Skip to Content

After the completion of the business blue printing phase and before starting SAP Spend Performance Management (SPM) implementation, hardware sizing is highly recommended. The official Quick Sizer project for SPM helps you to determine the right hardware sizing. Refer to note 1253768 for further explanation of the usage of Quick Sizer tool.

 

Before using tool, it will help to do some groundwork and find data volumes for Transaction and Master Data. Make sure that you take into account multiple source systems, external data feeds like bank feeds for expenses etc based on the business scenario. The time frame for extracting the data should also be considered based on feedback from Business User. For example last five years or 2008 to current date etc.. Based on the industry segment/type of business, an appropriate time frame can be chosen. Best practice recommendation is to go back to the point of relevance for Analyzing Spend and avoid “data overloading”. An example is that Freight spend which is heavily impacted by oil prices, suffers lot of volatility which is not necessarily predictable, hence a timeframe going back only a couple of years might suffice. Following is the list of information that needs to be made available to be able to use the SPM quicksizer. 

 

Transactional information:
1)      Number of Invoice Lines
2)      Number of pCard and T&E expense lines
3)      Number of Contract Lines
4)      Number of Scheduling Agreement Lines
5)      Number of Purchase Order Lines
6)      Number of Delivery Lines
7)      Number of Budgeting and Forecasting Lines
8)      Number of Project Lines
9)      Number of lines for commodity/market pricing information 

 

Master Data information:
1)      Number of Item/Products
2)      Number of Categories
3)      Number of Suppliers
4)      Number of Buyers
5)      Number of Buying Organizations
6)      Number of Cost Centers
7)      Number of GL Accounts
8)      Number of Site/Plants
9)      Number of Management Orgs 

 

The above list represents majority of the sizing volume inputs that would be required. Remember to inflate the above numbers appropriately in order to be able to accommodate for future loads until the next hardware upgrade is planned. In addition, consider the number of concurrent users productively using the SPM application. Now, armed with the above information, log into the SPM quicksizer tool and follow the instructions of the note 1253768 to get the recommended hardware specifications. For more information on how to use/understand the quicksizer tool check out the following blog links, Efficient SAP Hardware Sizing: Quick Sizer and Quick Sizer – Getting started

 

Typically, SPM can be  deployed either on an existing BW instance (which includes other analytics) or in a standalone setting where it doesn’t compete for system resources. Both are valid deployment scenarios. The second scenario is straight forward, all you need to do is take the recommendations from the Quicksizer tool, validate them with the system administrators and commission the machine. In the first scenario, the hardware sizing results can be used to extend the existing resource allocations. The disk space is additive to the existing configuration, but the SAPs and RAM might not be, especially if the business users of the other analytics (that SPM shares the box with) do not have the same work hours. Remember it’s the concurrent usage that typically drives RAM and SAPs numbers.  

 

Happy sizing!!! 

 

(Remember, additional information can be found at the SPM doc location in Service Market Place, link

To report this post you need to login first.

2 Comments

You must be Logged on to comment or reply to a post.

  1. Padmini Ranganathan
    This is a useful guidance to calculate the hardware impact and/need.  If you already have sufficient hardware capacity, you may not need to add new hardware to deploy Spend Performance Management.  Newell Rubbermaid for example did not have to invest any additional hardware or IT resources to deploy Spend Performance Management.(Reference: http://www.sap.com/usa/about/newsroom/press.epx?pressid=13282, or get a copy of the customer reference slide from your account representative).

    However, some customers may like to take an appliance approach and deploy Spend Performance Management in a separate instance to take advantage of visibility and savings opportunities without having to wait for a BI upgrade for example. 

    (0) 
    1. Rajesh Prabhu Post author
      Thanks Padmini.

      That’s absolutely correct, a sizeable number of customers who have implemented the solution have actually deployed it on their existing boxes. Many of them didnot have to change the specs.

      We quite often find that the existing systems have sufficient capacity (or can be brought up to sufficient capacity by minor efforts), especially if the existing systems were aggressively spec’d and have more than whats required for existing applications. Here are some tips to get started:
      * Do we have more storage than what we need already? Tip> Industry survey – http://www.computerworld.com/s/article/9178781/Survey_finds_storage_systems_underutilized_as_companies_add_more_capacity
      * Have we factored in the fact that not all historic data (not only for Spend but also for other applications that coexist on the same system) is relevant and that our archival process will reduce that volume going forward.
      * CPU cycles and Memory: Are all the applications being utilized in the same timezone? Potentially, one application might be run from plants in China and Spend might be run from US, where the global procurement dept might be located. So in terms of time zones the applications compete verry little (from an overlap perspective).
      * Based on usage statistics, are the same number of concurrent users logging onto those existing applications as originally spec’d?

      We need to ensure that we provide the adequate resources for the applications to function, and function with satisfactor performance. But we can be definitely be smart about it!

      (0) 

Leave a Reply