PaPM Tech
PaPM Tech
1 SAP Profitability and Performance Management Cloud – System Sizing, Tuning and Limits
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3
1.1 System Sizing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Memory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Storage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Optional Data Lake. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Optional Network Transfer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Sizing by Use Case. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5
Sizing by Data Size. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Mass Sizing and Performance Test. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.2 System Tuning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Limits. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.4 Background and Expert Information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
This Sizing Guide provides guidelines regarding the sizing and tuning of your SAP Profitability and Performance
Management Cloud system. It also informs you about the limits for acquiring data.
The Sizing Guide also provides support on how to use the Capacity Unit Estimator.
SAP Profitability and Performance Management Cloud leverages HANA Cloud as the underlying database and
infrastructure. Consider the following factors for sizing:
● Memory [page 3]
● Storage [page 4]
● Optional Data Lake [page 4]
● Optional Network Transfer [page 4]
● Sizing by Use Case [page 5]
● Sizing by Data Size [page 5]
● Mass Sizing and Performance Test [page 6]
1.1.1 Memory
HANA trusts that there is always enough memory available. As a result, the system immediately cancels a
running process in case of lack of HANA Memory.
The HANA Cockpit allows administrators or tenant owners to activate an alert when memory peak
consumption exceeds the critical threshold of 80%.
● Master data and transaction data, including dimension tables and fact tables which are managed by SAP
Profitability and Performance Management Cloud to cover the respective use case(s)
● Log, audit and comment tables
● Temporarily buffered and cached data from HANA Smart Data Access and HANA Smart Data Integration
The system immediately cancels the respective running process in case of a lack of HANA disk space.
The HANA Cockpit allows administrators or tenant owners to activate an alert when memory peak
consumption exceeds the critical threshold of 80%.
Data Lake is optional and only recommended if you need to acquire and store large amounts of data from other
applications at a low price. Working with such data is at least 30% slower compared to data in storage, so it is a
trade-off between speed and cost. If data lake capacity is zero, this is not relevant.
The HANA Cockpit allows administrators or tenent owners to activate an alert when memory peak
consumption exceeds the critical threshold of 80%.
Network transfer is optional and only needed if you want to use micro services, such as Workbook Calculation,
JavaScript Adapter or Python Adapter from outside the SAP Profitability and Performance Management Cloud
system, for example, when you want to call it directly from:
Note
This option is only available in the capacity unit estimator. It is not needed in the Provisioning Agent.
The Sizing Estimator allows for sizing by use case. Simply select a use case, and the system will automatically
suggest sizes for memory and optional data lake and network transfer. Sizing based on use cases proposes
typical sizes, which are enough on average.
Example
The Tax Calculation and Reporting use case typically reads directly from your general ledger and financial
statement items. Therefore, the data volume is rather small and acquisition of larger amounts of data is not
needed.
You can adjust the sizing proposal, if needed, based on more specific requirements. For more information, see
Sizing by Data Size [page 5].
The Sizing Estimator allows for a custom manual sizing based on specific requirements for a given standard or
non-standard use case.
Example
If you want to use the Tax Calculation and Reporting use case based on acquired data, you can increase the
optional data lake to one or more TB.
High-level statistics show that 500 GB memory is sufficient for 50% of customers, and 1000 GB memory is
sufficient for 90% of customers to cover their use case. This might already be a good indicator for your sizing.
On a medium level, you can assume that one year of calculation-relevant data should not occupy more than
40% of the available memory. For example, if 200 GB of data per year is read during the processing run, then
the total memory should be at least 500 GB.
The estimation of the optional data lake is simple: Sum up the TB of data that you want to store in the data lake,
and double it. This way, you make sure you stay on the save side. For example, if you plan to acquire 3 TB of
data, then go for a 6 TB sizing.
The estimation of the optional network transfer is just as simple: Sum up the GB of data relevant for the
calculation by the micro service times the number process runs per month, and double it to stay on the save
side. For example, if you want to use the workbook calculation from within an own application to calculate 100
GB of data 4 times per month, then choose 800 GB.
For more detailed information about sizing and tuning, see System Tuning [page 6].
Comprehensive scenarios processing millions of data records or more require a proper sizing and performance
test to ensure that the calculation process in your production environment runs smoothly in time and capacity.
SAP Profitability and Performance Management Cloud allows you to easily increase the memory as well as the
optional data lake and network transfer, if needed.
You can use the following options to improve the performance of your SAP Profitability and Performance
Management Cloud system:
1. Column pruning
Typically, not all columns of your data tables are really needed for a calculation, which opens up potential to
reduce the necessary memory and increase the processing speed, if only needed columns are read and
processed.
2. Row pruning
Typically, not all rows of your data tables are actually needed for all functions in a calculation process. The
golden rule is to filter the data as early and as much as possible. This ensures that the functions only
consider data rows that are actually needed.
3. Aggregation
Depending on the use case, you can aggregate data as much as possible. This ensures that the functions
work with the smallest data set possible.
4. Optimized function use
○ Make use of function rules
Instead of using a long chain of functions with one rule each, use few functions with multiple rules
each.
○ Use the Join function to aggregate data from different tables or sources
Instead of using a long chain of views and joins, use one Join function with multiple rules to collect and
aggregate data from different sources, and to distill exactly the columns, rows and aggregation you
need.
○ Make use of specialized functions
Instead of covering your business logic in a long chain of views and joins, leverage the power of specific
business functions and rules, such as Allocation, Calculation, Derivation, Transfer Structure etc. can
cover your business requirements in an efficient and easy-to-configure way.
5. Package parallelization and sequencing
If your use case deals with dozens or hundreds of millions of records or more per run, use partitions to
chunk the data into logical pieces. This increases the parallel throughput tremendously. It also gives you
more control over the overall memory consumption because it does not only lead to a higher parallelization
throughput, but also to less memory consumption for data sequencing.
6. Use and learn from the content packages
SAP Profitability and Performance Management Cloud provides content packages for a large variety of use
cases. The content packages can be imported via content network. They provide a set of best practices.
The content packages provide support on good model design, in particular:
1.3 Limits
● After 15 minutes of user inactivity (no requests have been sent), a central logout will be triggered due to
session timeout.
● Model Table import of data from an Excel file on the client computer is limited to 20 MB. We recommend
using CSV files in case you need to import amounts of data.
● Memory: 5970 GB
● Data Lake: 90 TB
● Network Transfer: 20000 GB per month
This section provides further background and expert information that help to fundamentally understand the
solution architecture.
Data Tiering
SAP Profitability and Performance Management Cloud can work with different kinds of local data offered by
HANA Cloud, in particular HANA Memory (Hot Store), HANA Storage (Warm Store) and optionally HANA Data
Lake (Cold Store):
HANA Memory is used for all processing and calculation functions. It contains the most frequently used
(buffer) data. This is what makes SAP Profitability and Performance Management Cloud so fast.
It is recommended to locate frequently used data in the HANA Storage. If your use case works with large
amount of data from other applications, we recommend to use the HANA Data Lake which can then be
integrated to SAP Profitability and Performance Management Cloud's underlying SAP HANA Cloud Database.
Some customers prefer to separate the tenants of SAP Profitability and Performance Management Cloud, for
instance:
The sizing depends on the tenant because the usage is metered across all tenants, and the sum of capacity
units across all tenants is compared to your subscription. As a result, you can flexibly decide about the number
and capacity of the tenants you need.
Overuse is not blocked by SAP to ensure that no operation on customer side is endangered. Instead, overusage
is reported in the customer cloud reporting.
Hyperlinks
Some links are classified by an icon and/or a mouseover text. These links provide additional information.
About the icons:
● Links with the icon : You are entering a Web site that is not hosted by SAP. By using such links, you agree (unless expressly stated otherwise in your
agreements with SAP) to this:
● The content of the linked-to site is not SAP documentation. You may not infer any product claims against SAP based on this information.
● SAP does not agree or disagree with the content on the linked-to site, nor does SAP warrant the availability and correctness. SAP shall not be liable for any
damages caused by the use of such content unless damages have been caused by SAP's gross negligence or willful misconduct.
● Links with the icon : You are leaving the documentation for that particular SAP product or service and are entering a SAP-hosted Web site. By using such
links, you agree that (unless expressly stated otherwise in your agreements with SAP) you may not infer any product claims against SAP based on this
information.
Example Code
Any software coding and/or code snippets are examples. They are not for productive use. The example code is only intended to better explain and visualize the syntax
and phrasing rules. SAP does not warrant the correctness and completeness of the example code. SAP shall not be liable for errors or damages caused by the use of
example code unless damages have been caused by SAP's gross negligence or willful misconduct.
Bias-Free Language
SAP supports a culture of diversity and inclusion. Whenever possible, we use unbiased language in our documentation to refer to people of all cultures, ethnicities,
genders, and abilities.
SAP and other SAP products and services mentioned herein as well as
their respective logos are trademarks or registered trademarks of SAP
SE (or an SAP affiliate company) in Germany and other countries. All
other product and service names mentioned are the trademarks of their
respective companies.