Databases Reference
In-Depth Information
incorporate the same data feeds in the warehouse, build out that schema appropriately,
and provide access to KPIs and data via their standard business intelligence tools.
Oracle Exalytics
Oracle Exalytics
is an engineered system designed to run the Oracle Business Intelli‐
gence Suite, Endeca Information Discovery, and Hyperion Planning applications. It is
a rack mountable unit that contains 40 processing cores and terabytes of memory. The
large number of cores delivers a unique visualization experience for the Oracle Business
Intelligence tools where results are returned as a mouse is moved across a dashboard
without the need to hit a “go” button.
Performance speed-up for queries and discovery occurs because needed data and data
structure is stored in memory. With the Oracle BI Server, results can automatically be
cached in-memory or a summary advisor can be used to persist aggregates in-memory
in the TimesTen for Exalytics database. For Essbase applications (including Hyperion),
designated Essbase cube subject areas are stored in memory. For data discovery with
Endeca, the Endeca Server is housed in memory.
The Metadata Challenge
On the one hand,
metadata—
or descriptive data about data—is incredibly important.
Virtually all types of interactions with a database require the use of metadata, from
datatypes of the data to business meaning and history of data fields.
On the other hand, metadata is useful only if the tools and clients who wish to use it
can leverage it. One of the great challenges is to create a set of common metadata def‐
initions that allows tools and databases from different vendors to interact.
There have been a number of attempts to reach an agreement on common metadata
definitions. In 2000, a standard was ratified that defines a common interface for inter‐
change of metadata implementations. Named the Common Warehouse Metadata In‐
terchange (CWMI) by the Object Management Group (OMG), this standard is based
on XML interchange. Oracle was one of the early proponents and developers of CWMI;
however, there has been limited adoption and metadata exchange capabilities remain
very vendor-specific today.
As noted earlier in this chapter, an emerging complementary solution—one in which
ETL into a single data warehouse is not the entire solution—is the leveraging of master
data management and data hub solutions. Today, most organizations are still a long way
from consolidated metadata, and when they have tried to do this as an IT best practice
project, they generally have not been successful. Such projects are usually adopted only
when delivered within a business intelligence project that delivers business value.
