Chapter 9. Understanding the Limitations of Oracle Products

This chapter reviews most of the Oracle Database features, options, and add-ons and discusses how well they work with unstructured data. Each section also covers potential directions the Oracle Database could move in to better work with this type of data.

This chapter isn't designed to focus on what Oracle can't do but rather shows the potential for the product to grow considering the sound and sturdy architecture it's based on.

What has become apparent when working with Oracle is that the database has a solid foundation which helps when working with unstructured data. From the use of SecureFiles, its support for object types, its locking mechanism, security, memory management, optimizer, and tight integration with capabilities such as XML and Spatial, the database offers a very secure base. From this secure base we can build and expand, to support and scale large volumes of different types of unstructured data.

The basic requirements

A database management system that is centred around storing and managing unstructured data with multimedia should satisfy the requirements given later.

Acting as more than a filesystem

The traditional filesystem has severe limitations associated with it. Most filesystems including Windows NTFS and Unix UFS do not scale. They have limitations on file size's length, filename's length, number of files per directory as well as performance issues searching against them, performing maintenance, and navigating through them. Security varies is complex to manage and in some cases hard to audit and track when changes are made. They do not efficiently manage versioning, natively index the data, and enable multi-dimensional views of the data.

Most operating systems are woefully inadequate for the storage of unstructured data, and yet they are still needed, because a large number of applications (such as Adobe Photoshop) are centered around the concept of reading and writing to a...

A comparison

As technology is always changing, especially software, trying to accurately compare two different database vendors is fraught with dangers. Each vendor is releasing new versions of their product. In a number of cases, a direct comparison of any two features might be an invalid comparison because of the core differences in the architecture between them.

The following scale highlights a personal observation based on the capabilities for managing unstructured data within the database. Though Oracle is a long way ahead of other vendors, there is still a lot it can do to enhance and improve the database to support unstructured data. Refer to the following diagram:

Oracle products

Oracle has a large variety of products which integrate with the database, and allow it to be managed, as well as, the general features within it. This section breaks down those products and features into groups and then looks at each group to see the strengths, weaknesses, and ideal enhancements the product needs to work best with unstructured data and multimedia.

The comments against each product and tool are my own personal comments based on usage and observation. They should in no way be used to evaluate the strengths and weaknesses of the database, as by the time of publication, some of these tools are likely be updated and offer new capabilities. The observations are accurate as of Oracle 11gR2. Though enhancement ideas are suggested, they are by no means a full set of potential features that can be added and just represent the tip of the iceberg.

Development

With the release of Oracle8i came the support for objects, which included abstract data types, repeating field structures...

Summary

The Oracle Database has a lot of features and capabilities that make it an ideal platform for managing all the data for a business. The features available for managing unstructured data are miles ahead of the other database vendors in the market.

To fully scale and support the growing and changing requirements of unstructured data, the Oracle Database needs to be enhanced to address these needs. This includes rearchitecture of the SGA, the optimizer and the data dictionary, as well as building new storage structures into tablespaces to handle read-only LOBs and LOBs that are updated and require versioning. The tuning focus of the database needs to factor in the network. The RMAN and data pump backup tool needs to be enhanced to address the issues of long-term archival, and all Oracle products and tools need to fully support object types unconditionally. The database needs to adapt to handle the complexities of managing large numbers of digital objects that can individually grow into...