Ibm Db2 Vs Oracle

For this week's paper I have chosen to do a comparison between IBM's DB2 database management system (DBMS) product and Oracle. Realistically, making such a comparison is akin to discussing religion with friends. Each person will have their own biases based on their own experiences. Looking for an unbiased opinion is much like looking for the Holy Grail, every bit as speculative and subjective with the results jaundiced by the outlook of the reporter (Lim, 2002).

A decision between DB2 and Oracle should be based on what resources are available for use at the site. Is there specific hardware or software already onsite that can or will be used in the installation, if so this will have a major impact on the decision making process. Does the staff have an existing knowledge base that favors one or the other DBMS and is there a team that has unallocated time to dedicate to this project or will another project be put on hold to allow them to make this project their focus until completion. Finally, have the costs of ongoing maintenance for fixed and soft assets been considered? All of the above will impact the decision making processes prior to purchasing one of the DBMS's. These factors aside, either DBMS will run as fast as the other provided that the environment has been optimized for it (Lim, 2002).

A side by side analysis of the hardware and software requirements for both DBMS's shows some remarkable similarities as well as some differences. Similarities in baseline processors across the various hardware platforms but different amounts of disk space and both applications desiring as much memory as possible to improve performance (Chigrik, 2003).

For DB2 v8.1 when installed into a windows environment you will need a minimum of a Pentium or Pentium compatible processor, at least 256mb of ram, and between 100 and 350mb of hard disk space depending on whether the installation is compact, custom, or typical. Depending of the file format of the disk drives additional space may be needed by DB2 because of cluster sizes (Chigrik, 2003).

Oracle 9i on the other hand requires a minimum of 128b of ram in a windows environment, although 256mb is recommended. Oracle also requires a minimum of 200mb of virtual memory for file swap space. In terms of processors, Oracle specifies a minimum of a Pentium 166 or higher. Oracle also needs 140mb of disk space on the system drive and an additional 2.8 to 4.5 GB of disk space for the Oracle home drive, depending on the format of the drive it is installed on (Chigrik, 2003).

Oracle 9i comes in three versions enterprise, standard, and personal. Oracle also requires the following patch levels on Windows operating systems (OS) for installation; Windows NT, service pack (sp) 5; Windows 2000, sp 1; Windows XP, none needed. Oracle 9i is also available for HP AIX, Compaq Tru64 Unix, Linux, and Sun Solaris (Chigrik, 2003).

The DB2 Universal Database v8.1 comes in six flavors, DB2 Enterprise Server Edition, DB2 Workgroup Server Edition, DB2 Workgroup Server Unlimited Edition, DB2 Personal Edition, DB2 Universal Developers Edition, and DB2 Personal Developers Edition. The following patch levels are required based on the Windows OS it is installed on. Windows NT, sp 6a; Windows 2000, sp 2 is required for terminal services; Windows XP, none. As with Oracle, DB2 is also available for the following UNIX based OS's, AIX, Linux, Sun Solaris, and adds support for HP UX (Chigrik, 2003).

Performing a flat out performance comparison between these two RDBMS is difficult mostly because both can be used to build a stable and efficient system. Most of the difference in performance can usually be traced back to the developer and the administrators for the respective systems. But, that said, a direct comparison is possible once a set of typical transactions have been defined, that can then be run under the different DBMS on different hardware and software platforms (Chigrik, 2003).

In fact, this is what has been done by the Transaction Processing Performance Council (TPC). The TPC specifies the kinds of typical transactions that would be run in inventory control systems, airline reservation systems, and banking systems, as well some general rules these transaction should satisfy. Then the TPC can produce benchmarks to measure transaction processing and database performance. Typical performance measures are transactions per second and transactions per minute on a given