In this blog, which succeeds Efficient SAP Hardware Sizing: Quick Sizer and Quick Sizer 2005 – Getting started, I’ll focus on interpreting the different sizing results. For reasons of better analysis, the Quick Sizer allows you to analyze sizing results at very different levels of detail. Here are the most important ones.
Different Result Levels
There are seven different result levels that show different perspectives on the sizing results. The ones that provide most important information for analysis purposes are described in detail; the others share one sub chapter.
One important principle of the Quick Sizer is that it calculates the results for each hour over a 24-hour time period and displays the highest result at any given hour. For example, on sizing project level, the result displayed is the highest over all solutions. At solution level it shows the highest result for each solution. Another principle is that the sizing results for user sizing and for throughput sizing are calculated and displayed separately.
Result by solution
By default, the Quick Sizer displays the sizing result by sizing method (user and throughput) and by SAP solution, for example SAP ERP or SAP NetWeaver. For each solution the Quick Sizer displays the required:
- CPU power in SAPS (hardware-independent, see above) at a target utilization of 65%, rounded in units of 100
- Disk space in MB, rounded in units of 1000
- Disk I/O in operations in seconds in units of 100
For easier analysis, there are also result categories, ranging from small (S) to extra large (XXL). The categories are to provide some estimate what size your project is in relation. XXL does not really indicate a very large sizing, it only means that if you have sizings beyond 30,000 SAPS and/or 1 TB you should not rely on a tool such as the Quick Sizer alone but check with your hardware provider.
Table: Sizing categories
|Category||Up to … SAPS||Up to … MB disk||Up to … I/Os per second|
|XXL||Contact hardware vendor or SAP for detailed sizing analysis|
If you only have a user-based sizing and if this result exceeds category S, you should either perform a throughput sizing or contact your hardware vendor or SAP. If you did both sizing methods and the category exceeds XL, you should also contact the hardware vendor or SAP for help.
The results at solution level include disk and memory offsets to reflect the resources you need for the standard software. These offsets may be different for each solution.
Result at software component level
To plan your system landscape it is helpful to know the hardware requirements for each software component, for example Enterprise Core Component (ECC) or Portal Server. We define software component as any a separately installable deployment unit.
At this level, the Quick Sizer provides more detailed information, for example the ratio between database server and application server, an information which may become important if you are sizing a multi-tier environment and need to now how to split up DB and application resources. The results for CPU and memory consist of three columns, total, DB and App.
|Total||Total displays the highest requirements at a given point in time|
|DB||Highest DB requirements at any given point in time (individual peak)|
|App||Highest application requirements at any given point in time (individual peak)|
However, it may be that the individual peaks for DB and application server may be at different times, so in order to better configure CPU and memory, the values for DB and App reflect the individual peaks. As a consequence of this independence, total is not necessarily the sum of the results for DB and App but instead may be higher when added. See an example below, where the (rounded) individual results for DB and App are higher than the total result.
If you size mySAP SCM, you’ll find additional columns for livecache memory.
The results at software component level also include disk and memory offsets to reflect the resources you need for the standard software. These offsets are different for the different software components. Note that we assume one physical application server only. If you plan to install more application servers you particularly need to add memory.
Inputs, Statistics and Results overview
This function is predominantly useful for documenting the sizing project as it contains all the data you entered and all the results of the Quick Sizer. With the print function you can download this information to save on your PC. This function is also helpful if you want to analyze the sizing data in more detail. More often that you think you can detect erroneous entries by simply doing a plausibility check on the highest CPU or disk contributors. This result level also provides detailed information on the top disk contributors, the table names, the projected growth after one year and according to the specified retention period, available archiving objects, and so on.
Other Result levels
- “Project” displays the result for all questionnaires added.
- “Key capabilities” displays the result at questionnaire level. This level does not include offsets.
- “Sizing element” displays the result at sizing element level. This level includes the archiving objects specified for the respective sizing element. It does not include offsets.
- “Line results and inputs” is helpful for cause-effect analyses of your entries and the respective results.
Interpreting the results
Interpreting SAPS or disk results has a lot to do with applying some common sense to what you see. A very good method to spot discrepancies is to compare the results for users and throughput. Huge differences may indicate wrong assumptions. Usually, the user activity is rated too high. The screenshot below shows a typical example from the solution level of the result screen.
The CPU has a difference of a factor 16, and the disk shows a difference of a factor 6. That there is no memory result for the throughput is ok, since memory in CRM is user context-driven.
Let’s take a look at the inputs on the result level for inputs and results.
The Results table shows how the result splits up into different sizing elements. The highest contributor is SLS-USER, the user in sales transactions. When you compare the results between the users and the respective throughput you can see that the factors are very high as well.
The input data show the data entered. For example, there are 130 active users in Activity Management (ACT-USER). When you compare this data with the 50,000 activities in the throughput (CRM-ACT), you can understand the discrepancy. Assuming 220 workdays per year, approximately 30 activities are being created per hour (50,000/220*8). If you assume 130 active users as done above, then there is something rotten. The other sizing elements follow suit in this argumentation. The question that now arises is: what data is incorrect, the volume numbers or the user figures? Experience shows that the user activity is often overrated. Often, customers enter named users instead of active users. Even if users fill in screens rapidly, the time between the saves or other actions that create load on the application server is termed think time.
Frequently asked Questions on Quick Sizer results
Q: There are results for users and throughput. Do I have to add them?
A: No. They are obtained separately and should not be added. Ideally, they are similar. If there is a strong divergence between both results, there are probably entry errors in the questionnaires or wrong assumptions about the business case or user activity.
Q: Does the Quick Sizer consider Unicode?
A: Yes. just make sure that the corresponding SAPS ratings of the hardware configuration include Unicode as well.
Q: I only have user information (3000 ERP users); can I do a user-only sizing?
A: We recommend not to. User-only sizing can be quite treacherous because it doesn’t consider background activity, system-system communication and others, for example.
Additional functions and options on the result screen
A list with the relevant functions is shown below.
|Display charts||To get a graphic impression of the 24-slope of CPU requirements, you can use this function.|