Oracle Db11gr2 vs Sybase Ase Cost 1587043

7/23/2019 Oracle Db11gr2 vs Sybase Ase Cost 1587043

http://slidepdf.com/reader/full/oracle-db11gr2-vs-sybase-ase-cost-1587043 1/38

625 Avenue of the Americas, 2n FloorNew York, NY 10011www.orcinternational.com

Database Manageability and ProductivityCost Comparison Study

Oracle Database 11g Release 2

vs.SAP Sybase ASE 15.7

Apr i l 2012



2

Table of Contents

Executive Summary ................................................................................................................... 4

About this Report ...................................................................................................................................... 5 Report Audience ....................................................................................................................................... 6

Testing Methods ....................................................................................................................................... 6

Comparison Study ..................................................................................................................... 8

Workload Weighting .................................................................................................................................. 8

Tasks ........................................................................................................................................................ 8

Comparison Testing .................................................................................................................................. 9

Administration Workload ......................................................................................................................... 10

Backup & Recovery Tasks ...................................................................................................................... 14

Performance & Tuning Tasks ................................................................................................................. 16

SAP Considerations ................................................................................................................................ 18

Conclusion .............................................................................................................................................. 18

Summary Compilation ...............................................................................................................19

Comparison Summary Results ............................................................................................................... 19

Final Results ........................................................................................................................................... 20

Detailed Comparison.................................................................................................................21

Initial Setup & Installation ....................................................................................................................... 21

Daily Administration Tasks ..................................................................................................................... 24

Backup & Recovery ................................................................................................................................ 30

Performance and Tuning Tasks .............................................................................................................. 33

Appendix A: Test Platform Details .............................................................................................35

Hardware ................................................................................................................................................ 35

Operating System ................................................................................................................................... 35

Oracle ..................................................................................................................................................... 35

Sybase .................................................................................................................................................... 35

Appendix B: New Features .......................................................................................................36

Oracle Database 11g Release 2............................................................................................................. 36

Sybase ASE 15.7 .................................................................................................................................... 36



3

Appendix C: Detailed Task Descriptions ...................................................................................37

Testing: Administration Workload ........................................................................................................... 37

Testing: Backup & Recovery Tasks ........................................................................................................ 38

Testing: Performance & Tuning Tasks ................................................................................................... 38

Index of Tables and Equations

Tables

Table 1 - Workload Weighting ....................................................................................................................... 8

Table 2 - Testing: Administration Workload Tasks ..................................................................................... 10

Table 3 - Testing: Daily Administration Tasks............................................................................................. 12

Table 4 - Testing: Backup & Recovery Tasks ............................................................................................. 14

Table 5 - Testing: Performance & Tuning Tasks ........................................................................................ 16

Table 6 - Summary of Testing Results ........................................................................................................ 19

Table 7 - Summary of Weighted DBA Time and Step Savings .................................................................. 20



4

Executive Summary

In today’s world, corporations are generating massive amounts of data, requiring advanced techniques toprotect, store and analyze this data. Database management systems have thus become an integral partof the day to day operation of a business. The purpose of this database manageability and productivitycost comparison study is to provide IT decision makers with factual information to address the rise incosts associated with managing a database platform on a day-to-day basis.

ORC International compared the database management and productivity costs of two enterprisedatabase management systems: Oracle Database 11g Release 2 Enterprise Edition and SAP Sybase Adaptive Server Enterprise 15.7 (herein referred to as Sybase ASE 15.7).

At ORC, we recognize the need to find affordable ways to manage database platforms and ways toreduce costs within an organization. By applying our knowledge of database systems, and themanageability required in order to adhere to the daily requirements of always-on systems, we have seenit as a responsibility to report on our findings. This detailed process allowed our analysts to calculaterelative management and productivity savings that a typical organization might expect to gain or losewhen implementing the compared products. In the comparison of the two systems, we found the Oracledatabase administrator (DBA) to be much more productive than the Sybase DBA.

Initial hardware and licensing costs may seem to be the primary recipient of database departmentalspend, however, the actual cost of an enterprise database management system is in the dailymanagement of these systems by the individuals assigned to perform the tasks outlined in this study. Thecost savings we outline in this study are realized in increased productivity which results in eventual dollarssaved.

Based on our weighted averaging of everyday DBA tasks, the study revealed significant time andresource savings when using Oracle Database 11g Release 2 Enterprise Edition over Sybase ASE 15.7.The study concluded:

Oracle Database 11g Release 2 Enterprise Edition resulted in an overall 71 percent timesavings when compared to Sybase ASE 15.7.

Oracle Database 11g Release 2 Enterprise Edition resulted in an overall 43 percentstep/complexity savings when compared to Sybase ASE 15.7.

Based on a productivity savings of 71 percent, the time savings amounted to $85,200 (US)dollars per year per DBA for an Oracle Database 11g Release 2 DBA versus a Sybase 15.7 ASE DBA.

Given the above outlined benefits of Oracle Database 11g Release 2 Enterprise Edition over Sybase ASE15.7 by such an overwhelming margin, it is a safe conclusion that the overall cost of SAP Business Suiteapplications from a database manageability standpoint would be greatly increased if Sybase ASE is used

instead of Oracle Database 11g Release 2 Enterprise Edition. While both products provide sufficient tools for DBAs to complete frequent and infrequent tasks efficiently,this report focuses on the differences between them for a common set of DBA tasks.



5

About this ReportThis in-depth comparative productivity cost study compares two products, employing pre-defined tasksand steps DBAs perform daily or regularly. DBAs were interviewed about the amount of time they typicallyspent in each of the four areas of study to determine the weighted workload value for each subject area,and multiple database disciplines were chosen, including 24x7x365 production DBAs as well asapplication and development DBAs.

To develop productivity cost comparisons and infer a dollar cost comparison, common DBA tasks werethe focus of this study. Some factors were excluded from this study which could directly affect the overallcost of ownership of either platform. Not included in this study were:

Physical Server and Hardware Costs: The costs associated with servicing the DBMS platformson the physical hardware. It is assumed however that physical server requirements are similar forboth platforms evaluated in this study.

Licensing Costs: The costs associated with servicing the DBMS platforms from a softwareperspective. These include operating system costs as well as the DBMS costs, or any other

software required to operate these platforms in a production environment.ORC International also looked at the overall usability of the tools that came pre-installed and availableout-of-the-box with each database, including the number of tools required to complete the tasks. Thesewere the only ones used, since external tools require a substantial amount of time to be spent forinstallation and configuration before use. Therefore, no third-party tools were used. In a few instances atool was not available for use and the task was completed using an SQL command, rather than a GUI-based tool.

Both Oracle Database 11g Release 2 Enterprise Edition and Sybase ASE 15.7 can be run on manyoperating systems, including Linux, UNIX, or Windows1. In order to manage a Sybase ASE 15.7 instanceor database, additional client applications must be installed on the DBAs machine. Oracle Database 11gRelease 2 Enterprise Edition Database Control (the management console) is browser-based, whichmeans it can be run from any location and on any computer with connectivity to Oracle Database 11 g

Release 2 Enterprise Edition server.

Sybase ASE 15.7 has several tools that make up the management and feature set of the Sybasedatabase. While many of the basic functions of each application can be accessed via the Sybase Centralmodule, to gain full power of each application, they must be launched separately. Oracle Database 11g Release 2 Enterprise Edition Diagnostics and Tuning Packs, though separately licensed, come pre-installed with the database kernel and therefore do not require any additional installation time from theDBA.

1 Oracle Database 11g R2 Supported Operating Systems: http://www.oracle.com/us/products/database/enterprise-

edition/index.html

http://www.oracle.com/us/products/database/enterprise-edition/index.html








6

Oracle Database 11g Release 2 Enterprise Edition

Oracle Enterprise Manager Database Control, including the Diagnostics and Tuning Packs

Sybase ASE 15.7 Sybase Central

Sybase Interactive SQL

Report AudienceThis report is intended for a variety of DBMS users:

IT Decision Makers who are responsible for managing IT centers including DBMS environments.

Database administrators who work on the DBMS daily.

System architects who have a stake in the success or failure of the DBMS.

Corporate product champions who are responsible for outlining the features of a corporateapplication, or system that may rely on DBMS features.

Corporate decision makers responsible for making financial and strategic choices for thecorporation.

Testing MethodsSeveral DBAs in multiple disciplines and environments were interviewed to determine the weighting of thesubject areas being tested. The tasks were chosen by a team of database architects and identified as themost common areas of time consumption for the DBA. The categories and tasks were defined as:

Initial Setup and Installation

o Out-of-the-box installation

o Setup of proactive monitoring

Daily Administration Tasks

o Tablespace/Data file creation

o User/schema creation

o Table creation

o Index creation

o Additional user creation

o Increasing Tablespace/Data file space

o Managing database fragmentation

o Loading data from external sources



7

o Configuring and processing workloads

Backup and Recovery Tasks

o Configuring and running full backups

o Recovery of a dropped table

o Recovery of a corrupt data file

o Recovery of an erroneous transaction

Performance and Tuning Tasks

o Problem diagnosis

o Instance memory tuning

o SQL statement tuning

Tasks were broken into steps, each of which required a user action to continue the workflow. Timingmethods were applied to track the time and the number of steps required to complete each task.Whenever available, defaults were selected in all test cases, and at the first sign of a “Finish ,” “OK” or“Submit” button, that button was selected.

The tasks were tracked individually for time and step count, so results are available at the detail andaggregate level. A workload weighting constant was applied to each task. The weighting constants weredetermined during interviews with DBAs of different disciplines.

These results were compared to the total available work hours in a year (40 hours X 52 weeks = 2080annual work hours), to calculate productivity results.

Dollar cost comparisons were also done to obtain task cost per occurrence; however, the DBA salary peryear is a constant in this study (the same for both platforms). Dollar costs can be compared at the

individual task level, but the productivity savings value is what will lead to overall dollar savings by anorganization.



8

Comparison Study

Workload WeightingUsing workload weighting, constants were applied to the tasks. These constants were based oninterviews with DBAs, and take into account the percentage of time a DBA might spend on a specific taskarea. Reports from various ticketing systems were analyzed for category assignment to also look at thefrequency of various requests to help balance the responses from the interviews. The workloadweightings are shown in Table 1.

Table 1 - Workload Weighting

Task Area Workload Weighting

A. Initial Setup and Installation 5%

B. Daily Administration Tasks 25%

C. Back and Recovery Tasks 10%

D. Performance and Tuning Tasks 45%

E. Administrative Functions *

(Not Measured or Evaluated) 15%

*15 percent of time was allotted for administrative/HR functions and notcounted towards evaluation of the actual DBA’s time and energy. All percentages of time mentioned in the remainder of the document reflectthe percentage of time that has been accounted for, 85 percent.

TasksEach task is a request or action a DBA performs based on a user request. To validate the chosen tasklist, it was compared to various ticketing system categories in production and application support queues.

The categories and issues reported were then looked at for frequency in order to ensure the tasks fit intothe most requested categories, and accounted for 85 percent of DBA work, minus administrativefunctions and meetings.

Time

Time was measured from the start of a given task, when the mouse began to move to complete the task.For tasks that could run in the background, time was measured when the job or process was submitted tothe DBMS. The job was tracked as a subtask and the run time was recorded independent of the task thatrequired the DBA’s attention.

For study consistency, any task that prevented a DBA from completing another task was measured forthe time required to complete the task.

ComplexityComplexity can be defined in many ways and each task has its own level of complexity. For this study, aformula was defined that weighted the step count to simulate the amount of DBA attention required tocomplete the task.



9

A step was defined as requiring a DBA to interact with the application to perform or finish a task,irrespective of clicking “Next,” “Finish,” or “Submit.” A step could include the typing of object names, orconfiguration of parameters such as passwords or instance names. These intuitive tasks received a step

count of one.

Depending on a step’s complexity, a step count penalty was applied. This included the inability to performthe task using a GUI-based tool. This penalty was the total number of steps required to perform the task,including clicking of menus, additional commands, and writing of scripts. If a script was required it wasviewed as a single step, since most DBAs maintain frequently used scripts in their own toolboxes.

Comparison TestingThe purpose of doing comparison testing between these database management systems is to determinethe long-term productivity and dollar savings or costs between the two systems. At first glance, hardwareand licensing costs can appear high. However, the real cost of a database management system is acombination of the daily management of the system and the people costs in time and dollars associatedwith performing the outlined tasks. The annual dollar cost of a DBA is constant for a given calendar year.

Therefore, the cost savings to consider is improved productivity — allowing the DBA to do more in acalendar year.

To compare dollar cost savings, rather than the dollar costs of each task, a $120,000.00 annual US salarywas taken from the report of a salary study of database administrators, architects, and developers whomainly focus on database development. The number was chosen from an available online Salary Study,and the number was chosen from the highest salaries reported in the study2.

Using the salary from the study and our own weighted productivity savings, we calculated the dollar costsavings of Oracle Database 11g R2 over Sybase ASE 15.7 as:

Salary ($120,000.00) * DBA Time Savings (71%) = $85,200.00

This number represents the cost savings per DBA, per year, within a corporate organization. In reality,the cost of a DBA’s yearly salary does not change for a corporation; the productivity of the DBA does. Thetest results below highlight how much more effective a DBA is when operating Oracle Database 11g R2compared to Sybase ASE 15.7.

This section provides the detailed comparative findings on the two platforms: Oracle Database 11 g R2and Sybase ASE 15.7. For the remainder of the document, Oracle Database 11 g R2 will be referred to asOracle Database 11g R2 and Sybase ASE 15.7 will be referred to as Sybase ASE 15.7.

2 The salary study is available from:http://salary-surveys.erieri.com/content/Architecture_Salary_Survey/Database_Architect_1856/US_National.htm

http://salary-surveys.erieri.com/content/Architecture_Salary_Survey/Database_Architect_1856/US_National.htm





10

Administration WorkloadThe general Administration Workload Testing takes into account tasks performed infrequently by theDBA. In our weighted workload value, these infrequent tasks account for only 5 percent of the DBA’s workhours, or 104 work hours per year, based on a 40-hour work week. The tasks included in this part of thetest are shown in the table below.

Table 2 - Testing: Administration Workload Tasks

A. Initial Setup and InstallationTime (Min:Sec) Steps Percent Difference

Oracle Sybase Oracle Sybase Time Steps

1 Out-of-the-box installation 15:59 32:59 1 3 52% 67%

2 Setup proactive monitoring 00:19 05:15 4 11 94% 64%

Total 16:18 38:14 5 14 57% 64%

Out-of-the-Box Default Installation

Installation of both systems was simple and straightforward. The system documentation for both DBMSwas complete and thorough. Both systems install all pre-requisites without issue or error. Oracle won oninstallation time by 52 percent. Oracle Database 11g R2 was less complicated to install and requiredfewer steps than Sybase ASE 15.7. As the results show (Table 2 - Testing: Administration WorkloadTasks), Oracle Database 11g R2 took only 1 step to complete, Sybase ASE 15.7 took 3 steps tocomplete.

No additional steps were required for either DBMS platform in order to put it to task. Both OracleDatabase 11g R2 and Sybase ASE 15.7 were ready for duty as soon as the installation was completed.

Both installations were easy. Oracle Database 11g R2 installation required little or no confirmation exceptfor a few text boxes (e.g., passwords, instance names) which required input. Sybase ASE 15.7 alsorequired only steps that were intuitive, similar to the Oracle Database 11g R2 installation (e.g.,

passwords, database name, instance name, etc.).

Both systems had few required parameters and adequate defaults to complete the installation.

Proactive Monitoring Setup & ConfigurationProactive monitoring setup and configuration were much faster and easier for Oracle Database 11g R2than Sybase ASE 15.7. For both products, four nearly-identical thresholds were modified with thefollowing results:

Oracle Database 11g R2 required 64 percent fewer steps for setup than Sybase ASE 15.7.

Oracle setup time was also 94 percent less than Sybase ASE 15.7.

Comparing the two platforms was difficult, as they approach threshold monitoring and configuration quitedifferently. As the numbers above reflect, using the provided tools for Oracle Database 11g R2 requiredless setup time and steps than using the provided tools with Sybase ASE 15.7.



11

For Oracle Database 11g R2, all monitoring thresholds were configured from a single screen, and allchanges enabled in a single submit/commit. In the case of Sybase ASE 15.7, the Monitor from SybaseCentral was employed. This was however a very manual process for the analyst, and a penalty wasapplied in both time and step count.

The task of setting up and configuring proactive monitoring is not performed regularly. This typicallyhappens on initial database creation, but sometimes must be changed until the DBA has identified thebiggest areas of contention within his or her database. In interviews, DBAs said of all the administrativetasks, this one must be performed most often. The type of database (e.g., production, implementation,quality assurance, development) plays a role in configuring this task due to monitoring requirements andSLAs being different based on the type of environment the database is in. There are separate tools thatcan be purchased and installed by a Sybase ASE 15.7 DBA, however this can lead to unforeseen costsby the organization and will lead to lost time from a productivity perspective.

One benefit of Oracle Database 11g R2 over Sybase ASE 15.7 is the installed Oracle Diagnostics Packand Oracle Tuning Pack. More specifically, the Automatic Database Diagnostic Monitor (ADDM) isintegrated into Oracle Database 11g R2 kernel. This self-diagnostic tool takes much of the guesswork outof performance-tuning Oracle Database 11g R2 instance. Performance tuning often comes from theory,and DBAs are likely to look for performance issues in areas they have encountered in the past, thereforemaking decisions on performance metrics that may be unfounded or inaccurate. Sometimes these metricconfigurations yield false positive alerts to the DBA and production support teams, thereby resulting in thealerts being ignored or turned off for lack of response.

With the ADDM performance problems are categorized into trees; root cause is located based on thecombined years of performance-tuning experience by Oracle’s trained architects. Key performanceindicators (KPI) are included for each ADDM-reported issue, so each issue can be prioritized andaddressed according to its impact on the database instance. The ADDM recommendations are based ona repository called Automatic Workload Repository (AWR) that enables Oracle Database 11g R2 toaccurately monitor and diagnose instance problems. The AWR contains statistics on how the instance isused operationally. The purpose of the AWR is to take periodic snapshots of the current databaseinstance load for processing by the ADDM; it forms the foundation for how the ADDM processes and

parses the data. The AWR allows the ADDM to identify and present performance issues and enableshistorical instance analysis to allow the DBA to isolate the instance in question and show performanceimprovement over time.



12

Daily Administrative Tasks

Daily administrative tasks account for 25 percent (520 work hours) of the DBA’s workload. The dailyadministration tasks identified by our DBA panel are listed in the table below:

Table 3 - Testing: Daily Administration Tasks

B. Daily Administration TasksTime (Min:Sec) Steps Percent Difference


3 Create Tablespace 00:13 00:25 2 1 48% -100%

4 Create User/Schema 00:06 00:10 2 1 40% -100%

5 Create Table 00:45 00:54 17 17 17% 0%

6 Create Index 00:43 00:13 3 3 -231% 0%

7 Create Additional User 00:09 00:17 2 4 47% 50%

8 Add Space to Tablespace 00:08 00:10 1 2 20% 50%

9 Reclaim Fragmented Space 00:33 05:124 6 89% 33%

10 Load Data from Text File (csv) 01:20 01:22 13 1 2% -1200%

11 Configure Adaptive Thresholds and

Perform Workloads00:46 06:14 0 7 88% 100%

Total 4:43 14:57 44 42 68% -5%

Object Creation: Tablespace/Data FileTablespace creation occurred 48 percent faster for Oracle Database 11g R2 when compared to Sybase ASE 15.7. The number of steps required for this task using Oracle Database 11g R2 required 1 more stepthan Sybase ASE 15.7 to perform the test, and because of the decreased amount of steps required forboth systems to be tested, this resulted in some artificially skewed numbers.

While both platforms performed the task quickly and without error, it is shown in the number that Sybase ASE 15.7 took significantly longer to perform from a time perspective.

Object Creation: User/SchemaCreating a new user/schema was easy, quick, and straightforward for both platforms. Sybase ASE 15.7performed this task 40 percent slower than Oracle Database 11g R2; however, Oracle took an additionalstep in order to complete. Both DBMS platforms created the additional user without issue or error.

Object Creation: TableTable creation was also very similar for both platforms, with identical tables being created. Because thedefault tablespace was a system-based tablespace when using Oracle Database 11g R2, an additionalstep was added to Oracle Database 11g R2 to create the user-based objects in a non-system basedtablespace. Oracle Database 11g R2 completed the task 9 seconds faster than Sybase ASE 15.7 usingthe same number of steps.



13

Object Creation: IndexOracle Database 11g R2 fell behind Sybase ASE 15.7 in speed and matched Sybase ASE 15.7 incomplexity during index creation. Default values were selected for both platforms. However, OracleDatabase 11g R2 performed this task slower than Sybase ASE 15.7 by 30 seconds. The index consistedof a single column in both tests.

Object Creation: Additional User An additional differentiating factor between the two programs was the creation of an additional user. It isORC’s opinion that both Oracle Database 11g R2 and Sybase ASE 15.7 each provide an intuitiveinterface for user/schema creation. Oracle Database 11g R2 took 47 percent less time to complete thistask than Sybase ASE 15.7, and also required 50 percent less steps. While this is a significantproductivity difference, most DBAs use prepared scripts for creating additional or bulk users, whichreduces screen navigation time. The user creation task is tricky to analyze fully, since users differ widelyand are created for different purposes. In our interviews, we found standard users are created only once.

Temporary users, on the other hand, are created often, so as to allow architects or developers to

diagnose production and quality assurance systems. Temporary users are often created and thendropped, or passwords changed once access is no longer needed. This human analysis speaks to thespeed and efficiency levels the Oracle DBA provides when performing this task.

Space Management

Add Space to TablespaceBoth DBMS platforms provide an intuitive user interface to grow the tablespace/data files, rather than addan additional file to the tablespace/data file. For this study, the amount added was 25 percent of the initialspace. Oracle Database 11g R2 out performed Sybase ASE 15.7 by 20 percent and Sybase ASE 15.7required an additional step, resulting in a 100 percent gain in step count.

Reclaim Fragmented SpaceOracle Database 11g R2 offers the ability to reclaim fragmented space by reorganizing the tablespacewith a very intuitive user interface. In order to accomplish the same task in Sybase ASE 15.7, a DBA mustmanually run a REORG script, or restore the database from a backup during a maintenance window. Theprocess was straightforward for Oracle Database 11g R2. In regards to Sybase ASE 15.7, individualtables can be defragmented using the same REORG command and tables can be isolated as requiringreorganization by using the MON tables. Oracle Database 11g R2 completed this task requiring 33percent fewer steps and 89 percent less time when compared to Sybase ASE 15.7.

Load Data from Text FileFor this task, ORC International’s testing representatives chose a small country-to-IP mapping table andidentified a 169,000-record dataset for loading. This table is commonly truncated, loaded and unloaded

into production systems. For this exercise, the technicians chose to create the table upon the first load,adding considerable steps to Oracle Database 11g R2’s loading of the file.

Oracle Database 11g R2 and Sybase ASE 15.7 require almost identical time in order to complete thistask. However, Oracle Database 11g R2 required the creation of the table, rather than the wizardintuitively knowing how the table should be created.



14

While the time to import the data was nearly identical, it was much more complex for Oracle Database11g R2 to import over Sybase ASE 15.7 due to the table creation process for Oracle Database 11 g R2.

Configure Adaptive Thresholds and Perform WorkloadsThis is an important task for any database administrator. In order to analyze and know what is happeninginside of the database instance on a day to day basis, it is critical to properly analyze how the databaseresponds to various application loads. It also enables the database administrators to communicate toother departments how the users behave in the system, and how they use a given database. The need toquickly access and consume this data is critical to the everyday interface for the database administrator.

Oracle Database 11g R2 was superior in this category and testing scenario. It accomplishedconfiguration, consumption and processing of the workload in 46 seconds, 88 percent less time than thecomparable test run on Sybase ASE 15.7. Oracle Database 11g R2 also accomplished this task with zerosteps, only increments. No further input was required of the database administrator.

It is the opinion of ORC this is a very laborious process for Sybase ASE 15.7. Specific trace flags must be

set on the Sybase ASE 15.7 server, and then MON tables must be manually looked at and queried,specifically the monProcessSQLText table. This leads to continuous time investment of a staff DBA for anorganization.

Backup & Recovery TasksBackup and recovery are vitally important to a DBMS platform and not surprisingly, both systemsperformed these tasks. Oracle Database 11g R2 showed significant productivity gains over the Sybase ASE 15.7. The following table outlines the tasks performed:

Table 4 - Testing: Backup & Recovery Tasks

C. Backup and Recovery TasksTime (Min:Sec) Steps Percent Difference


12 Configure and Run Full Backup 00:08 00:30 0 1 73% 100%

13 Recover Dropped Table 00:28 02:32 3 14 82% 79%

14 Recover Corrupt Datafile 01:32 02:00 5 10 23% 50%

15 Recover from erroneous transaction

using Flashback01:18 10:00 8 10 87% 20%

Total 3:26 15:02 16 35 77% 54%



15

Configure & Run Full Backup(s)Configuring and running full backups were very similar for both Oracle Database 11g R2 and Sybase ASE 15.7. Both have intuitive interfaces for accomplishing this task outside standard maintenance. Thistask is often required in the middle of the day or during a production time, when a release occurs, a

problem needs diagnosis, or QA requires a refresh of the most current data.

Oracle Database 11g R2 performed the backup process 73 percent faster than Sybase ASE 15.7 andrequired zero steps on the technician’s part other than selecting “Next/Submit.” Both platforms executedthe task quickly and without issue, however it took Sybase ASE 15.7 over 100 percent longer in order toaccomplish the backup.

Recover Dropped TableRecovery of a dropped table in Oracle Database 11g R2 was very simple. In fact Oracle Database 11g R2 outperformed Sybase ASE 15.7 with an 82 percent time and 79 percent step/complexity savings.Oracle Database 11g R2 provided out of the box options for the recovery of a table. The steps are clearlydefined and easy to follow, and in less than 30 seconds, our testing had recovered a table with over160,000 rows in it.

Sybase ASE 15.7 required the testers to restore the database to a new location, extract the data from thenewly restored database, truncate the destination table, then import the freshly exported data. This wassimilar to the process required by SQL Server 2008 R2, and is no surprise since SQL Server wasoriginally based on Sybase. It was found that there were multiple points at which this could causeproblems. For a database of significant size in a production environment where the database is > 100 GB,this additional restore could add hours to the overall restore process of a single table when using Sybase ASE 15.7, not to mention the requirement for additional space in a production environment.

Had this been an actual production instance of significant size, Oracle Database 11 g R2 would have stillperformed within seconds since the option to flashback a single table is present and available. The exactamount of time to restore a database of significant size in an environment utilizing Sybase ASE 15.7 isnot known since each production environment, including the computer and disk subsystem architecture

can vary from company to company.

It is ORC International’s opinion that based on this task, this alone can add a significant amount of timesaved each year by an organization if they choose to utilize Oracle Database 11g R2 over Sybase ASE15.7.

Recover Corrupt Data File

ORC International’s team of experts was not able to restore a Sybase ASE 15.7 datafile using the wizardsand tools provided by Sybase. The only thing our experts were able to accomplish was the restoration ofan entire database. It is our opinion that using Oracle Database 11g R2 provided Oracle AdvisedRecovery process was more intuitive, and provided the available functionality. Oracle Database 11 g R2wizard hand-held the user through the entire process and made reliable recommendations to correct anyknown issue.



16

Correct Erroneous TransactionThis scenario was very complex for Sybase ASE 15.7 testers to complete. In this test not only did ORChave to repeat certain steps to restore the database to a new location, export the table, import it back intothe destination table after a truncation, but ORC also had to repeat certain transactions, since we only

had our most recent backup to work with. By using the Flashback technology again, Oracle Database 11g R2 outperformed Sybase ASE 15.7 by 87 percent in time savings and 20 percent in step/complexitysavings.

It must be noted again in the case of a large production database this can result in many hours investedby a Sybase ASE 15.7 DBA in order to perform this recovery operation. Even with a DBA being 100percent focused on the recovery, the amount of manual steps in the Sybase ASE 15.7 recovery operationcan lead to further human errors that can potentially increase the amount of time it takes to recover froman erroneous transaction. It is therefore ORC’s opinion this task can lead to a significant amount of timeand productivity savings to an organization that chooses to utilize Oracle Database 11g R2 as its DBMSplatform of choice rather than Sybase ASE 15.7.

Performance & Tuning TasksPerformance and tuning tasks are vital to the scalability and survivability of applications so productivity inthis area is important. The DBA needs to be able to quickly diagnose and solve performance and tuningproblems. ORC International’s panel of database professionals revealed that in true 24x7 environments,this task accounted for nearly half of the DBAs’ time.

From a timing perspective Oracle Database 11g R2 out-performed Sybase ASE 15.7 by a substantialmargin. The high difference is due to the variance of all tasks, where Sybase ASE 15.7 had a much morecomplex process for tuning SQL statements as well as problem diagnosis and took much longer toprocess than Oracle Database 11g R2.

Table 5 - Testing: Performance & Tuning Tasks

D. Performance and Tuning Tasks

Time (Min:Sec) Steps Percent Difference


16 Problem Diagnosis 00:08 02:00 2 5 93% 60%

17 Tune Instance Memory 00:00 00:00 0 0 0% 0%

18 Tune SQL Statements 00:00 02:00 0 5 100% 100%

Total 00:08 4:00 2 10 97% 80%

Problem DiagnosisOracle Database 11g R2 showed significant gains over Sybase ASE 15.7 in problem diagnosis, with 93percent time savings. Performance diagnosability is built inside the Oracle Database kernel and hence

performance problems are automatically detected and reported. Also, the initial Oracle EnterpriseManager’s intuitive dashboard, with its KPIs and visualizations of database performance, makesproblems easy to diagnose and locate. Sybase has some standard informational reports that are easilyaccessible when using the Sybase Central, however the lack of a consolidated dashboard for easyidentification makes the tool more cumbersome to use.



17

In ORC’s opinion, the lack of an interactive dashboard within Sybase ASE 15.7 does not provide DBAs orother IT staff with a quick means to tell where a problem may be occurring. An overview dashboard of theDBMS system would enable quicker problem diagnosis and resolution. There are third-party monitoringproducts from other vendors are available for both DBMS platforms, however one of these additional tools

would have to be purchased in order to obtain the same level of ease of use that comes from a standardinstallation of Oracle Database 11g R2. This purchase would add to the overall cost of ownership forSybase ASE 15.7 and also lead to lost productivity in installation, configuration, and training on such athird-party tool.

In our opinion, Sybase ASE 15.7 cannot be compared on any level of problem diagnosis with OracleDatabase 11g R2 without the installation and configuration of third-party or custom built tools. OracleDatabase 11g R2’s additional productivity is gained by employing the ADDM and AWR. While Sybase ASE 15.7 does have standard reports that can be looked at, there is not an intuitive way for a DBA to geta high level view of database performance without a third-party tool.

Instance Memory TuningBoth Oracle Database 11g R2 and Sybase ASE 15.7 offer out-of-the-box instance memory tuning.

Therefore, this test will be removed from subsequent studies of this kind, since nothing was available forus to test.

SQL Statement Tuning

SQL statement tuning is an often-overlooked aspect of DBA work, but one that takes a good portion oftheir time. Sometimes DBAs are hired by organizations that have poorly-designed databases, ordatabases that require constant tuning. Other times, they find themselves part of a development team orsupporting a development team where the developers outnumber the DBAs 10 or 15 to one. In thesecases, DBAs tune and review a lot of SQL. Oracle Database 11g R2 showed a total domination inproductivity and time savings when compared to Sybase ASE 15.7. Oracle Database 11g R2 used nosteps or time in order to complete this task.

Oracle Database 11g R2 provides out-of-the-box tuning of SQL statements and reporting of problematic

SQL. The DBA need not wait for alerts on blocked objects or reports from operations that web pages arenon-responsive. The DBA can easily navigate within Oracle Enterprise Manager using the SQL ResponseTime chart on the performance dashboard to quickly dive into KPIs and look at the collection details toidentify poorly written SQL if they choose to not rely solely upon the Automatic Tuning features of OracleDatabase 11g R2. For the purposes of this test, we made the assumption SQL would be automaticallytuned, that none of our production SQL contained recursive statements, or that queries would not goparallel3.

Sybase ASE 15.7 required the DBA to perform a completely manual process in terms of monitoring andperformance tuning. Sybase Interactive SQL was employed by ORC’s experts and it was determined thata deep understanding of the various MON tables, as well as the specific database structure was requiredin order to fully tune and optimize the database for performance. It was also our determination that deepunderstanding of how an application interacted with the database would also be required in order to fullyoptimize the database for a specific application usage.

Oracle Database 11g R2 outperformed Sybase ASE 15.7 by 100 percent in time savings and 100 percentin step/complexity savings.

3 Oracle Automatic SQL Tuning Features and Usage:

http://download.oracle.com/docs/cd/E11882_01/server.112/e16638/sql_tune.htm#CHDHICFI






18

SAP Considerations

SAP has acquired Sybase and is putting efforts to make Sybase the underlying database for SAPBusiness Suite applications. Sybase ASE and Oracle Database 11g R2 are two potential DBMS platformsfor SAP Business Suite applications even though IBM DB2 and Microsoft SQL Server also have somepresence in this space. While this study did not focus on SAP specifically, it is apparent by the results ofthis study that Sybase ASE has tremendous hurdles to overcome before it can be a real contender in theSAP landscape due to the significant added cost of manageability. Especially due to the Initial Setup andInstallation, Backup and Recovery, as well as Performance Tuning, which account for a combined weightof 60 percent of a typical DBA’s time. Moreover, the smaller number of skilled Sybase DBAs under SAPcompared to Oracle Database should also factor into the cost of using Sybase ASE for SAP BusinessSuite applications.

Conclusion Any organization can have significant productivity savings by choosing Oracle Database 11g Release 2Enterprise Edition over Sybase ASE 15.7. Each IT organization is different, but based on our ownindependent interviews and lab tests, these tasks, and their respective weighted averages, have beenidentified as the most common tasks that make up the typical database professional’s day.

The two key areas that combined to account for 55 percent of a DBA’s time, based on our weightedvariables, included Backup and Recovery , and Performance Tuning Tasks. These two areas had thegreatest manageability differences between the two products. Oracle Database 11g Release 2 EnterpriseEdition exceeded Sybase ASE 15.7 in these two areas by 87 percent in average time saved, and 67percent in average step/complexity savings.

Based on our overall findings, Oracle Database 11g Release 2 Enterprise Edition provides nearly a 71percent productivity savings and is 43 percent less complex compared to Sybase ASE 15.7, which

translates into a cost savings of $85,200 per Oracle DBA.



19

Summary Compilation

Comparison Summary Results

Table 6 - Summary of Testing Results

Task DescriptionTime (Min:Sec) Steps

Oracle Sybase Oracle Sybase

A. Initial Setup & Installation 16:18 38:14 5 14

1 Out-of-the-box installation 15:59 32:59 1 3

2 Setup proactive monitoring 0:19 5:15 4 11

B. Daily Administration Tasks 4:43 14:57 44 42

3 Create Tablespace 0:13 0:25 2 1

4 Create User/Schema 0:06 0:10 2 1

5 Create Table 0:45 0:54 17 17

6 Create Index 0:43 0:13 3 3

7 Create Additional User 0:09 0:17 2 4

8 Add Space to Tablespace 0:08 0:10 1 2

9 Reclaim Fragmented Space 0:33 5:12 4 6

10 Load Data from Text File (csv) 1:20 1:22 13 1

11Configure Adaptive Thresholds and

Perform Workloads0:46 6:14 0 7

C. Backup and Recovery Tasks 3:26 15:02 16 35

12 Configure and Run Full Backup 0:08 0:30 0 113 Recover Dropped Table 0:28 2:32 3 14

14 Recover Corrupt Datafile 1:32 2:00 5 10

15Recover from erroneous transaction

using Flashback1:18 10:00 8 10

D. Performance and Tuning Tasks 0:08 4:00 2 10

16 Problem Diagnosis 0:08 2:00 2 5

17 Tune Instance Memory (Appendix C) 0:00 0:00 0 0

18 Tune SQL Statements 0:00 2:00 0 5



20

Final Results

Table 7 - Summary of Weighted DBA Time and Step Savings

Tasks Area Workload

Weighting

% Difference Workday Savings

Time Steps Time Steps

A. Initial Setup & Installation 5% 57% 64% 3% 3%

B. Daily Administration Tasks 25% 68% -5% 17% -1%

C. Backup and Recovery Tasks 10% 77% 54% 8% 5%

D. Performance and Tuning Tasks45%

97% 80% 44% 36%

Total 85%* 66% 33% 71% 43%

*15 percent of time was allotted for administrative/HR functions and not counted towardsevaluation of the actual DBA’s time and energy.



21

Detailed Comparison

Initial Setup & InstallationTask 1: Default Inst al lat ion

Oracle SYBASE ASE 15.7

Process Flow Time Steps Process Flow Time Steps

Task: Default Installation

In Explorer Navigate to DVD andstart setup by double clickingsetup.exe

00:00 0Splash Screen - Setuploading

00:18 0

Email Configuration: Defaultsselected/Warnings ignored

0:02 0Sybase Splash screen -More loading

00:08 0

Install Options: Default selected 0:02 0Introduction Screen -Click Next

00:01 0

System Class / Defaults selected 0:02 0Choose Install Folder -Defaults selected(C:\Sybase) - Click Next

00:02 0

Password Screen / Set password to123456

0:05 1

Destination Popup -Destination does notexist. Do you want tocreate it? - Yes selected

00:01 0

Installation Review / Finish Selected 0:08 0

Choose Install Set -

DEFAULT NOTSELECTED - Full choseninstead of Typical

00:01 1

Actual Installation / Binary Copy, etc. 8:31 0Choose Installation Type- Default selected

00:01 0

Oracle Database Configuration Assistant ran based on previousinputs

7:09 0

EULA - Selected AllRegions for licenseagreement - Selected Iagree - Clicked Next

00:09 2

Oracle installed without any issues orerrors

Pre Installation Summary- Clicked Install

00:01 0

Installing - Binaryinstallation

27:28 0



22




Remember ASEPassword - Default ofEnabled selected, nextclicked

00:01 0

Configure New Servers - All Defaults left checked.

00:01 0

Wait screen - SelfConfiguring

00:09 0

Configure New AdaptiveServer - Defaults selected- Next clicked

00:03 0

Page Size Warning - Okselected

00:02 0

System DatabaseConfiguration Screen -Defaults selected - Nextclicked

00:01 0

Configure New BackupServer - Defaults selected- Next clicked

00:01 0

Configure New XP server- Defaults selected - Nextclicked

00:01 0

Configure New Job

Scheduler - Defaultsselected - Next clicked

00:01 0

Configure Self-Management - Defaultsselected - Next clicked

00:02 0

Wait screen 00:09 0

Configure Web Services -Defaults selected - Nextclicked

00:01 0

Unified Agent - SelfDiscovery - Defaultsselected - Next clicked

00:01 0

Unified Agent - SecurityLogin Modules - Defaultsselected - Next clicked

00:02 0

Configure new serverssummary - Next clicked

00:01 0



23




Running Tasks - Aftertime elapsed, clickedDone

04:13 0

Oracle Total 15:59 1 Sybase Total 32:59 3

Task 2: Set up proact ive mon i tor ing for p erformance and space ut i l izat ion



Task: Setup Proactive Monitoring

Enterprise Manager homepage /Select Metric and Policy Settings atthe bottom of the page

0:01 0

PENALTY - No proactivemonitoring without thirdparty tools. All manualmonitoring.

05:00 10

Metric and Policy Settings / Edit fourvarious thresholds and select OK

0:18 4

Sybase Control Center -Connection Assumed -Right click the server -Select Monitor

00:02 0

Monitor - Select SQL Activity

00:03 1

Monitor - Manually look atall values 00:10 0




24

Daily Administration Tasks

Task 3: Create Tablespace

Oracle SYBASE ASE 15.7Process Flow Time Steps Process Flow Time Steps

Task: Create Tablespace

Server Screen / TablespacesSelected

0:01 0

Sybase Central -Connection Assumed -Expand server - right clickDatabase Devices - selectNew

00:03 0

Tablespace Creation / UnderDatafiles select Add (new tablespacenamed PRODDATA)

0:05 1

Add Database Device -Enter Device Name -Leave device path - ClickNext

00:05 1

New Datafile Creation Locationdefault selected. New file namedproddata.tbs / OK selected

0:06 1 Add Database Device -Click Next

00:02 0

Tablespace Creation Review /Summary screen viewed to ensurenew space was created

0:01 0 Add Database Device -Click Next

00:02 0

Add Database Device -Click Next

00:02 0

Add Database Device -Click Finish

00:02 0

Sybase Central -

Connection Assumed -Right click database ofchoice - select Properties

00:02 0

Database Properties -Select Devices

00:01 0

Database Properties -Click Add

00:01 0

Database Properties(Specify Device and Size)- Select newly createddevice - Click Ok

00:03 0

Database Properties -Click Ok

00:02 0




25

Task 4: Create User/Schema



Task: Create User/Schema

Enterprise Manager Homepage /select Security

0:01 0

Sybase Central -Connection Assumed -Expand Database - Rightclick Users - select New ->User

00:01 0

Security Page / Click Create 0:01 0 Add User - EnterUsername - Click Next

00:03 1

Enter username “PROD_APP”password 123456 and all other fieldsleft default / Select OK

0:04 2 Add User - Click Next 00:02 0

Add User - Click Finish 00:04 0


Task 5: Create Table



Task: Create Table

Enterprise Manager Home, go toSchema / tables 0:01 0

Sybase Central -Connection Assumed -

Expand Database, rightclick User Tables, selectNew -> Table

00:02 0

Tables change default schema fromSystem to PROD_APP / Click Create

0:02 1New Table - Columnsentered, defaults given,table name provided

00:40 17

Enter Table Name / 5 columnschange various default values / datatypes/ nullability / Click OK

0:42 16New Table - Save TableClicked

00:12 0




26

Task 6: Create Index



Task: Create Index

Enterprise Manager Home / go toSchema / tables

0:01 0

Sybase Central -Connection Assumed -Expand Database, ExpandTables, Expand SpecificTable, Right click Indexes,select New -> Index

00:06 0

Select Indexes / Create 0:02 0New Index - Type in nameof new index - click Next

00:04 1

Input index name, enter table nameand select Populate Columns

0:06 2

New Index - Pick singlecolumn, and click button toput it in right column -

Click Finish

00:03 2

Select single column to include in theindex. / Select OK

0:34 1


Task 7: Create Add it ional User



Task: Create Additional User

Enterprise Manager Homepage /select Security

0:01 0

Sybase Central -Connection Assumed -Expand server, right clickLogins, select New ->Login

00:02 0

Security Page / Click Create 0:02 0

Add Login - Type inUsername, Password, andConfirm Password - ClickNext

00:08 3

Enter username “PROD_MAILER”

password 123456 and all other fieldsleft default / Select OK

0:06 2 Add Login - Click Next 00:02 0



27



Task: Create Additional User

Add Login - Select userdatabase for the login tobe associated with - ClickNext

00:03 1

Add Login - Click Finish 00:02 0


Task 8: Add Space to Tablespace



Task: Add Space to Tablespace

Enterprise Manager Homepage /Datafiles

0:01 0

Sybase Central -Connection Assumed -Expand Databases - Rightclick database selectproperties

00:02 0

Pick an existing datafile / Additionalspace specified, OK selected

0:07 1Database Properties -Select Devices - Selectspecific Device click Edit

00:03 1

Edit Device Size - Type

10, select Ok, select Ok toclose properties

00:05 1




28

Task 9: Reclaim Fragmented Space



Task: Reclaim Fragmented Space

Enterprise Manager Homepage 0:00 0

Launch Interactive SQL -Manual reorg script writtento rebuild the table anddefragment the objects

01:12 2

Schema Tab / Reorganize Objects /Tablespace / Various wizard screensto complete

0:33 4Penalty - For manualprocess and time taken towrite script

4:00 4


Task 10: Lo ad Data from Text Fi le (CSV)



Task: Load Data from Text File (CSV)

Enterprise Manager Homepage /Data Movement

0:00 0

Sybase Central -Connection Assumed -Right click table - selectData Transfer ->Incremental -> In

00:03 0

Data Movement / Automaticallygenerate control file, host credentialsentered

0:02 0

Incremental Transfer In for

{X] - Click Browse -Navigate to file selectSave

00:07 1

Data Movement / Select Datafile /Create table, have to manuallyspecify tables, even though sourcefile has already been specified

0:51 13

Incremental Transfer In for{X] - Click Finish - Wait forload data process tocomplete.

01:12 0

Data Movement / Conventional Path/ Various screens related to jobsubmission, accept defaults.

0:09 0

Job Review 0:07 0

Actual Load 0:11 0



29



Task: Load Data from Text File (CSV)


Task 11: Conf igure Adapt ive Thresholds and Perform Wo rkloads



Task: Configure Adaptive Thresholds and Compare Workloads

Enterprise Manager Homepage /Select Baseline Metrics

0:04 0

Penalty - No way ofperforming workloads.100% manual process to

do tracing.

04:00 4

Threshold Configuration / QuickConfiguration Selected

0:06 0

Interactive SQL - issuevarious DBCC statementsto turn on the correct traceflags (302, 310, 3604)

00:14 1

Profile Selected 0:05 0

Interactive SQL - Querythe monProcessSQLTexttable, to manually digestthe information (Timepenalty)

02:00 2

Review Profile / Finish Selected 0:04 0

Wait Time for Consuming 0:27 0




30

Backup & Recovery

Task 12: Conf igure and Run Ful l Backup


Task: Configuring and Run Full Backup

Backup Configuration / Schedule aFull Backup

0:01 0

Sybase Central -Connection Assumed -Right click database -select Backup

00:01 0

Schedule Oracle Suggested Backup 0:01 0Backup Database - Typename of backup

00:06 1

Disks left as default 0:02 0Backup Database - ClickFinish - Wait for process tocomplete.

00:23 0

Default time of 2:00 AM selected 0:04 0

Job submitted 0:00 0


Task 13: Recov er Dropped Table



Task: Recover dropped table

Enterprise Manger Homepage / Availability

0:02 0

Sybase Central -Connection Assumed -Penalty - Database mustbe created (Time + 1Minute; Steps + 5)

01:18 8

Availability / Perform Recovery 0:00 0

Sybase Central -Connection Assumed -Expand Databases ->Expand Tables -> Rightclick table -> Select DataTransfer -> BCP -> Out ->Select Finish

00:09 0



31



Task: Recover dropped table

Perform Recovery / Change thescope to tables / Operation Type toFlashback Dropped Tables / ClickRecovery

0:03 3

Interactive SQL - Truncatethe table in question thatmust be restored - Penaltyfor manual process + time

01:00 5

Defaults selected / Object Recovered 0:23 0

Sybase Central -Connection Assumed -Expand Databases ->Expand Tables -> Rightclick table -> Select DataTransfer -> BCP -> In ->Enter Error File Location -

> Select Finish (Note, itwould typically be inanother location)

00:05 1


Task 14: Recover Corrup t Datafi le



Task: Recover Dropped Datafile

Cannot connect to the DatabaseInstance / Select Perform Recovery

0:00 0

No Concept of recovery ofa datafile. Only adatabase. - Penalty +10steps + 2 minutes time

02:00 10

Perform Recovery / Provideusername / password combination

00:06 2

Perform Recovery / OracleSuggested Recovery

00:02 1

Manage Failures / Review / DatafileMissing / Select Advise

00:04 2

Restore File / Recover 1:20 0




32

Task 15: Recover from erroneous transact ion



Task: Recovery from erroneous transaction

Enterprise Manager Homepage /Select Schema / Select Tables

0:04 1

Cannot Restore fromerroneous transaction,without resorting an entiredatabase, and movingspecific transactions intothe newly restoreddatabase. - A databasecan be restored to aspecific time, however nota specific transaction:Penalty Time + 10 mins;Steps + 10

10:00 10

Table Search – EnterOwner/Schema / Select Go / Selectthe table to flashback

00:09 2

View Table: Select Flashback Tablein Actions Go / Pick a point in timeto flashback to input it intoFlashback to a timestamp

00:12 2

Perform OLR / Flashback Tables /Defaults Selected / Select Next

00:04 2

Confirmation / Select Next / Review

/ Select Submit 00:49 1




33

Performance and Tuning Tasks

Task 16: Problem Diagno sis


Task: Configure System to Identify Top Resource-Consuming SQL

Enterprise Manager Homepage /Review latest ADDM findings andORA errors in the alert log

0:02 0

Penalty - Usability -Interactive SQL must beemployed and it is amanual process throughthe various mon tables.

02:00 5

Drill into each KIP to find root cause 0:06 2


Task 17: Tune Instance Memory



Task: Tune Instance Memory

Automatic Memory Tuning isenabled as part of the installation

0:00 0 Automatic Memory Tuningis enabled as part of theinstallation

0:00 0




34

Task 18: Tune SQL Statements



Task: Tune SQL Statements

Automatic SQL Tuning used for thisTask / No penalty / Nothing to do

0:00 0

Penalty - Interactive SQLmust be employed tomanually look at variousmon tables, + deepunderstanding of thedatabase structure mustbe obtained in order toproperly tune thestatements, or a third partytool must be employed(i.e. SQL Expert Option for ASE or Quest, etc.)

02:00 5




35

Appendix A: Test Platform Details

Hardware

Processors: Dual Xeon 5520

Memory: 12 GB DDR, 6x2 GB

Storage

o Primary system drives

2 146 GB SAS 15k RPM RAID 1

o Data drives

3 146 GB SAS 15k RPM RAID 5

Operating System Sybase Windows Server ASE 15.7 R2 Enterprise

Oracle

Oracle Database 11g Release 2 Enterprise Edition (64 bit)

Oracle Enterprise Manager Diagnostics and Tuning Packs

Oracle Database Configuration Assistant

Sybase

Sybase ASE 15.7

Sybase Profiler

Sybase Agent

Sybase Database Engine Tuning Advisor



36

Appendix B: New Features

Oracle Database 11g Release 2

The list of new features contained in Oracle Database 11g Release 2 Enterprise Edition can befound at: http://download.oracle.com/docs/cd/E11882_01/server.112/e17128.pdf

Sybase ASE 15.7

The list of new features contained in Sybase ASE 15.7 can be found at:http://msdn.Sybase.com/en-us/library/bb500435.aspx

http://download.oracle.com/docs/cd/E11882_01/server.112/e17128.pdf



http://msdn.microsoft.com/en-us/library/bb500435.aspx






37

Appendix C: Detailed Task Descriptions

Testing: Administration WorkloadOut-of-the-Box Installation A standard installation of both DBMS platforms, accepting all defaults and only providing input whenrequired in order to complete the steps. These installation steps do not account for any customization forthe server platform for use in a production environment.

Setup of New Database InstanceUsing provided tools and wizards, only defaults were selected and input was provided only when requiredin order to complete the steps. This step focused on the creation of a new database/instance for theDBMS.

Proactive Monitoring Setup & ConfigurationUsing the provided tools, monitoring was configured for both DBMS platforms. Defaults were selected.

Daily Administrative Tasks

These tasks were identified as daily administrative tasks that DBAs typically perform.

Object Creation: Tablespace/Data File

This task tested the creation of additional space for the database, in a separate file.

Object Creation: User/Schema

This task tested the creation of an additional user for the database. This could be viewed as either atemporary user account for access to the database, or a new application user of a more permanent

nature.

Object Creation: Table

This task tested the creation of a new table.

Object Creation: IndexThis task tested the creation of a new index on a given table.

Space ManagementThese tasks tested the ability and ease to add and remove space from existing data files.

Add Space to Tablespace/Data FileThis task tested the ease of adding space to an existing tablespace that had grown to capacity.



Reclaim Fragmented SpaceThis task tested the ability to remove unused space from a data file.

Testing: Backup & Recovery TasksThis set of tasks tested various backup and recovery options available in both DBMS platforms.

Configure & Run Full Backup(s)In this test we checked the ability to run a single backup at user request, which was not scheduled.

Recover Dropped Table

In this test we compared the DBMS platform’s ability to recover a single table that had been lost eitherthrough user error or through corruption.

Recover Dropped Datafile

In this test we compared the DBMS platform’s ability to recover a single data file that had been lost either

through user error, or through corruption.

Recover from Erroneous Transaction

In this test we recovered from a transaction that was run in error to either mass update a table or group oftables.

Testing: Performance & Tuning TasksIn this group of tests we compared the DBMS platform’s ability to quickly diagnose and resolveperformance problems related to the everyday usage of the database platform in question.

Problem Diagnosis

For this test we looked at the ease in which problems could be located, and how quickly user responsecould be determined.

Instance Memory Tuning

This test was not run since both platforms provide the out of the box instance memory management. Thistest will be removed from future iterations of this study.

SQL Statement Tuning

In this test we determined the ability for each DBMS platform to quickly isolate and resolve problematicSQL statements.

Documents

Oracle Db11gr2 vs Sybase Ase Cost 1587043