For this week’s paper I have chosen to do a comparison between IBM’s DB 2 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 DB 2 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 DB 2 v 8.
1 when installed into a windows environment you will need a minimum of a Pentium or Pentium compatible processor, at least 256 mb of ram, and between 100 and 350 mb 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 DB 2 because of cluster sizes (Chigrik, 2003). Oracle 9 i on the other hand requires a minimum of 128 b of ram in a windows environment, although 256 mb is recommended. Oracle also requires a minimum of 200 mb of virtual memory for file swap space.
In terms of processors, Oracle specifies a minimum of a Pentium 166 or higher. Oracle also needs 140 mb 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 9 i 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 9 i is also available for HP AIX, Compaq Tru 64 Unix, Linux, and Sun Solaris (Chigrik, 2003). The DB 2 Universal Database v 8. 1 comes in six flavors, DB 2 Enterprise Server Edition, DB 2 Workgroup Server Edition, DB 2 Workgroup Server Unlimited Edition, DB 2 Personal Edition, DB 2 Universal Developers Edition, and DB 2 Personal Developers Edition. The following patch levels are required based on the Windows OS it is installed on.
Windows NT, sp 6 a; Windows 2000, sp 2 is required for terminal services; Windows XP, none. As with Oracle, DB 2 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 system and database (Chigrik, 2003). For Online Transaction Processing (OLT P), the TPC-C benchmark is the most current and is the standard that DBMS’s are measured against.
When Oracle 9 i and IBM DB 2 v 8. 1 were ran head to head processing the same transactions, Oracle held all of the top 5 spots in the TPC-C. Unless an organization runs a very large database, the key areas that Oracle won the TPC-C benchmarks on won’t matter to the same degree as they would in for a large database. Newer versions of both RDBMS have made the field more competitive with the top five positions split between them, Oracle holding only three of five currently (TPC, 2005). With a statistical dead heat in performance except in very large databases, pricing becomes an issue for both DBMS. Comparing the standard edition of Oracle 9 i to the work group edition of DB 2 v 8.
1, Oracle is double the cost all the way from one processor up to thirty two. When comparing Oracle 9 i Enterprise Edition to DB 2 Enterprise Edition v 8. 1; pricing for Oracle is just slightly less than double that of DB 2 until you go above four processors then difference shifts to approximately one third more for the Oracle product. As with anything, discount can be negotiated and this will affect the overall cost to the enterprise (Chigrik, 2003). In terms of features, both DBMSs support ANSI-SQL 92 entry level and do not support it on the intermediate level.
The dialect used by Oracle is called PL-SQL and the dialect used by DB 2 is called DB 2 SQL dialect. The main difference between the dialects is that Oracle supports function based and domain indexes, while DB 2 supports block indexes and dimension block indexes. IBM DB 2 has higher limits in most of the common database features except most columns in an index key, longest index key, max char () size, max table row length, longest SQL statement, and recursive sub queries (Chigrik, 2003). In the end there is no clear cut winner, as stated previously, a lot of the outcome is determined by the experience of the developer of the database and the administrator of the database. Both DBMSs can be used to build stable and efficient transaction processing systems, with the level of experience of the personnel working with the system having a greater impact on performance, than the vendor supplying the DBMS (Chigrik, 2003). ReferencesChigrik, A.
(2003). Oracle 9 i Database vs DB 2 v 8. 1. Retrieved May 11, 2005, from MSSQL City Web site: web vs db 2. htm#part 3. Lim, C.
(2002). Oracle vs DB 2 vs Teradata. Retrieved May. 11, 2005, from Experts. About. com Web site: web Processing Performance Council, (2005).
Retrieved May. 11, 2005, from Top Ten TPC-C by Performance Version 5 Results Web site: web per results. asp.