BA202 – Know How and Where to Fine-Tune Your

Business Intelligence Platform

Public
Speakers

Las Vegas, Oct 19 - 23 Barcelona, Nov 10 - 12

 Harjeet Judge  Henry Banks

© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 2

Disclaimer

This presentation outlines our general product direction and should not be relied on in making a
purchase decision. This presentation is not subject to your license agreement or any other agreement
with SAP. SAP has no obligation to pursue any course of business outlined in this presentation or to
develop or release any functionality mentioned in this presentation. This presentation and SAP's
strategy and possible future developments are subject to change and may be changed by SAP at any
time for any reason without notice. This document is provided without a warranty of any kind, either
express or implied, including but not limited to, the implied warranties of merchantability, fitness for a
particular purpose, or non-infringement. SAP assumes no responsibility for errors or omissions in this
document, except if such damages were caused by SAP intentionally or grossly negligent.

© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 3

Agenda

Overview Processing Tier

 BI Platform conceptual tiers  Configuring the processing servers
Platform  Fine-tuning the details
 Infrastructure & Architecture & Sizing Clients Tools & Data Access
 Hardware: Physical & Virtual  Usage & Efficiency of reporting
Web Tier  Semantic Layer & Federation
 Splitting, Sizing & distributing Web-components  Supportability & Monitoring Tools
Intelligence Tier
 Central Management Server & System Database(s)
 Disk Tuning & File Repository Servers

© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 4

Overview

Public
BI Platform Conceptual Tiers – process flows

© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 6

Platform considerations

Public
Infrastructure requirements
Architecting BI systems

Sizing & Landscape

Demands of BI applications means bursty & dynamic allocations that hit processor & memory.
 Do you have enough CPU power?
 Do you have enough RAM Memory? Be generous for Java applications.
– Consult the Sizing Guides and professional services. Revisit your Sizing frequently!

 Are your server processes logically distributed across nodes?

– Be prepared to scale-out (extend) your systems, & cluster across new servers.
 Design your system include load-balancing & fail-over tolerances:
– Have extra services on each node, and duplicate again across other hosts.

 Benefits of fewer, larger machines (good for processor & memory management)
– is offset by demands of local IO requirements & availability.
o Note that Virtualization leans towards more, smaller VMs, rather than a few very large nodes (>16 Core).

© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 8

Hardware: Physical or Virtual?

Know about best-practices for virtualized environments

 Ensure strict CPU reservations and Memory pre-allocation:
– Avoid use of shares, limits, affinity, or other artificial mechanisms to
manage capacity across hosts.

 Ensure virtualization hosts can handle the aggregate requirements

from multiple tenants:
– Hardware needs to be properly designed to cope with the IO, in
particular Disk activity & Network.
o Consider SSD & RAID for best disk speed, & multiple Network Interface
Cards.

 VMs offers advantages for HA & DR among things, but traceability

can be an issue: Are other VMs jamming the I/O paths from the
host?
– Watch out for communications timeouts due to overloaded VM
network bridge (software, as opposed to hardware).
© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 9
Web Tier

Public
Splitting, Sizing & distributing the web components
Failover, Load-balance & performance

Accelerate static content:

Add an Apache web server to the deployment
 Split the Web & Application Tiers to improve the user-experience.

Add Multiple Tomcat application servers, approximately every 500 CAL.

 Introduce software load-balancers into the topology.

 Increase -Xmx via the Tomcat console (Default 2048MB to 10GB+ per node)
 Increase java -Xmx parameter value in web.xml properties on the AppServer.
 Increase Tomcat maxThreads to 500 from default 200.

Other settings:
 Upgrade Tomcat & JVM versions (BI4.2)
 Evaluate proprietary JAVA & Tomcat .xml parameters (such as DEFLATE)
 Maximize caching opportunities at browser, proxy, & web-server levels.
© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 11
Intelligence Tier

Public
System Database(s)
Central Management Server, Auditing & Monitoring

BI is a bandwidth-heavy application
Beware: Latency from poorly provisioned CMS databases with have a
cascading effect!

 A dedicated CMS DB running on separate hardware is recommended.

 Increase the number of requested connections from default 14 to 50 threads.
 Increase the CMS „Max number of objects in Cache‟ in-line with actuals
– Cluster CMS when planning over 500 Active Concurrent users.

 Make sure the CMS DB is available to provide low latency & high throughput
 Check vendor-specific optimisations with your DBA, statistics are up to date.
– Monitor CMS response times, & check for long-running infostore queries.

© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 13

System Database(s)
Central Management Server, Auditing & Monitoring

 Configure Auditing setting to limit grown of Audit database

– Reduce “Delete Events older that (days) parameter
– Only Audit events that you require
– Monitor Audit Thread Utilization (%)
o Consistent high % usage could indicate problem with
auditing system

 Adjust Hot-Backup feature:

– Mirror requirements for “Number of Objects in CMS”
o Understand growth requirements on FRS

 Increase –maxobjectsincache CMS command line

– from default 250000 esp. nearer 1M
 Review <CMSname>.cms registry settings for:
– MaximumObjectsToKeepInMemory &
– SizeOfLargestObjectAllowedInCacheInBytes

© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 14

Disk Tuning the File Repository Servers

An underperforming FRS can starve the entire BI system

• I/O Bottlenecks on the disks & network cause severe effects.
– Your file-system should be tuned for „Write‟ & „Read-after-write‟
operations.
– Plan for capacity increases due to hot-backup & scheduling output.

Ensure good performance of inter-node communication links,

and use quality SAN storage.
 For FRS, run checks on NAS connectivity (vendor-specific).
 Use local, fast storage for WebI Cache & Temp-file directories.
– If response times are too low (<=10 ms) your environment will
slow.

© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 15

Processing Tiers

Public
Configuring the processing servers

Each server can happily consume 100GB of RAM

 Consult the Sizing Companion Guide! Distribute additional processing & job servers across nodes:
– WebI, Visualization, Data Federation, Scheduling, Promotion, Monitoring, Auditing etc. (25+ APS types!)
– Remember that Add-Ons for Lumira, Design Studio, Analysis Office etc. require significant additional sizing!

 Split the Adaptive Processing Servers (APS) using the Configuration Wizard templates
– Adjust the application settings (i.e. capturing of details, events, frequency) for Auditing, Monitoring &
Platform Search.

 Refine configuration using APS Best Practices, to make further adjustments to the heap size (-Xmx cmd line)
– Do this for all Java servers (DSL_Bridge etc. up to 30GB each) & CR/Dashboard child sub-processes.
– WebI “memory analysis” thresholds: Adjust Upper & Maximum properties.
o Approx. 1 WPS per CPU works, allow maybe 30GB per WPS.
o Increase Binary Max Stream & Character stream values etc.
Evaluate alternative Garbage Collection parameters in APS command line properties (parallelGC, newGC).

© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 17

Fine-tuning the processing details

 Co-locate the Processing Servers with the data-source in the same network
 Schedule reports to cached instances, rather than refresh on-demand.
– Use Server groups to distribute schedules.
– Use Monitoring probes to watch system metrics & alert administrators in real-
time.

– Adjust Idle document & connection timeouts, as required.

– Increase Heap values in ConnectionServer.cfg & MDAS.properties.
– Adjust MaxLovSize parameters in client descriptor.xml files.
– Consider use of LOVChunk & ResultChunkSIZE settings for BICS.

© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 18

Clients Tools & Data Access

Public
Clients Tools & Semantic Layer

Careful usage & efficiency of reporting

Complex query operators can have larger memory footprints: Be careful
when using Unions, Cartesian products, Full-scans, Grouping Sets etc.
Good query design includes:
 Use query-stripping, drill, and guided navigation. Look to refactor multi-query
documents.
 Use pre-defined and dynamic filtering, with default selection-options for variable
and parameters inputs.
 Avoid 100s of heavy client-side calculations (e.g. Nested Sections with
Conditions).
 Avoid very wide tables (50-200 objects) & data-dumping queries (500K-1M rows).
Activate Caches.
 Evaluate viewer & editor preferences: HTML vs. Java. (Adjust memory heap
settings using JRE console + WebI tips & tricks)
 For big performance gains, keep your BW & HANA updated to coincide with
enhancement in BI releases, and cache & statistics updated!

© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 20

Data Access, Semantic Layer & Federation

Consider efficiency & proximity of reporting Database &

connected network speed
 Ensure good relational Universe design & revisit:
– Index awareness (on keys rather than labels)
– Aggregate awareness (on summary tables)
– Delegated aggregation (measure projection)
– Use of Shortcut joins & views
– Tune Multisource universes & optimise Federation join-engine using DFAT

 Edit connection properties to:

– Adjust Thread pool & Array size & tweak advanced query parameters
(BEGIN_SQL, ConnectInit, JOIN_BY_SQL,
MAX_CONCURRENT_MEMORY_CONSUMING)
– Define Safety-belts & timeouts to avoid data-dumping
– Configure hierarchy depth, members selection, exclude unassigned nodes.
– Avoid too many Details or navigational attributes

© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 21

Supportability & Monitoring Tools

Tools
• BIPlatform CMC Monitoring & OSCol (SMDAgent)
• Platform support tool 2.0 & E2E traces with client plugin
• Wily Introscope & Solution Manager
• ST0x transactions & BW BIPTools

© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 22

Supportability & Monitoring Tools

Utilities
• GLFViewer
• HTTPWatch
• Jmeter, JVMMon
• WebIAdminTool

© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 23

Thank you
Contact information:

Henry Banks & Harjeet Judge

Analytics Product Management RIG
Products & Innovation @ SAP

© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 24

Feedback
Please complete your session evaluation for
BA202

©©2015
2015
SAP
SAP
SESE
oror
anan
SAP
SAP
affiliate
affiliate
company.
company.
AllAll
rights
rights
reserved.
reserved. Public 2525
SAP TechEd Online
Continue your SAP TechEd education after the event!

 Access replays of keynotes, Demo Jam, SAP TechEd live interviews, select lecture sessions, and more!
 Hands-on replays

http://sapteched.com/online

© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 26

Further information

SAP Public Web

scn.sap.com
www.sap.com
SAP Education and Certification Opportunities
www.sap.com/education
Watch SAP TechEd Online
www.sapteched.com/online

© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 27

© 2015 SAP SE or an SAP affiliate company. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or for any purpose without the express permission of SAP SE or an SAP affiliate company.

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. Please see http://global12.sap.com/corporate-en/legal/copyright/index.epx for additional trademark information and notices.

Some software products marketed by SAP SE and its distributors contain proprietary software components of other software vendors.

National product specifications may vary.

These materials are provided by SAP SE or an SAP affiliate company for informational purposes only, without representation or warranty of any kind, and SAP SE or its
affiliated companies shall not be liable for errors or omissions with respect to the materials. The only warranties for SAP SE or SAP affiliate company products and
services are those that are set forth in the express warranty statements accompanying such products and services, if any. Nothing herein should be construed as
constituting an additional warranty.

In particular, SAP SE or its affiliated companies have no obligation to pursue any course of business outlined in this document or any related presentation, or to develop
or release any functionality mentioned therein. This document, or any related presentation, and SAP SE‟s or its affiliated companies‟ strategy and possible future
developments, products, and/or platform directions and functionality are all subject to change and may be changed by SAP SE or its affiliated companies at any time
for any reason without notice. The information in this document is not a commitment, promise, or legal obligation to deliver any material, code, or functionality. All forward-
looking statements are subject to various risks and uncertainties that could cause actual results to differ materially from expectations. Readers are cautioned not to place
undue reliance on these forward-looking statements, which speak only as of their dates, and they should not be relied upon in making purchasing decisions.

© 2015 SAP SE or an SAP affiliate company. All rights reserved. Public 28