SAP HANA Application Accelerator - hdespot.comhdespot.com/wp-content/uploads/2015/09/SAPHDE_Webinar-SAP_HANA...Redirect database reads for ABAP programs or customer namespace (Z programs)

SAP HANA Application Accelerator Use cases and implementation

Justin Molenaur MoleHill Technology Solutions, LLC. July, 2015

©  2015 MoleHill Technology Solutions, LLC.. 2

Content

SAP HANA Use Case Overview

SAP HANA Application Accelerator §  Why is this useful? §  When would we use this?

Implementation §  Overview §  Implementation Steps §  Constraints §  Configuration Notes

Results and Recommendations §  Sample Results §  Methodology for acceleration candidates §  Lessons Learned §  Performance Guidelines

References

SAP HANA Application Accelerator Use Cases Why is this useful? When would we use this?


SAP HANA Use Cases

Application Database

•  S/4 HANA •  BW on HANA

Process Acceleration

• CO-PA Accelerator

•  FI Accelerator •  Application

Accelerator

Analytics / Native

• Data Warehousing Platform

•  XS Apps


Why is this useful?

HANA is FAST

Increase performance of ABAP processes by reading HANA instead of AnyDB


When would we use this?

Need for Speed

Have HANA investments

already

Don’t want to change

code

1.  Need to accelerate custom code

2.  Existing or planned HANA investment

3.  No or limited code changes

4.  Near real time latency is acceptable

SAP HANA Application Accelerator Overview Architecture and Implementation Steps Constraints Configuration Notes


HANA Application Accelerator Summary

§  Redirect database reads for ABAP programs or customer namespace (Z programs)

§  Targets specifics program names and tables

§  Only SELECT operations are accelerated

AnyDB

SAP Netweaver

System Redirect?

SELECT Y

N

HANA


SAP HANA Application Accelerator Detailed Architecture

Source System Redirect Configuration

Replication SAP HANA Sources: •  SAP ERP •  SAP APO •  SAP CRM •  Any Netweaver

application

AnyDB

HANA SAP

Netweaver System

SELECT INSERT UPDATE DELETE

SELECT ONLY

SAP LT Server

Sybase Rep Server

Program Name / Background Job Name

AND

Table Name SDI


SAP HANA Application Accelerator Implementation Steps

Sources: •  SAP ERP •  SAP APO •  SAP CRM •  Any Netweaver

application

AnyDB

SAP Netweaver

System


1.  Evaluate Pre-Requisites (OSS Note 1696402) •  SAP NetWeaver 7.0, 7.0 with EhP 1- 3 OR SAP NetWeaver 7.4.

•  SAP Kernel 7.21

2.  Acquire and Install software - SWT2DB add on

3.  Set parameter to switch on RDA (Redirect Database Access)


SAP HANA Application Accelerator Implementation Steps cont.

Source System Sources: •  SAP ERP •  SAP APO •  SAP CRM •  Any Netweaver

application

AnyDB

SAP Netweaver

System


SAP LT Server

4.  Ensure there is active replication to HANA for required tables

5.  Create secondary database connection in source system; use SLT replication user (if applicable)

HANA

SAP LT Server SAP LT Server

Sybase Rep Server

SDI



Source System Redirect Configuration Sources:

•  SAP ERP •  SAP APO •  SAP CRM •  Any Netweaver

application

AnyDB

SAP Netweaver

System


SELECT ONLY

Program Name / Background Job Name

AND

Table Name

6.  Create or acquire configuration file for scenario

7.  Upload configuration file

8.  Maintain database connection for scenario

9.  Activate scenario



Source System Sources: •  SAP ERP •  SAP APO •  SAP CRM •  Any Netweaver

application

AnyDB

SAP Netweaver

System


10.  Test Redirection

-  Activate trace in source system OR SQL trace in HANA


Redirection: Information and Constraints

§  Queries

§  Redirected: SELECT, OPEN CURSOR … FETCH

§  NOT Redirected: Queries with CONNECTION addition, SELECT … FOR UPDATE, Queries prepared with OPEN CURSOR WITH HOLD

§  Objects

§  Redirected: Transparent tables, Cluster tables, Database views

§  NOT Redirected: Pool tables, Table pools, Table clusters

§  Connect HANA database as the secondary database connection

§  The HANA database must be available, no DB fallback is possible without code change

§  SAP official statement is standard programs cannot be accelerated


Configuration Notes

§  The “program name” as defined within the Acceleration configuration file is based on the variable SY-CPROG during ABAP runtime.

§  In RFC calls the main program SY-CPROG is always set to SAPMSSY1.

§  In Web Dynpro applications the main program SY-CPROG is always set to SAPMHTTP.


Accelerator Maintenance

§  Transaction Code RDA_MAINTAIN – all maintenance activities performed here

SAP HANA Application Accelerator Sample Results Candidate Methodology Lessons Learned Performance Guidelines


Sample results

§  Across 30 scenarios, 2 – 60% decrease in DB read time can be expected with no code changes

§  Increase performance significantly if code can be changed and optimized

After Acceleration ~60% runtime reduction Before Acceleration

Join operation that got a huge improvement


Candidate Methodology

§  Look at Early Watch report or other monitoring to find programs which have a high DB read time

§  Talk with functional leads to find existing processes that have very long run times

§  Talk with the business to find processes that could have big business benefit

§  Execute a trace before acceleration and check the major tables affecting DB read time; SELECT, “LOOPs with For All Entry” statements


Lessons Learned

§  Engage process experts or stakeholders early to help prioritize acceleration benefits

§  Ensure that you perform tests with adequate volumes to mimic production volumes

§  Remove tables from acceleration that does not improve performance

§  As a general rule, perform an analysis of the program for obvious faults (binary search, sorting, etc) and correct those first.

§  Any function modules that are called from the main configured program will also be redirected

§  Look for specific indicators where code changes would bring a large benefit


Performance guidelines

§  Inefficient code will still be inefficient after acceleration

§  HANA will generally outperform AnyDB in following cases

§  When it comes to larger volume processing, bulk SELECT with specific column list

§  On SELECTs that contain any type of join, especially for joins on non-key/non-index fields

§  Aggregation operations

§  Any SELECT with a JOIN operation

§  HANA will generally perform the same as AnyDB in following cases

§  Many tiny reads from single tables (SELECT SINGLE)

§  SELECT * (all columns)

§  With programs that are very heavy % in ABAP processing already; large internal tables and loops


Reference information

§  SAP HANA Application Accelerator – OSS Note 1694697

§  This also contains the Accelerator Guide for customizing

§  SWT2DB Installation – OSS Note 1696402

§  Kernel 7.21 – OSS Notes 1713986 and 1716826

§  SWT2DB Troubleshooting – OSS Note 1949508

§  Secondary Database connection – OSS Note 1597627


Get Started - Implementation Overview

1.  Evaluate Pre-Requisites (Netweaver 7.0, Kernel 7.21)

2.  Acquire and Install software - SWT2DB add on

3.  Set parameter to switch on RDA (Redirect Database Access)

4.  Ensure there is active replication to HANA for required tables

5.  Create secondary database connection in source system; use SLT replication user

6.  Create or acquire configuration file for scenario

7.  Upload configuration file

8.  Maintain database connection for scenario

9.  Activate scenario

10.  Test redirection

Questions?

©  2015 MoleHill Technology Solutions, LLC..

Thank you

[email protected]

http://www.hdespot.com

http://scn.sap.com/community/distinguished-engineers

SAP HANA Application Accelerator - hdespot.comhdespot.com/wp-content/uploads/2015/09/SAPHDE_Webinar-SAP_HANA...Redirect database reads for ABAP programs or customer namespace (Z programs)

Documents