Oracle Essentials [Electronic resources] : Oracle Database 10g, 3rd Edition نسخه متنی

اینجــــا یک کتابخانه دیجیتالی است

با بیش از 100000 منبع الکترونیکی رایگان به زبان فارسی ، عربی و انگلیسی

Oracle Essentials [Electronic resources] : Oracle Database 10g, 3rd Edition - نسخه متنی

Jonathan Stern

| نمايش فراداده ، افزودن یک نقد و بررسی
افزودن به کتابخانه شخصی
ارسال به دوستان
جستجو در متن کتاب
بیشتر
تنظیمات قلم

فونت

اندازه قلم

+ - پیش فرض

حالت نمایش

روز نیمروز شب
جستجو در لغت نامه
بیشتر
لیست موضوعات
توضیحات
افزودن یادداشت جدید










9.5 Managing the Data Warehouse


Once
you've built a data warehouse topology, you could
deploy multiple Oracle databases to implement the data warehouse and
its data marts. Enterprise-wide warehouses are more common on Unix
servers and are beginning to appear on clustered (RAC) Linux
platforms. Smaller data marts are common on Windows and Linux. Many
organizations are increasingly consolidating data marts and
enterprise data warehouses on the more scalable platforms.

Oracle Enterprise Manager provides a common
GUI for managing these multiple instances regardless of the
underlying operating system. EM is browser-based with a multiuser
repository for tracking and managing the Oracle instances. (EM is
discussed in much more detail in Chapter 5.)

In warehousing, in addition to basic management, ongoing tuning for
performance is crucial. Enterprise Manager 10g
supports many of the automated diagnostics and tuning features in
Oracle Database 10g.

Within the largest warehouses and
data marts, you may want to manage or maintain availability to some
of the data even as other parts of the database are moved offline.
Oracle's Partitioning option enables data partitions
based on business value ranges (such as date) or discrete values for
administrative flexibility, while enhancing query performance through
the cost-based optimizer's ability to eliminate
access to non-relevant partitions. As an example,
"rolling window" administrative
operations are used to add new data and remove old data based on
ranges of time. A new partition can be added, loaded, and indexed in
parallel, and optionally removed, all without impacting access to
existing data.

Range
partitioning first became available as an Oracle8 option. Hash
partitioning was added to the Oracle8i
Partitioning option enabling the spread of data evenly based on a
hash algorithm for performance. Hashing may be used within range
partitions (composite partitioning) to increase the performance of
queries while still maintaining the manageability offered by range
partitioning. Oracle9i introduced
"List"
partitioningpartitions based on discrete values such as
geographies. A composite partitioning type of Range-List, such that
you can partition by dates within geographies, first appeared in
Oracle9i Release 2.


Data Warehouses and Backups


Early
data warehousing practitioners often overlooked the need to perform
backups. Their belief was that because data for the warehouse was
extracted from operational systems, the warehouses could easily be
repopulated from those same systems if needed. However, as warehouses
grew and the transformations needed to create and refresh them
evolved, it became evident that backups of data warehouses were
necessary because the transformation process had grown extremely
complicated and time-consuming. Today, planning for warehouse
availability includes not only an understanding of how long loading
will take, but also backup and recovery operations.


/ 167