Business transactions are mainly conducted using enterprise software. The use of this software is responsible for a large percentage of the carbon emissions of an enterprise software provider such as SAP. To understand and reduce the energy consumption induced by software, quantifying it and relating it to the source code is crucial. This paper addresses this imperative by presenting a modular approach to measure the energy consumption of code statements written in SAP’s proprietary programming language ABAP. Various measurement approaches were evaluated, leading to the development of a prototype based on an estimation model. The prototype, implemented in an SAP S/4HANA environment running on an IBM Power architecture, demonstrates a mean absolute percentage error of less than 1%. Initial measurements show that it is possible to measure differences in energy consumption between ABAP implementations, such as iterative versus recursive approaches for solving the Fibonacci problem. This research contributes an evaluated prototype tailored to ABAP statements and serves as a valuable resource for future empirical investigations in the domain.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Towards an Energy Estimation Model for ABAP Statements on SAP S/4HANA Software

  • Ann-Christin Fleischle,
  • Laurenz Baumgart,
  • Holger Wittges,
  • Helmut Krcmar

摘要

Business transactions are mainly conducted using enterprise software. The use of this software is responsible for a large percentage of the carbon emissions of an enterprise software provider such as SAP. To understand and reduce the energy consumption induced by software, quantifying it and relating it to the source code is crucial. This paper addresses this imperative by presenting a modular approach to measure the energy consumption of code statements written in SAP’s proprietary programming language ABAP. Various measurement approaches were evaluated, leading to the development of a prototype based on an estimation model. The prototype, implemented in an SAP S/4HANA environment running on an IBM Power architecture, demonstrates a mean absolute percentage error of less than 1%. Initial measurements show that it is possible to measure differences in energy consumption between ABAP implementations, such as iterative versus recursive approaches for solving the Fibonacci problem. This research contributes an evaluated prototype tailored to ABAP statements and serves as a valuable resource for future empirical investigations in the domain.