Technology Blogs by Members
Explore a vibrant mix of technical expertise, industry insights, and tech buzz in member blogs covering SAP products, technology, and events. Get in the mix!
cancel
Showing results for 
Search instead for 
Did you mean: 
TammyPowlas
Active Contributor

***UPDATED*** 2013

PLEASE SEE SAP.COM/BISIZING for the latest guidance or Sizing and Deploying SAP BI 4 and SAP Lumira

Jason DeMelo, SAP, provided an ASUG webcast back in July (sponsored by the BITI team) and also presented a BI 4.0 sizing session at TechED Las Vegas last month.  The usual disclaimer applies as this is informational and can change at any time.

Agenda:

Sizing Concepts

     

Sizing tools in BI4:

T-shirt Sizing Guides

Quick Sizer

Companion Guide

 

Sizing Concepts

Figure 1, Source: SAP

As Figure 1 shows, previous sizing for BusinessObjects was done in CPU units.  Starting in 4.0, processing requirements are now stated in SAPS.  SAPS are a hardware agnostic measure of processing capacity and used by several other vendors.

   

How do you calculate SAPS? 1000 SAPS = 2,000 fully processed order line items per hour.

Figure 2, Source: SAP

How are SAPS calculated? First a test deployment is selected and SAP standard SD benchmark is run (step 1, Figure 2. Then, on that same hardware the software is installed (BI 4.0) and tests are run.  The CPU time consumed is recorded.  Then in step 3  calculates CPU time per user per hour.  Step 4 is to scale the SAPS score for # of users and the CPU time the user will consume in an hour.

Figure 3, Source: SAP

When doing a sizing, start with the number of users.  Of that users, how many will be logged in at same time?  A rule of thumb is 10/10/10 rule: 10% of total user population will be logged in at any one time, and of those 10%, only 10% will be actively clicking away at the same point in time.

Sizing Tools in BI4

Figure 4: Source: SAP

Figure 4 shows the sizing tools.  T-Shirt sizing is good for smaller deployments.  The Quick Sizer, used for other SAP technologies,  is an online tool on Service Marketplace. 

 

The companion guide provides recommended settings and tuning tips.  The approach to testing is included.

Figure 5, Source: SAP

Figure 5 explains T-shirt sizing guides.  These are based on only number of concurrent users – those who are logged in and actively using the system at any point in time.  There is a T-Shirt guide for each client tool.

Figure 6, Source: SAP

Figure 6 shows the available T-Shirt Sizing Guides available (as of TechED last month)

Figure 7, Source: SAP

At the heart of the T-Shirt sizing guides are tables.  Based on size of deployment, SAPS and memory requirements are estimated are the frontend and back end.  ON the front end there is the application tier, processing tier, intelligence tier.  On the back end is the database.

Figure 8, Source: SAP

Figure 8 is an example of how this works.  Take the 50 users and find the closest match and record the SAPS and memory recommendations, and send them to your hardware vendor who will come back with recommendations based on SAPS and memory requirements.

Figure 9, Source: SAP

According to Jason, these are worst case user recommendations.  They do not want you have overstressed hardware and not size for multiple products on the same machine.  Figure 9 provides a caution.

Figure 10, Source: SAP

The Quick Sizer is for larger deployments and more detailed requirements.    Figure 10 shows three types of users:

  1. Information Consumer  - the lightest of the users – simple charts, table, light report document usage; little interaction
  2. Business User – click more than Information Consumer
  3. Expert User – take little time between clicks

Figure 11, Source: SAP

Figure 11 shows the products to be included in the BI 4 Quick Sizer.

 

Figure 12, Source: SAP

Jason said the sizing estimates for Quick Sizer have more fidelity and are broken out by tier.  The extra information will give you a better idea of how to plan landscape.

Figure 13, Source: SAP

First specify inputs and then the Quick Sizer web application will return sizing information.

 

What is different this time is that the sizing is for a mixed deployment, the users are broken by user type and documents are specified by small, medium, and large (see Figure 13 above).

Figure 14, Source: SAP

Based on the inputs into the Quick Sizer, the application will return SAPS and memory estimates by tier for each of the products used in the deployment.

In this example, to support a 50,000 user deployment, tell hardware vendor that you need just under 39K SAPS of processing capacity and 45 GB of memory.

Figure 15, Source: SAP

Figure 15 shows the availability of the Quick Sizer.

Figure 16, Source: SAP

The Companion Guide lets use understand the differences and it contains tuning and performance tips.

 

Links Provided in Presentation:

Quick Sizer Fact Sheet

Hardware benchmark results in SAPS

BusinessObjects BI

Where to go in Service Marketplace:

http://service.sap.com/quicksizing

Jason recommends checking the Service Marketplace frequently for up-to-date information.

In the SAP world we hear about SAPS all the time; at this session I learned all about them.

4 Comments
Labels in this area