Skip to Content
Technical Articles
Author's profile photo Rami Kandimalla

SAP S/4HANA End-to-End Performance

This blog post targets SAP System Administrators and Technical Architects to provide some insights into End-to-End Performance of SAP S/4HANA

What is Performance?

The most important indicators for performance are the End-to-End Response Time and throughput.

In SAP systems, many factors can cause performance bottlenecks. Hardware bottlenecks and bad configuration can cause problems, Insufficient hardware, like slow CPUs or not enough main memory, can make an SAP system unusable. The resource consumption (CPU, Memory, I/O, Disc) on all involved components has also to be considered for all types of processing.

System Sizing

Sizing is the process of translating business process requirements into hardware requirements; sizing exercise determines the hardware requirements of the SAP S/4HANA system, SAP HANA database as well as the disk size, the required memory and required network throughput.

For more Information on Sizing, depending on the “To Be” system release and scenarios, please refer to the following sources

The correct sizing of the SAP system is very important, otherwise the performance of SAP system will be below its potential.

End-to-End Performance

The end user performance is mainly correlated to the End-to-End response time; several factors loosely constitute End-to-End response times,  it is the sum of backend response time (the time to retrieve data and process the requested response), the time to transport request and response data between client and server, and the time to finally render the result in.

SAP S/4HANA End-to-End Performance

SAP S/4HANA Architecture consist of 3-tiers, a SAP Fiori UI5 frontend, an SAP S/4HANA ABAP backend and a SAP HANA database

Database and ABAP processing times are two primary components of ABAP response time. Effects of pushing code to SAP HANA changes standard metrics of ABAP and database time as well as the common SAP R/3 standard of 1 second per Dialog step.

Some SAP HANA features and ABAP coding that contribute to backend response times.

  • core data services (CDS)
  • ABAP Managed Database Procedures
  • Compatibility views
  • Parallel processing in SAP HANA
  • Columnar in-memory database

End-to-End performance of an SAP S/4HANA system is spread across 4 layers: SAP Fiori UI, SAP S/4HANA  ABAP layer, CDS views and SAP HANA

SAP S/4HANA ABAP

Potential ABAP Application related areas which impact End-to-End performance of SAP S/4HANA

  • ABAP coding errors
  • Inactive ABAP objects
  • Inactive Business Functions
  • Missing / Wrong Functional configurations
  • Missing Backend Authorizations
  • Inactive CDS views (DCL, DDL)
  • High OData response times
  • Inactive OData services

Refer to the following sources for additional information on SAP S/4HANA performance Best Practices

  • 2689405 – FAQ: S/4HANA Performance Best Practices – Collective Note
  • 1794297 – Secondary Indexes for S/4HANA and the business suite on HANA

SAP HANA

Due to the column-based architecture, SQL access to tables with a high number of entries can be very fast on SAP HANA, even without additional secondary indices. By observing the general symptoms of SAP HANA systems, resource bottleneck (CPU, Memory and I/O), paging or column store unloads are possible causes of poor performance.

Refer to the following sources for additional information on SAP HANA performance Best Practices

  • 2000000 – FAQ: SAP HANA Performance Optimization
  • 2600030 – Parameter Recommendations in SAP HANA Environments
  • 1999997 – FAQ: SAP HANA Memory
  • 1698281 – Assess the memory consumption of a SAP HANA System

SAP S/4HANA Core Data Service (CDS) Views and VDM

Core Data Services (CDS) is an enhancement of SQL which provides a data definition language (DDL) for defining semantically rich database tables/views (CDS entities) in the database. The enhancements include annotations, associations, and expressions. The Virtual Data Model based on Core Data Services (CDS) exposes business relevant data in a semantically rich and harmonized way. It provides HANA-optimized access to business data via SQL. The S/4HANA Core data model is semantically and structurally leading, ensuring correct cross-references between traditional and new business processes.

Accessing CDS views impacts quite significantly the resource consumption on the persistency layer. From an architecture point of view there are two main differences compared to the way we have accessed business data before. In the past only transactional data and some master data had been accessed from persistency; all kind of Metadata (customizing, texts, etc.) have been cached on the Application Server. Using CDS views, simple direct access to a (maybe even buffered) table may be replaced by a complex (layered) view including joins with Metadata and authorizations.

For additional details on troubleshooting performance of ABAP CDS views

SAP Fiori

SAP Fiori is a collection of apps with a simple and easy-to-use experience for broadly used SAP software functions that work seamlessly across devices such as desktop, tablet, and smartphone.

One major aspect of good user experience is a responsive UI. This means we must respond to a request or interaction triggered by a user (on any supported device) in a way that continuous non-disruptive working is possible everywhere.

Potential SAP Fiori related areas which impacts End- to-End performance

  • Rendering errors
  • Wrong app routing
  • Wrong app navigation
  • Missing tiles
  • Missing Frontend Authorizations
  • High UI load times

For additional details on troubleshooting performance issues for SAP Fiori.

Summary

A good architecture does not implement everything that is possible, but only what is actually required. Tuning SAP S/4HANA system performance requires knowledge and collaboration from different areas, performance problems often result from a combination of different factors. Good KPIs should reflect the performance both from a user and system point of view, break the End-to-end response time down to component time and stay in defined budget.

– Brought to you by the S/4HANA RIG –

Assigned tags

      13 Comments
      You must be Logged on to comment or reply to a post.
      Author's profile photo Joseph Chinnabathini
      Joseph Chinnabathini

      Great Article Anna.

      Author's profile photo Rami Kandimalla
      Rami Kandimalla
      Blog Post Author

      Thank you, Joseph.

      Author's profile photo sunil patil
      sunil patil

      Good Article

      please share more such articles

      Author's profile photo Rami Kandimalla
      Rami Kandimalla
      Blog Post Author

      Thank you Sunil

      Author's profile photo Syambabu Allu
      Syambabu Allu

      Excellent Blog.

      Thank you,

      Syam

      Author's profile photo Rami Kandimalla
      Rami Kandimalla
      Blog Post Author

      Thank you Syambabu

      Author's profile photo Omprakash Senapathi
      Omprakash Senapathi

      Good one. thanks.

      Author's profile photo Rami Kandimalla
      Rami Kandimalla
      Blog Post Author

      Thank you Omprakash

      Author's profile photo Frank Dorr
      Frank Dorr

      Good overview! Thank you.

      Author's profile photo Sai Yerrapragada
      Sai Yerrapragada

      This is very nice article. Thanks for your time

      Author's profile photo Hitesh Kumar
      Hitesh Kumar

      Very nice article. Thanks Rami.

      One question- In traditional R3 world there was Dialog Response time benchmark as 1 Second. Do we have any benchmark for S4HANA Dialog Response time, recommendation from SAP? Is there any benchmark for individual layers UI, SP HANA, CDS or S/4 ABAP?

      Author's profile photo Rami Kandimalla
      Rami Kandimalla
      Blog Post Author

      Thanks Hitesh. Regarding SAP S4HANA's dialog response benchmark, 1s(0.5-1.5s) is for simple tasks and for complex tasks it is 3s.

      You can refer to this collective note for additional details.

      2689405 - FAQ: S/4HANA Performance Best Practices - Collective Note

      Author's profile photo Hitesh Kumar
      Hitesh Kumar

      Its helpful, thanks Rami. So it becomes little bit subjective to say which task is easy and which is complex tasks. May be complexity of a task impacts HANA layer response time, other 3 layers would not have much difference, be it simple or complex scenario.

      It would be better if SAP can share some % benchmarking for response time at each layer level, as we had with traditional R3- CPU/DB time should not be more that X% of Total Dialog response time. It helped us to identify problematic area,  whether it is buffer or CPU or DB.