

Thursday Oct 01, 2009

In-Database MapReduce (Map-Reduce)

[Read More]

Sunday Sep 06, 2009

11gR2 - What is in it for me?

[Read More]

Tuesday Aug 18, 2009

Partitioning and/or/with Exadata?

Got a little busy there with all the comments on the Netezza posting, but now we're back into some of the outstanding topics. This post is the next one on discussing some of the ODTUG session questions (see this post). One of the questions was about the use of partitioning and whether it is made obsolete by Exadata off-loading... In other words, should you look at one, the other or both? The answer is that you will want both, and there are a variety of reasons for that. First of all, on the query side you will hopefully be using partitioning (often range partitioning) for partition elimination. From an I/O perspective that looks roughly like this*:
Partition_pruning.JPG
In essence, partition pruning allows you to reduce a 5TB I/O operation to a much smaller I/O operation and therefore much faster return of the information. Compression is something that may or may not be used. In this example we are compressing the data and further reducing the I/O numbers. So far there is nothing new here, with Exadata however you will see a further reduction. After applying a smart scan, both the rows returned (remember Exadata is smart storage and actually knows rows and columns should be returned!) and the columns returned are further reduced. This is on top of partitioning. You will get something like this:
ExadatawithPartition_pruning.JPG
The conclusion from a query and I/O perspective is therefore that you will benefit from both.[Read More]
About

The data warehouse insider is written by the Oracle product management team and sheds lights on all thing data warehousing and big data.

Search

Archives
« April 2014
SunMonTueWedThuFriSat
  
2
4
5
6
7
8
9
10
11
12
13
14
16
18
19
20
21
23
24
25
26
27
28
29
30
   
       
Today