The Maximum Capacity of the Oracle Database
In the 21st century, Oracle has databases that hold anywhere from one terabyte -- equal to a billion kilobytes -- to 10 or even 100 Tb. Databases have grown because business needs have grown. Mergers and acquisitions mean one company has more data to store, and regulations in many businesses mandate extensive electronic records. The challenge to using a very large database (VLDB) isn't finding storage space, it's managing that much memory efficiently.
Other People Are Reading
-
Size Problems
-
Storing data is of limited use if you can't retrieve it efficiently. One problem with employing a VLDB is that it makes cost-effective performance difficult, as there's so much space in the memory that the computer has to search for data. Bigger systems with more storage also mean more places where a problem can develop and shut down operations. A database of 100 terabytes is physically attainable if you can afford it, but it's an impractical option if you can't manage it properly.
Partitioning
-
Partitioning makes VLDBs more manageable. A partition is a section of a data table or an index where the contents all share a common attribute such as the same datatype or column name. When the database receives an SQL query, it prunes the partitions that aren't relevant to the query from its search parameter. In a VLDB, or even an ordinary database, this reduces the search time for the query, which improves performance. If technical problems force you to shut down one partition, queries to other partitions can still operate.
-
Backing Up
-
In a crisis, a VLDB must identify the information to protect, back it up and recover it later. The size of a VLDB makes this a challenge: A 10 terabyte data warehouse may take 100 times longer to back up and require 100 times more storage than a 100 gigabyte database. Your data recovery plan has to include sufficient hardware to hold all your important data, and finding the time for the backup process. If your system has any downtime, you may be able to record it byte by byte by taking advantage of any slow moments.
Lifecycle Management
-
The data in your Oracle Database has a life cycle. Users create it, use it regularly, archive it when it gets old and finally delete it. Information Lifecycle Management stores data based on where it is in the life cycle: Current data has to be stored where it's easily available and cost-effective to recover, while rarely-used data can be stored in less accessible memory space. This approach helps manage VLDBs by determining which information has to be accessible. You then place that information in a partition where SQL queries find it easily.
-
