PLSContents Who should use this document? .. What You Need to KnoW.snnnnnninnnnnnnnnnttnennnnai 6 Get the Latest Version of this Document wninnnnnnntnnnnnnai 6 Pre-Sizing Checklist .s.ssnmnnnnnnnnnnsnannnnn nnnnnnnnnnnnnnnnnnnnnssnnnase 7 Post-Sizing Checklist. seve vn seen B B14 Machine Requirements. Introduction Components and Platforms in Scope.. Disclaimer... Co sen sen seen 10 Sizing T0OS nnn nnssnmnnnnnnnnnnnannnnn snnnnnnneseesessesnnnnne LL Bid Resource Usage Estimator. BI4.1 Configuration Wizard... B14 Architecture User Class Definitions wn. Users, Active Users, Active Concurrent User$ wnnnns Conceptual tiers and services. enn 1 Deployment and Sizing Methodology. Sizing Steps and Methodology. Processing and Memory Requirements Calculations Prerequisites nics Step 1: Resource Usage Estimator SetUB nnn Step 2: Intelligence 08 Tier a sen one 18 Step 3: Intelligence Tier. 18 step 4: Application Tier... 18 Step 5: Processing Tier... 19 Analysis OLAP ssnssnn Crystal Reports 2011 Crystal Reports for Enterprise... Dashboard Design. Web Intelligence Lifecycle Management Service: Platform Search Services, Data Federation Service... 26 2 © SAP AG ~ Business Intelligence 4 Sizing GuideAdaptive Connectivity Service son. snnenenernnnnns 26 Step 6: Sizing Analysis and Seale-out.. semnnancnnnenine 27 Step 7: APS Configuration 28 APS Service Groupings 28 Basic APS Configuration Instructions emu se se 28 Adaptive Job Server sou nnmnnnninnininninininnnininnnnninnnmnncnmminennnnanesnnnsnes 30 Step 8: Deployment and Monitoring .nisnmnnmninnininininninninnnninnnnnennmennnnenne BL Scaling-out B14 Servers. a a semen 31 Building the System .... on Monitoring Bld nn ses nnn seve 32 Using Audit Reports wun snnnnensennnnne 3S Monitoring CPU and Memory Usage. ssn sone 36 BI Tool Simulation Workflows. Suggested Service Settings and Limits... Central Management Service (CMS).. Crystal Reports Cache Service nn Dashboard Cache Service, Dashboard Processing Service... Crystal Reports Processing Service File Repository Service (FRS)... Web Intelligence (Webl) Processing Service, DSL Bridge Service. Connection Server.. AL Web Application Server Scale-out Memory Expectations Sizing Example. Scheduling and Publishing... for Performance Scheduling and Publishing Best Practi Use Dedicated Machines ... YO sannnnisrsnnnnn ssinnnnnnnnsnnnnnnnnnnnannnnnnnansesnnnnne D2 Service Tuning ns serene ronnnne evn SB Java Garbage Collection... 58 Setting an APS Garbage Collector .rsnnssnnn nnn seen SB Setting the Tomcat Garbage Collector on Windows..nnsennnnmnnnnnnmannnnnnnsesnnnnnn SA Setting the Tomeat Garbage Collector on UNIX and LIAUK win 3 © SAP AG ~ Business Intelligence 4 Sizing GuideMinimum Memory Settings. snmnnennnnnnen SS Web Intelligence Tuning... sone 56 Visualization Service .. 56 Web Application Server... 56 Crystal Reports for Enterprise Tuning snr nnn seen ST Dashboards Tuning sn. nnnnnannnnnnnnnnnnnnn SS Analysis OLAP Tuning. .nn ‘SAP BW Considerations and Recommendations... seen seems 61 BW Configuration. 81 Limiting Query Result Size. sn sn seve BEx Query Performance sneneernesnnnnne Navigational Attributes. ssn sons BW Specific Tuning and Configuration Set Specific Properties Activate OLAP Cache and Delta Cache... Additional Content suns sess ee) CMS Database Tuning wnnnn Virtualization: Service Level and Performance wn a nnn 65 Desktop Virtualization. 66 Sizing SAP BusinessObjects Live Office... 87 Sizing SAP BusinessObjects Explorer.. = 68 Sizing SAP BusinessObjects Mobile... 69 Sizing SAP BusinessObjects Design Studio nu se ene 68 Appendix... 70 Bid P&R Benchmark Testing Details. 70 Test report documents and data characteristics 70 Report data description... 73 CMS repository data description venues nnn evens Test hardware specifications ...unmnnnninnnnnnnnnensnnnnnnannnnnnnninnnnnrsesnnnnns TA Document Update History .unninniinnnnnnnneiinnninnnnannnnunnnninnnnssnnnnnnasssn 7S 4 © SAP AG ~ Business Intelligence 4 Sizing Guide© Copyright 2023, 2014 SAP AG. 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 AG. The information contained herein may be changed without prior notice. Some software products marketed by SAP AG and its distributors contain proprietary software components of ther software vendors. Microsoft, Windows, Outlook, and PowerPoint are registered trademarks of Microsoft Corporation, 18M, DB2, DB2 Universal Database, 05/2, Parallel Sysplex, MVS/ESA, AIK, S/380, AS/400, 08/380, (05/200, iSeries, pSeries, xSeries, zSeries, /05, AFP Intelligent Miner, WebSphere, Netfinity, Twoll, and Informix are trademarks or registered trademarks of IBM. Corporation in the United States andjor other counties Oracle isa ragistered trademark of Oracle Corporation. UNIX, X/Open, OSF/1, and Motif are registered trademarks of the Open Group, Citi, ICA, Program Neighborhood, MetaFrame, WinFrame, VideoFrame, and MuliWin are trademarks or registered trademarks of Citrix Systems, Inc HTML, XML, XHTML and W3C are trademarks or registered trademarks of W3C*, World Wide Web Consortium, Massachusetts Institute of Technology. Java is a registered trademark of Sun Microsystems, Inc JavaScrit is a registered trademark of Sun Microsystems, Inc, used under license for technology invented and implemented by Netscape. MaxDB is a trademark of MySQL AB, Sweden. SAP, R/3, mySAP, mySAP.com, xAPpS, xAPP, SAP NetWeaver, SAP BusinessObjects and other SAP products and services mentioned herein as well as their These materials are subject to change without notice. These matorials are provided by SAP AG and its affliated companies ("SAP Group") for informational purposes only, without representation or warranty of any kind, and SAP Group shall not be lable for errors or omissions with respect to the materials. The only warranties for SAP Group 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 Disclaimer Some components ofthis product are based on Java™. [Any code change in these components may cause unpredictable and severe malfunctions and is therefore expressively prohibited, as is any decompilation of these components. SAP Library document classification: CUSTOMERS & PARTNERS Documentation in the SAP Service Marketplace ‘You can find this documentation at the folowing address: bnto://service sap com/sizin respective logos are trademarks or registered trademarks of SAP AG in Germany and in several other countries all over the world. ll ether product and service names mentioned are the trademarks of their respective companies. Data contained in this decument serves informational purposes only National product specifications may vary. 5 © SAP AG ~ Business Intelligence 4 Sizing Guide ayWho should use this document? This document provides recommendations and best practices to help with deploying and scaling the various services in SAP BusinessObjects Bl 4. We recommend consulting this guide if you are Working on a Sizing strategy, especially in conjunction with the Bid Resource Usage Estimator. Be sure to use the most recent version of the Resource Usage Estimator. «Planning to deploy SAP BusinessObjects BI 4 software Optimizing or tuning an existing SAP BusinessObjects Bl 4 deployment For all deployments, itis necessary to consult ¢ Sizing Expert in order to validate sizing exercises after initial planning using the Resource Usage Estimator and this Sizing Guide. Sometimes customers, partners, or consultants develop their own applications, strongly modify SAP's out-of-the-box solutions, or implement complex integrations involving multiple systems. In these cases, when sizing for the largest deployments or when your sizing requirements need to be extremely precise, you should consider an “Expert Sizing”. An Expert Sizing is @ hands-on exercise performed by or in collaboration with an SAP Professional, where customer-specific data is analyzed and used to achieve a sizing result with greater precision. The main objective is to determine the resource consumption of customized content, applications, and usage patterns by taking comprehensive measurements. For more information, see htto://service.sap.com/sizing -> Sizing Guidelines -> General Sizing Procedures > Expert Sizing, What You Need to Know To Size your deployment effectively and accurately, you need to know the following, which are explained further in this document: The SAPS rating of your hardware. This is the performance rating of the hardware you are going to use for your deployment, You also need to determine the number of actual* cores in the system and from that determine the SAPS-per-core value. «Ifyou are using virtualized hardware, you need to know if the IT department will allow CPU and memory reservations. Without reservations, the SAPS number will not be accurate and your sizing calculations will be incorrect + The number of users of each 8! Tool that you intend to use. Within that context, the types of each user (consumer, business user, expert). This is explained further in this document. 4 The relative document sizes for each BI Tool. * The data sources the BI Tools will access ** "Actual’ cores refer to ‘full’ cores in a CPU. Hyger-threaded cores should not be counted for the purposes of sizing, Get the Latest Version of this Document You can get the latest version of this document an www.sap.com/bisizing. 6 © SAP AG ~ Business Intelligence 4 Sizing GuidePre-Sizing Checklist These are the things you need to do before you start your sizing exercise: [How many Active Concurrent users will you have for each Bl tool? See “Users, Active Users, Active Concurrent Users” in this document for more information. [What types of users will be using each tool? Information Consumers (low workload), Business Users (medium workload) or Experts (high workload)? See “User Class Definitions” in this document for more information. HZ Doyou know how users will use each BI tool? Will they refresh documents rapidly? Will they schedule them at off-peak time? Will they open three documents at once? Will they build dashboards that consume multiple reports? How users use the tools affects the workload they produce. What types of machines will you be deploying to? What is the SAPS rating of each machine? See “Bld Machine Requirements" in this document for more information. What types of data sources will your users access? See “Data Sources” in this document for more information. EZ Will you be deploying to a virtualized environment? See “Virtualization: Service Level and Performance” in this document and wws.sap.com/bivirtualization for more information. Donat assume you are finished sizing if you only fill in the Resource Usage Estimator or the Quicksizer. These tools do not cover all deployment scenarios nor all data sources. You must follow the Deployment and Sizing Methodology section below to accurately size your deployment. 7 © SAP AG ~ Business Intelligence 4 Sizing GuidePost-Sizing Checklist Here are the things you should consider once you have followed the sizing steps in this guide: HZ Doa sanity check on your deployment landscape. Are there too many I/O-intensive services running on the same machine? Is it configured to allow for peak usage? L.e., do you think it will run comfortably (around 65% average utilization) so that peak times don't cause issues? Are the other parts of your IT infrastructure ready and sized for the load the Bl system will place oon them? Has your BW system been resized and patched? Are the relational databases tuned and ready for the load? Check the BI Pattern Books for best practices in how to configure your deployment. Check www.sapbusinessobjectsbi.com for the latest information on Bld, especi area if that applies to you. ly the Upgrade To simulate user load, check out the Performance Testing SAP BusinessObjects BI Platform Web site. HZ —_Tolearn more about getting the most out of your infrastructure, see the resources available at To ensure good performance in a virtualized environment, see “Virtualization: Service Level and Performance” in this document and wwsv.sap.com/bivirtualization for more information. Install auditing reports so you can monitor your system. A sot of auditing raports is available for download at http://scn.sap.com/docs/DOC-6175, Schedule a re-sizing in six months. See the Resizing section in this document for more information. 8 © SAP AG ~ Business Intelligence 4 Sizing GuideBI4 Machine Requirements For Bi, the minimum hardware for evaluation purposes is 8000 SAPS of processing power and 16 GB of memory. This is defined in the BI 4.1 PAM, found here: http:///scn.sap.com/docs/DOC-41355. Hardware requirements for deployments of any kind are determined by following the remainder of this guide. Introduction This document discusses the Sizing of the Business Intelligence 4 suite of services, hereafter referred to a5 BI4, The Bid suite is designed to serve customers from small business to SAP's largest customers, The umber of users, types of users, usage patterns, number of BI tools included in the suite, number of data sources supported by the suite as well as the deployment options supported by the suite ll factor into a series of variables that affect the successful deployment of the Bl suite. No configuration fits all customers. The purpose of this document is to help guide you through the Sizing Exercise for BI4, Sizing Bl is very different compared to sizing of other types of Enterprise software. Bl is a very resource Intensive task. The act of extracting information from a potentially large amount of data requires adequate amounts of processing power and exercises all the important subsystems of a computer: CPU, memory, disk and network, Having the right amount of capacity of these four aspects of your system is key to success. BI can also be very bursty, since the load can rely a lot on the schedule of its users. ABI system is used much like Google is: searching for information interactively. If Google required 2 number of seconds to respond, most users would consider “the Internet broken”. They have an expectation and often a business need to have a responsive system available to them. The purpose of, this document is to help ensure your users can access BI decision-making information in a timely, No tool or decumant can replace human judgment. So while this document attempts to cover as many aspects of the Sizing Exercise as possible, Sizing Experts at SAP should always be consulted. Make sure you have the latest version of this document and the other Bld Sizing resources by visiting www.sap.com/bisizing Components and Platforms in Scope The guidance and estimates referenced in this document are based on the following: © Analysis OLAP © Crystal Reports 2011 + Crystal Reports for Enterprise * Dashboard Designer + Web intelligence 9 © SAP AG ~ Business Intelligence 4 Sizing GuideGeneral sizing tips are also provided for: © Explorer © Mobile 81 «Live Office «B14 Platform services: Life Cycle Management, Search, Data Federation «SAP BW It is assumed that readers of this document are already familiar with core concepts from the Business Intelligence Platform Administrator Guide (Admin Guide). There you can find conceptual information and technical details on wide number of topics which are intentionally not covered in this guide. Disclaimer This document demonstrates how someone might perform sizing of a BI4 system. The methodology and sample walkthroughs offered here are examples of the tasks and thinking involved. The performance and functioning of an actual system may vary for many reasons. The examples offered here should not be considered a guarantee of success of a particular deployment. You should work with your SAP account representative and ensure you get the advice of a Sizing Expart when making final decisions with regards to a deployment of Bi4. Resizing Its recommended that you resize your environment every six months or sooner if you have had significant changes in the scope of your deployment. It can be difficult to anticipate how yaur users will make use of the BI environment at the outset ofa project. Resizing is recommended to ensure your systems are configured for how your users are using the system, In order to be informed accurately about the usage of the system, it is recommended that you take advantage of the auditing features end auditing reports available for BusinessObjects BI4. A set of auditing reports is available for download at http://scn.sap.com/docs/DOC-6175 . These reports will help you determine how your BI services are being use Monitoring and probe reports are also recommended. Probe reports can give you alerts when time to run expectations are longer than expected. See the BI4 Admin Guide for more information on creating probes. CMS database performance should also be monitored. Monitoring and recording of the basic CPU and memory usage of the machines in your deployment is. also recommended. The goal is to have an average CPU usage of 60% to 65% in order to handle peaks in the range of 80%. Analyzing the historical usage of your deployment can help you determine if more resources are needed for a particular node. Similarly, if memory is frequently being fully used, you may be experiencing reduced performance and need to add more. 10 © SAP AG ~ Business Intelligence 4 Sizing GuideSizing Tools BI4 Resource Usage Estimator The sizing tool reference in this document is the BI4 Resource Usage Estimator. It was previously named the Bid Sizing Estimator. The estimates it provides are based on Performance and Reliability (P&R) sizing benchmark tasts performed in SAP's test labs. The results and this tool help to estimate load based on performance results in these lab tests. They are only estimates and should not be used as simple predictions of performance or deployment guidance for your project. The Resource Usage Estimator is just a tool. It and this guide must be used together to provide effective sizing guidance. BI4.1 Configuration Wizard Starting with SAP BusinessObjects 814.1, the Configuration Wizard helps the sizing exercise by defining a number of APSes based on the deployment size chosen. This document explains how to perform a custom sizing exercise, The Configuration Wizard's APSes can be used as a time saver when it comes, time to define the APSes for your deployment. For more information, see the Bld.1. Admin Guide, Br © SAP AG ~ Business Intelligence 4 Sizing GuideBI 4 Architecture ‘The Bld architecture can be a complex set of services and technologies. The following is a representation of the system from a technology perspective: For the purposes of sizing Bld, the essential aspects of the BI4 architecture are this: ‘The Bid Platform: This is a set of base services that enable BI services to run in a distributed deployment. Each node in a Bid cluster runs a Server Intelligence Agent (SIA) which runs selected services on that node. Nodes in a cluster communicate over a service bus. ‘The Bid Platform and suite of services are designed using a 64-bit architecture to ensure scalability to the largest customers. ‘The CMS: This is where all the objects and configuration information that control and secure the platform are located. The CMS is a combination of service plus database for storing and retrieving this information. Basic Services: Core services such as the File Repository (FRS) need to run somewhere in the cluster. APS: The Application Processing Service (APS) is a service host. 4 large number of the BI services run inside APSes, There can be more than one APS running on a given node, and there can be many nodes running APSes in a Bl landscape. Services running in the APS are Java-based and the APS controls the JVM that the services run in, There is one JVM per APS. R © SAP AG ~ Business Intelligence 4 Sizing Guide ayNon-APS Services: Not all Bl services run inside an APS. Those that don't are run separately by the SIA. BI Servcies: The Bl services in the system have specific architectural considerations: Crystal Reports: Crystal Reports services are available in two versions: Crystal Reports for Enterprise is designed to have great Universe and SAP connectivity. itis a newer version of Crystal Reports. It uses data access technology common to the other tools in the suite. Crystal Reports 2011 is the original Crystal Reports product line, maintained and available for backward compatibility. Web intelligence: Webl provides analysis and reporting capabilities to the suite. It provides compatibility with old Universes (UNV), new Universes (UNX) as well as SAP BW. The Visualization Service (aka CVOM) provides data visualizations (charting, graphics) services to Web and the DSL and DSL Bridge house the data services for Webl. User Class Definitions Three user classes are defined below. Class definitions are used to identify how much the system will be used and how much load will be placed on it. Information Consumers use the system a little and Expert Users use the system a lot. Knowing the mix of user classes is important since it directly affects the performance and resources required by the system. Be aware that customers often underestimate the amount of use that a user will make of Bl system, Information Consumers - The least active of all the user types. Information consumers spend an average of 300 seconds (5 minutes) idle in between navigation steps. These users typically view predefined and static content and perform relatively litte drilling and filtering on their own, Business Users - These users perform some moderate amount of drilling and filtering on their own, Business users spend an average of 30 seconds idle in between navigation steps. Expert Users - The most active of all the user types. Expert users spend an average of 10 seconds idle in between navigation steps. These users are much more likely to perform resource-intensive operations in the system including ad-hoc analysis and customization of reports, retrieving a large number of rows, and heavy client-side filtering. Users, Active Users, Active Concurrent Users Many sizing exercises begin with a total number of users, for example: the customer needs a system sized for 10,000 users. The next step is to determine how many of those users will be logged-in to the system at the same time. The bast way is to work with the customer to understand their usage patterns. IF customer can’t provide guidance on this number, it’s common to estimate 10% as 2 minimum. We call these active users. In this example, there would be 1,000 active users. Because users can spend much of thelr time being idle after logging-in, you also need to determine how many of the active users in the systam will be concurrently generating load. This is another opportunity to take customer usage patterns into account, but if there is no information available on this, the most common 3 © SAP AG ~ Business Intelligence 4 Sizing Guide ishestimate is 10% as a minimum, We call these users who are logged-in and concurrently generating load, active concurrent users. In our example, there would be 100 active concurrent users. 0 When you use the B14 sizing tools, specify your inputs in terms of Active Concurrent Users. In the example above, we went from 10,000 users to 100 active concurrent users, which translate to 1% concurrency. It's important to understand that the typical concurrency recommendations vary depending on the size of deployment and other factors including the kind of work users perform. If you expect or experience typical concurrency higher than these nominal ratios, you should expect a heavier load and should compensate accordingly. For guidance in such a case, you may want to use the Resource Usage Estimator and specify more users as ‘expert users’ to account for the increased use of the system. Conceptual tiers and services The SAP Bi platform can be thought of as a set of conceptual tiers. A subset of each list of services is provided here for convenience. Refer to the BI4 Admin Guide for the complete list. Application Tier © Web Application Service (typically Tomcat) © CMC~Central Management Console © Bilaunch pad Intelligence Tier (Referred to in the Admin Guide as “Management Tier”) «vse + Crystal Reports Cache Service * Dashboard Cache Service + FRS Processing Tier + Adaptive Job Service + Adaptive Processing Service (APS) © Analysis OLAP and Multi-Dimensional Analysis Service (MDAS) © DSL Bridge Service © Visualization Service (VOM) © Crystal Reports Processing Service © Web Intelligence Processing Service *1 The database which the OMS uses for its repository may sometimes be referred to as the “ineligence Tier DB" 14 © SAP AG ~ Business Intelligence 4 Sizing GuideDeployment and Sizing Methodology Sizing a BI4 deployment requires a reasonable degree of planning so that calculations and predictions can be made about the needs of the system. The number of users and the needs of those users can be used to predict load on the system. The types of data sources used have an effect on the load and needs of the system. The needs of the users include the Bl services that they need to use. Some users use the services a lt, some use them lt, Some reports are scheduled to be processed at night and viewed during the day, Some reports need to be refreshed when viewed, which causes more load as the number of users increases Once the user requirements are defined, the system can be defined that will achieve the required amount of processing. The final step in the sizing is to apply it to the hardware landscape. The deployment hardware can range from many small machines to one large machine. The Sizing Exercise includes the allocation of BI services to the nodes in the system, taking into account the CPU, memory, disk and network capabilities of nodes to be used in the construction of the system landscape. Sizing Steps and Methodology The basic approach to sizing is working through the requirements of all the tiers in the deployment, accounting for the processing and memory requirements of each subsystem. Most subsystems and services have special considerations that need to be taken into account Once the processing and memory requirements are determined, the next major step is to fit the services to the hardware landscape. This requires knowing the SAPS per core rating of the computer to be used as well as the amount of memory available to the machine or machines, You must determine the SAPS rating of the machines that will make up your deployment. You cannot assume any SAPS rating from this or any other document. You can determine the SAPS rating of machines by visiting the Web site: SAP SD Standard Application Benchmark Results The Bid Resource Usage Estimator is used as a tool to help calculate the processing load to be handled by each tier in your system. The Estimator is used to get initial numbers for processing and memory needs. Depending on the BI tool under consideration, additional calculations may need to be made with corresponding changes and/or additions to the sizing. Processing and Memory Requirements Calculations For each tier, especially the Processing Tier, the processing power and memory requirements need to be determined. This is the core of the Siting Exercise. Once these numbers are determined, scale-out to deployment hardware can be done. The processing power is calculated and specified in a processing unit called SAPS. This is a processor and computer independent unit that describes the processing power of a CPU in a given computer. SAPS is an SAP-derived processing rating that allows you to calculate the power required by a deployment. it a5 © SAP AG ~ Business Intelligence 4 Sizing Guide ishtakes into account the I/O processing capabilities of the computer in addition to its raw computation abilities. Computers can have different SAPS processing capabilities even with the same type of CPU, Prerequisites The goal of the sizing exercise is to calculate the peak load that will be placed on the system. In order to proceed with the sizing steps, you need to know the following information listed here. Gathering this data accurately is the most important part of the sizing exercise since all the sizing calculations derive from this information Users: How many Information Consumers, Business Users and Expert Users of each type of BI tool. An average user workflow is also important to know. If you expect users to open five BI documents and refresh them all at the same time, that’s five times the load of one user. That needs to be accounted for since the system needs to handle that load. It is very important to know if the common workflow of the users is going to include refreshing reports and if so, how frequently. Will the reports be scheduled to run at night and only be viewed during the day or will they be refreshed by users avery hour? That's an important part of the load prediction and thus the sizing estimate. Data Sources: What types of data sources will be used: direct-access SOL databases, UNV Universes, UNKX Universes, SAP BW, BW on HANA or HANA direct? It can be a mix. It’s important to know which types of data sources will be used for the majority of Bl processing so that peak load can be predicted. Some customers expect their mix of data sources to change over time, This is an important thing to conside Document Size: The relative size of each document is important to know. Will most documents be small and cause little impact on the system or will most be large and require a lot of processing? This should bbe determined for each Bl tool expected to be used. SAPS Rating: In order to know how much computing resources are needed for a deployment, you need to know the computer's performance rating. This is measured in SAPS, You may be building the system Using one computer or many. You may have the computer(s) specified in advance or not until the initial sizing has been calculated. The SAPS rating of the system's intended hardware allows you to determine the required amount of hardware to meet the system's processing requirements. The SAPS rating takes into account CPU processing power as well as the computer's I/O capabilities. To learn the exact amount of SAPS produced by your hardware, check the SAP SD Standard Application Benchmark Results Web site For the sizing exercise, the amount of SAPS per core needs to be determined. The benchmark shows the SAPS rating for the computer. You then need to divide by the number of actual cores of the CPU. Note: Hyper-threaded cores should not be used for this calculation, 16 © SAP AG ~ Business Intelligence 4 Sizing GuideStep 1: Resource Usage Estimator Setup ‘The Resource Usage Estimator should be initialized with the number of users of each type for each type of BI tool. The report size sliders should also be set to reflect the sizes and types of documents that will be processed by the system. It might look like this: ee ae oe En ru (sars ia) Memory (6B (91) ] TT —— Restore Systern Defaults Set Configuration Template Note: Analysis OLAP is also referred to as the ‘Multi Dimensional Analysis Service’ or MDAS. See the Bl4 Admin Guide for more information on changing Analysis OLAP settings. ‘Din your sizing document, add in the SAPS and memory required for Analysis OLAP. as © SAP AG ~ Business Intelligence 4 Sizing GuideCrystal Reports 2011 To size Crystal Reports 2011, set the user numbers appropriately and set the user numbers for all other tools to zero, rocesna Ter SST eo TT If your deployment will be to one large machine, you can record these requirements in your sizing document and move to the next tool. Proceed here if you will be scaling out your deployment. The Crystal Reports 2011 services required resources are determined by the processing requirements. The required memory is calculated as follows: Number of CPU cores x 1.25 GB/core The number of cores is determined by the SAPS shown in the Resource Usage Estimator for the Processing Tier (while having set only the Crystal Reports 2011 user numbers). To determine the required CPU coras, divide the SAPS needed by your SAPS per core calculation (sae the Prerequisites section above for more information). You should round up when you encounter fractional cores. Once you know the number of cores needed, you can calculate the memory requirement by multiplying by 1.25 GB per core. ‘Din your sizing document, add in the SAPS and memory for Crystal Reports 2011, Crystal Reports for Enterprise To size Crystal Reports for Enterprise, set the user numbers appropriately and set the user numbers for all other tools to zero. TT a The Crystal Reports for Enterprise services required resources are determined by the processing requirements. The required memory is calculated as follows: Number of CPU cores x 1.5 GB8/core The number of cores is determined by the SAPS shown in the Resource Usage Estimator for the Processing Tier (while having set only the Crystal Reports for Enterprise user numbers). To determine the required CPU cores, divide the SAPS needed by your SAPS per core calculation (see the Prerequisites section above for more information). You should round up when you encounter fractional cores. 20 © SAP AG ~ Business Intelligence 4 Sizing Guide(Once you know the number of cores needed, you can calculate the memory requirement. For example, 20 cores of processing power requires 20 * 1.5 = 30 GB of memory. Din your sizing document, add in the SAPS and memory for Crystal Reports for Enterprise. Data Sources Crystal Reports for Enterprise is sized differently depending on the data sources expectad to be used. Reporting directly on relational data sources or reporting on Universes requires no additional sizing. When reporting on SAP BW including SAP BW on HANA, additional sizing is required, BW Data Source Sizing Querying report data delivered via BW requires additional processing and memory resources. Sizing is done relative to the amount of data you expect to be retrieved by an average query and also by the frequency of queries submitted to BW. Note: itis very important to have a fully patched BW system and also for your BW system to be sized and tuned to your Bl query needs. It is highly recommended that your BW system be checked and tuned by a BW expert before connecting a BI system to it. ‘An example query is 100,000 rows with 10 fields, equivalent to 1 million cells. Your data set(s) may vary in shape and size. With a 2000 SAPS per core CPU, 1 million cells take approximately 10 seconds to process and require 11GB of memory for the processing. You need to judge the processing and memory requirements of your queries accordingly, taking into account the SAPS per core rating of the computer. The processing time is linear, so 500,000 cells should take about 5 seconds to process, for example. Similariy, if your CPU is rated at 1000 SAPS, then 1 million cells will take about 20 seconds to process. (Once you know the amount of processing to be done for each query, you need to determine how often queries will be submitted. This can be new queries or report refreshes. The refresh frequency may be quite low in the case of scheduled reports, or may be quite high if a sales team needs to refresh their reports all at the same time of day. Itis important to get an accurate prediction of the peak load, since under-rating this requirement can result in significant degradation of performance as the system attempts to process multiple processing and memory-intensive operations at the same time. 21 © SAP AG ~ Business Intelligence 4 Sizing GuideDivide the number of queries expected per minute by the time is takes to process a query to get the number of cores needed to process BW queries. 60s per minute jueries per core per minute =———S P&T ES _ v Pp Pp Single query process time (5) required queries per minute cores required = 4 TS PO 4 queries per core per minute For example, to support 30 queries per minute where the query returns 1 million cells: cores required Scores To size the memory requirement, multiply the expected query size by 1000 and round up to the nearest GB, This amount of memory will be needed for each core to process the BW queries. For example, 1,000,000 cells per query x 1000 = approximately 1GB per core, > In your sizing document, add in the SAPS and memory required for BW query processing. JDBC Considerations For medium to high usage Universe scenarios that make use of JDBC connectivity, including HANA JDBC connectivity, additional memory should be allocated for the Java environment created by the Connection Server, You should add 2GB of memory to the node that hosts the Connection Server and change the configuration of the Connection Server to allocate that additional memory for the JVM. See the Connection Server tuning section Connection Servar on page 41 for datails. Dashboard Design To size Dashboard Design, set the user numbers appropriately and set the user numbers for all other tools to zero, rece Tar SOREL recess er Din your sizing document, add in the SAPS and memory for Dashboard Design. 22 © SAP AG ~ Business Intelligence 4 Sizing GuideWeb Intelligence To size Web Intelligence, sat the user numbers appropriately and set the user numbers for all other tools to zero, Processing Ter 2 Prcesteg ier [TT If less than 6GB of memory is shown, 6GB is recommended, Din your sizing document, add in the SAPS and memory for the Webl Processing Server. Data Visualization The Data Visualization (CVOM) service is used by Webl. This service is recommended to have a memory allocation of 268, > In your sizing document, add in 268 of memory for the Visualization service. DSL Bridge The DSL Bridge service contains the data connectivity for UNX Universes as well as SAP BW. When Web! uses either of these types of data sources, additional sizing for the DSL Bridge is necessary. See the two following sections depending on the type of data sources that are expected to be used, The DSL Bridge is an APS-hosted service. Dota Sources Web is sized differently depending on the data sources expected to be used. Reporting on UNV Universes requires no additional sizing, When reporting on UNX Universes or SAP BW including SAP BW on HANA, additional sizing is required. Apply sizing using the data source specific sections below as appropriate. UNX Universe Data Source Sizing The sizing calculation for the DSL Bridge for UNX Universes is 0.25 GB per active concurrent user, 8GB minimum, 30GB maximum. If the calculation results in more than 30G8, additional DSL Bridge service instances should be created so that no DSL Bridge instance has more than 30GB allocated to it. > In your sizing document, add in the memory for the DSL Bridge service to handle UNX Universes. Note: The processing power (SAPS) for UNX Universes is accounted for by the SAPS allocated to the Web! Processing Server. 23 © SAP AG ~ Business Intelligence 4 Sizing Guide ishBW Data Source Sizing Querying report data delivered via BW requires additional processing and memory resources. Sizing is done relative to the amount of data you expect to be retrieved by an average query and also by the frequency of queries submitted to BW. Note: itis very important to have a fully patched 8W system and also for your BW system to be sized and tuned to your BI query needs. Itis highly recommended that your BW system be checked and tuned by a BW expert before connecting a Bl system to it. An example query is 100,000 rows with 10 fields, equivalent to 4 million cells. Your data set(s) may vary in shape and size. With @ 2000 SAPS per core CPU, 1 million cells take approximately 10 seconds to process and requires 1GB of memory for the processing. You need to judge the processing and memory requirements of your queries accordingly, taking into account the SAPS per core rating of your computer. The processing time is linear, so 500,000 cells should take about 5 seconds to process, for example. Similarly, if your CPU is rated at 1000 SAPS, then 1 million cells will take about 20 seconds to process. Once you know the amount of processing to be done for each query, you need to determine how often queries will be submitted. This can be new queries or report refreshes. The refresh frequency may be quite low in the case of scheduled reports, or may be quite high if a sales team needs to refresh their reports all at the same time of day. Itis important to get an accurate prediction of the peak load, since Uunder-rating this requirement can result insignificant degradation of performance as the system attempts to process multiple processing and memory-intensive operations at the same time. Divide the number of queries expected per minute by the time is takes to process a query to get the number of cores needed to process BW queries. 60s per minute jueries per core per minute = Per minute _ v Pe Pp ‘Single query process time (s) required queries per minute cores required = queries per core per minute For example, to support 30 queries per minute where the query returns 1 million cells 30 30 cores required = 735 =2?= 5 cores To size the memory requirement, multiply the expected query size by 1000 and round up to the nearest GB, This amount of memory will be needed for each core to process the BW queries. For example, 1,000,000 cells per query x 1000 = approximately 1GB per core, 24 © SAP AG ~ Business Intelligence 4 Sizing GuideThe additional processing and memory is to be allocsted to the DSL Bridge service. In case the amount of processing and/or memary exceeds the machine on which is runs, multiple DSL Bridge instances can be created to-scale-out the load, In your sizing document, add in the SAPS and memory for BWW query processing for the DSL Bridge. When receiving and processing data from BW, the DSL Bridge requires memory te hold the query results. This has the same memory requirement as UNX Universes: 0.25 GB per active concurrent user, EGE minimum, 30GB maximum. the calculation results in more than 3068, additional DSL Bridge service instances should be created so that no DSL Bridge instance has more than 30GB allecated to it In your siting document, add in the memacy for the DSL Bridge service to buffer the BW query results DBC, Consideration For medium to high usage scenarios that make use of JOBC connectivity, including HANA JDBC connectivity, additional memery should be allocated for the Java environment ereated by the Connection Server: You should add 268 of memory to the node that hosts the Connection Server and change the configuration of the Connection Server to allocate that additional memory for the VM. See the Connection Server tuning section Connection Server on page 41 for details, Lifecycle Management Services ‘Life Cycle Management! refers to the following group of services: Visual Difference Service * Lifecycle Management ClearCase Service + Lifecycle Management Service * Trace Log Service It is recommended that these services be grouped into a LCM-specific APS for medium and larger deployments. For medium to large sized deployments, this APS should be allacated at least 750 MB of memory. For very large deployments, this APS should be allocated at least 168 of memory. For pracessing calculations, a nominal amaunt of SAPS of 1000 should be used. Pin your sizing document, add in-the SAPS and memory for the LOM Nate: For deployments prior to Bl 4.1$P3, onlyane instance of the LCM service should be run ina cluster. 25 © SAP AG — Business Intelligence 4 Siting Guide =Platform Search Services Siting of the Search Service depends on your company’s intended ute of search, Itcan be turned aff, used lightly or used heavily If you intend to use search, the amount of resources to allocate to it depends on how much and how frequent content is changed in your system. For a system where content doesn't change frequently during the day, you cen schedule search to index at night. It is recommended that a separate APS be created just far the Search Service. This will allow for easy monitoring and configuration of resources for search. For light usage of search, allocate 500 MB of memory and 1000 SAPS. For heavier use of search, allocate 3 GB of memory ang 3000 SAPS. Din your sizing document, add in the SAS and memory for the Platfarm Search Service. Data Federation Service Ifyou intend ta use multi-seuree Universes, you needa size for the Dats Federation service. It is recommended that an APS ba crested fer the purpase of running tha Date Federation, especially far large deployments, This allows the data federation services to be monitored and configured easily, For medium deployments, allocate 1 GB.of RAM. Fer larger deployments, allocate 2 GB of RAM, Din your sizing document, add in the SARS and memory for the Data Federation service, For more thoraugh siting of the Data Federation service, see the Date Federation Service Sizing end Tuning Companion Guida, located on SAP Service Market Place (SMF) Adaptive Connectivity Service For medium and larger deployments, the Adaptive Connectivity service needs sizing. For medium deployments, 500 MB of memory is recommended. For large deployments, 1 GB of memory is recommended. Itisrecommended that an APS be created for the purpase of running the Adaptive Connectivity services, ‘especially for large deployments. This allows the data federation services to be monitored and configured easily. ‘Pin your sizing document, add in the SAPS and memory for the Adaptive Connectivity service ifneeded 26 © SAP AG — Business Intelligence 4 Siting Guide =Step 6: Sizing Analysis and Scale-out Scaling-out your deployment is often necessary, where the size of the deployment is larger than a single machine that is expected te run the Bl system. Additionally, some analysis and thought is needed ta selectively place services on nodes in a cluster so that optimal perfarmance and reliability is schieved. ‘Once you have the basic sizing as calculated in the above sections, you should have a list ofall the services with SAPS and memory requirements for each. The next step is to ensure the hardware used in nods of the deployment can handia the load Befare assigning services to nodes in the system can occur, you need to determine what the constraints of the system are. There wil vary likely be processing power and memory limits on each node. Some nodes may be different, For each of the tiers: Intelligence D8 tier, Intelligence tier and Application tier and for each of the Bl tools to be used in your deplayment, do the following for each of the warkloads: 1. Start by assigning the warkload to anew node in the system, 2. Add in memory te account for the operating system and basic Gl infrastructure components such as the SIA. 2GB) of memary and 2000 SAPS are recammended for these functions. 3. Check to see if either of the processing power of the nade or the memory imit of the nade has been excesdad for the workload that has been assigned to it. fit fits, you can move onto the next workload, fit doasn’t fit, you need to add additional nodes until the full processing power and/or memory forthe workload have been accommodated. For workloads made up of APS-based servicas, an ABS specific to that workload should be crasted on the node for those services, A nade may have ather services defined on it, possibly as a result of the standard installation, Any services not intended for use an # node should be disabled. See the APS Configuration section below for more information, Incases whore more than ane workload can fit on a node, its fine to put multiala workloads on the same node, However, be sure to teke inte account expected future growth and also mixing of similar services. For exemple, Crystal Reports and Webl are bath very I/O intensive so they shouldn't be put on the same node unless you don't expact them to be used at the same time, Similarly, the Application Tier hosting the Web server can be bound by memory and processing power, which the DSL Brilge can also be — these services shouldn't be put on the same node without some thought to the usage of the system "Mote than 26B of memory may be needed depending on the hardware platform and operating system used, Consult your machine vender for more information. a © SAP AG — Business Intelligence 4 Siting Guide =Step 7: APS Configuration The Adaptive Processing Server (APS) isthe aplication that hosts a let efthe Bl services, Multiple APSes may be defined on multiple nodes within a deployment. In almost all cases mare than one APS will be needed inthe system, both for management and maintenance of the running services. Often nodes in the system need to run only a small subset of the available services inthe default APS created at install time, making at least some amount of APS configuration necessary an every machine. Configuring APSes is an escantial part of sizing Bld ln BI 4.0, one APS is defined at installation, The rest of this section shows the basics of APS configuration. APS Service Groupings When establishing your APSes, the following APS definitions are recommended nsignt to Aceon Sennce je Imengence Montonng Service "Web Intengence Wontonng Serve er Pang Pay Se ala Federalion Service ene ‘Vsti erence Sec TRGyoS Nernagrne CCE Serves Liecyee Managerneal Sanne ‘Mul Dimarsional Analysis Service 'BEK Web Application Serves 28 © SAP AG — Business Intelligence 4 Siting Guide =Note: The color-caded groupings shown above can be used when deploying APSes for redundancy and fall-over with minimal machine dependencies, The red group (right-hand column) services are related ‘the blue grcup (left-hand column) are mostly unrelated but have no dependencies on the red group. In caces where you want to duplicate services and also reduce or eliminate machine dependencies, this partitioning can be used. Basic APS Configuration Instructions It is strongly recommended that you not alter the original APS defined by the installer. If that APS is not needed, it should be left intact, stopped and disabled. How to Create a New APS + levthe CMC, go to the Services listing «On the Manage menu, choose New and then New Server Manage > Actors SB3%\/co;>- me Be BF serverstist Ts = Gece [fe ee eee © crystal Reports Services Enocee © ata Feceration Services WD service Cotman @_eenloner Services RE ween core: Properties Create Server Group | = Se 1 = On the Create New Server dislog box, choose the most appropriate Service Category Select the first of potentially many services from the Select Service list, then press Next. * Todd mare services to the APS, select them in the Available Additional Services list press >to add them to the Selected Services list: Press Next wrhen done, Give the APS a unique name. Chose Create, At this point the APS will exist but will net be enabled or started. You can enable and start your new APS: by right-clicking on the new service in the CMC and selecting the appropriate options. How to Permanently Disable an APS 29 * Inthe CMC, go to the Services listing ‘© Right-click on the server (APS) that you wish to modify * Choose Disable Server © Right-click on the server (APS) that you wish to modify * Choose Properties # Find the option: ‘Automatically start this server when the Server Intelligence Agent starts! + Unchack the eption, © SAP AG ~ Business Intelligence 4 Sizing Guide aHaw to Change the Memory Setting of an APS Inthe CMC, go to the Services listing Right-click en the server (APS) that yeu with to medify Choose Properties Locate the Command Line Parameters Serollta the right until you locate -mnxlG (er similar it may have a different number Change the number after the -Xmnx, For example to set the APS to use 6GB, change the command line ta read -Kmx6G Haw te Change the services running in an APS © Inthe CMC, go to the Services listing ight click on the server (APS) that you wish to modify © Choose Select Services See the 51d Admin Gi depth information on configuring APSes. Adaptive Job Server Note that the Adaptive lob Server does not require “splitting” or configuration as described above for the APS. This is because the Adaptive Job Server runs services as separate dedicated processes. 30 © SAP AG — Business Intelligence 4 Siting Guide aStep 8: Deployment and Menitoring [An important part of BI4 deployment is monitoring the operation of the various services tomake sure they are running at a reasonable level of utilization, For BI systems, an average utilization goal Is 60% of a node's resources. This goal is prescribed because of the bursty nature of BI, which doesn’t run in a steady state but is very influenced by user workload. During initial installation and testing, monitoring of the system is very important since the best sizing. exercise can’t fully know hour a system will behave and perform. Simulating user load is highly recommended, using products suen as Apache JMeter or HP LoadRunner, Start with a small number of simulated users and bring the workload up slowly, watching how the system reacts. Proceeding gradually will uncover any trouble spots that can be dealt with before progressing to a full warkload Itis critical to menitorall aspects af resource usage in the systern: = Disk © Network = CPU © Memory This neads to be done across the entire Bld landscape, including the CMS database and reporting databases. The monitoring services in BIS provide a subset of the needed tools ta do this. Depending an yeur monitor solution, you should enable the fellewing: 8, Proactively address Scaling-out BH Servers ‘A very important aspect af sizing that is inferred by the seanarios above is the creation and deployment of dedicated 814 servers, The goal ists have dedicated machines running just the services that are necessary for its role. This is 8 common requirement of 2 scale-out deployment. See the Bid Admin Guide for guidance on creating clustered deployments, The Sizing Guide suggests that Bld be installed as a full installation on each server and then any unnecessary services be turned off, This allows for a more simple installation procedure and also allows for future changes to the deplayment. For exemple, if the role of a machine changes, all the B14 components are thera ta be enabled and disabled as necessary Building the System It is vary important to methodically “build out” the system. 18, Start with a smaller landscape, using smaller number of users te gain confidence in your configuration. Gradually increase the load in increments af $0 to 200 users, only adding services/servers as necessary. b. Carefully monitor and analyze the performance and resource usage across the entire landscape, including the CMS repository database, the Web application tier, and any other SAP BI 4 platform servers involved in the test to identify bottlenecks in either the underiving infrastructure or server configuration 31 © SAP AG — Business Intelligence 4 Siting Guide a¢. Take the appropriate actions to resolve problem areas at you find them. For example, adding. another CMS hast if CMS CPU utilization is above 80%, Testing gradually and solving issues as, they are found is important, since progressing immediately to a full workload can make The CMS database is key to the overall performance and scalability af the Bl platform. A dedicated CMS database running on dedicated hardware is always recommended. Work with your DBA to ensure the CMS database server is correctly sized, configured, and monitored by referring to database vendor materials on sizing. The underlying infrastructure, including machines and network, is eritical to the overall performance ang scalability of the BI platform; work with your infrastructure administraters to ensure the environment is correctly sized, configured, and monitored When starting your SAP Bid landecape, it's recommended to methodically start each of your server intelligence agent (SIA) nodes and ensure all services are correctly started hefore starting anather node. Monitoring 814 The Bid suite has a number of built-in monitoring capabilities that can be used to measure the performance of the system as you build it up and continue into production, Check the Processing Server Metrics Itis recommended that you frequently check the Processing Server Metrics for the Bl processing services that your deployment uses, especially in the setun phase of your project. Metries are maintained for each processing service such as Average Processing Time, Maximum Processing Time and Minimum Processing Time. Here is an example for the Web Intelligence Processing Server: Server tame state GB _ MASTERSHAKE.AdaptvezobServer & running |B _ MASTERSHAKE.AdaptiveProcessingServer & Running G Ss Properties Start Server Restart Server Stop Server Force Termination Enable Server Disable Server 2 © SAP AG — Business Intelligence 4 Siting Guide =Some of the information shown includes: Web Intelligence Processing Service Metrics Cache size (Kb) 14764 umber of out-of-date documents in cache Cache high mark count (CPU usage (26) Total CPU ume (seconds) Memory high threshold count Memory max threshold count Virtual memory size (Mb) 2 Current number of citer calls ° Number of remote extension errors ° Current number of tasks ° Total number of client calls 0 Total number of tasks ° Idle time (seconds) 1869 Current number of active sessions ° umber of documents opened fram cache 0 umber of cocuments ° Current number of sessions © Number of document swap 0 Number of swapped documents 0 ‘umber of sessions tmeout ° Total number of sessions ° Number of users 0 Mumber of active threads 0 Total number of threads 5 ES © SAP AG — Business Intelligence 4 Sizing GuideConfigure Report Processing Time Alerts The Monitering festures of BI4 allow you te receive alerts when specific reports take too long to process. Youcan configure probes that run a Web Intelligence or Crystal Reports document periadically. You can be alerted if any of the documents take toa long to run, Monitoring Step 1: In the Monitoring area of the CMC, open the properties for the Crystal Reports Service (Processing Server) to configure the probe for Crystal Reparts. Gpen Interactive Analysis to configure the probe for Web Intelligence, You sheuld slready have documents ereated that you would like te maniter the performance of, See the Bl Admin Guide fer more information on creating and configuring probes. While the system isin testing and validation phases, report processing probes should be run frequently [every few minutes) in order to effectively monitor the performance of the system. (Once a system is placed inte production, it is suggested that you run report probes hourly so you ean be alerted if system performance degrades You can change the schedule far a repart prabe as shown here: Schedule: Crystal Reports Service (Processing Server) Fide Navigation Properoes Recurrence > Default Settings fun object: [Houry SSCS ~ Schedule ‘Object wil run every Whours and x minutes. Instance Tile Hourtn) =[1 =] minute) = [0 =] start oateTime: foa.=] [19] [Po] rtor201 3) Schedule For End Date/Time: [03 =] [19 =] [Pm =] [5/20/2023 &) History mite Number of retries allowed: (0 Retry interval in seconds: 1800 et © SAP AG — Business Intelligence 4 Siting Guide aStep 2: In the Watchlist, create a new Watch that has rules based on the execution time of the probe. Set the Caution to 5000 (milliseconds) to have a caution alert occur when the report takes longer than S seconds, This is the most time a user expects # repart to require to process. a lar Admibtratoe | Pttsces ee | 0908 wa roy tw oan Set the Danger Rule to be seme larger amount of time, such as 19 seconds. See the Bld Admin Guide for mare information, including information on creating CMS Database alerts, The performance of the CMS is vital te the overall operation of the BusinessObjects environment and setting probes to monitor its performance is recammended. Using Audit Reports Usage information for yaur Bld deployment can be ebtained by using audit reports. This can be wery helping in determining hovr many users are using certain Bl tools, certain documents, ete, You can download a.set of reports that are helpful for reporting on the BusinessObjects audit database here: httei//sen.sap.com/dees/D06-6175 Es © SAP AG ~ Business Intelligence 4 Sizing Guide aMonitoring CPU and Memory Usage Monitering and recording of the basic CPU and memory usage of the machines in your deployment is also recommended. The goal s to have an average CPU usage of 60% to 65% in order to handle peaks in the range of 80%. Analyzing the historical usage of your deployment can help you determine if more resources are needed far a particular nade, Similarly, if memory is frequently being fully used, you may be experiencing reduced performance and need to add more BI Too! Simulation Workflows Creating simulated user workflows are very important in the testing of s:new deployment. The following workflows are starting points for your user simulation. 8s sure your simulations make use of the features that most users will use, Crystal Reports 1. Logon to OpenDocument web application 2. View the repertin HTML viewer 3, Enter the Dynamic Cascading Parameters for live data and view first page 4, Gotathe 2nd page 5. Gotathe page in the middle of the report Dashboards 1. Open Bl launch pad nage 2. Logon to Bi launch pad 3, Navigate te Document List Tab 4. Expand Public Folder 5. Opan the folder containing the Dashboard 8. View the Dashboard 7. Refresh the dashboard 8 Close the Dashboard 9. Logot Web Intelligence 1. Open 8! launch pad page and login 2. Navigate to the folder that contains the Webl report 3. Openthe report 4. Open Advanced Prompt and Run Query for retrieve live data 5. Navigate to Sub Report, “Nation” 6, Start Drilling Mode, Drill Down, and End Drill 7. Close Document 8. Logoff Analysis OLAP 9. Open BI launch pad page 20. Logen to BI launch pad 36 © SAP AG — Business Intelligence 4 Siting Guide =Li, Navigate to Document List 42, Open Saved Workspace 13, Sap Axes 14, Print Analysis 35, Close Workspace 16. Logett Suggested Service Settings and Limits The following is a lit of specific services and associated configuration information and suggested best practices. Central Management Service (CMS) The Central Management Server (CMS) can scale horizontally and vertically. The CMS provides server, user, session management, and security (access rights and authentication) management. A single CMS instance can service a large number active concurrent users, depending on the resources of the machine on which an instance runs, 800 active concurrent users per instance is.a conservative number useful far the purpases of planning yaur deployment. The recommended method for adding OMS instances ta your deployment is to monitor the CPU and memory eansumation of each CMS instance with a target goal of around 68% utilization. Adjust the number af ChSes while under representative user load to achieve approximately 65% utilization. Too many CMSes can add unnecessary synchronization traffic betureen instances. Toa feu CMSes will hinder performance of the deployment. Keep in mind that all CMS instances share the same repository datebese. If your repository DB vendor doesn't offer low-latency horizontal scalability, you'll need to ensure you follow your 0B vendor's guidance an howto size and scale vertically, expecislly far deployment with thousands of active concurrent users. ** The performance of the CMS database is very important to the overall performance of the system, It is strongly recommended that your CMS database be sized and tuned fer high performance aperation by ‘a qualified DBA, See the section, CMS Database Tuning CMS Database Tuning, for more information. Crystal Reports Cache Service The Crystal Reports Cache Service scales both up end out and can support 400 active concurrent users per instance, You will very likely not need more than 1 instance on any given machine, since the Cache Service is never enough of a bottleneck to make this necessary. Because of the volume and frequency of communication betwaenthe Processing Service and the Cache Service, t's recammended that you deploy in pairs, That is, on each machine where there is a Cache Service, also deploy a Processing service and vice versa. This can yield better performance than having the Cache Service and Pracessing Service on different machines. 37 © SAP AG — Business Intelligence 4 Siting Guide =Dashboard Cache Service, Dashboard Processing Service Because the technology used for the Crystal Reports and Dashboard Cache and Processing Services are largely the same, the scalability characteristics and recommendations are also the same. Refer to the Crystal Reports Cache Service and Crystal Reports Processing Service and use the same information for Dashboards, Required memory is calculated as follows: # of children = # of logical cores +2 Each child may use up to 1.5 GB of RAM Crystal Reports Processing Service When sizing for workflows involving a small number of users viewing Crystsl Reports, you may notice high values for recommended memory. You may also notice that for a larger number af users the requirad memory is almost constant (it does not increase much when more users are added). This is the expected behavior and itis due to the internal functionality of the Crystal Reports processing servers and engine, which are optimized to offer smoother overall performance, especially for s higher number of users over a longer period of time. Theaptimal allocated memory is Based an an algorithm that uses the number of CPUs on the machine rather than the number of users in the system, Because of this, it's not recommended to have more than one Crystal Reparts Processing Service on any given machine. As you scale-out your deployment, it's batter to add instances to other machines that dan’t current have Required mamery is caleulsted as follow Crystal Reports 2011 # of children = # of logical cores "2.5 Each child can use upto 500 MB of RAM. > 1.25 GB percore Crystal Reports for Enterprise: # of children = # of logical sores +2 Each child can use upto 3 G3 of RAM > 15GB per core Because of the volume and frequency of communication between the Processing Service and the Cache Service, it's recommended that you deploy in pairs. That is,an each machine where there is @ Processing Service, also deploy a Cache Service and vice versa. Data from the database is written to disk in the temporary directory during processing. The speed of ‘this-disk will have an effect on precessing performance, Note: Each Crystal Reports child process (as noted above) can procass a number of report processing, jobs at once. A job can potentially be shared among multiple user requests depending on sharing criteria, Crystal Reports 2011 child processes default to 40 jabs each while Crystal Reparts for Enterprise child processes default to 20 jobs each. The number of jabs par child determines how many child processes are crested. The number of users served by 2 child process can be greater than the job limit if 38 © SAP AG — Business Intelligence 4 Siting Guide =job sharing is possible. Sharing can occur when a job's report processing criteria match, such as data security and parameter prompts. It is the role of the cache server to control job sharing, File Repository Service (FRS) While your deployment can have multiple Input and Qutput FRS instances, the first FRS pair i output] to register with the CMS become the only active FRS pair, Other FRS instances are.considered passive backups. Although all FRS instances run simultaneously, only the active FRS pair handles requests. Ifan active FRS fails, a passive FRS is changed to active status. When the previously active FRS becomes operational again, its registered as a passive FRS with the CMS, Good disk and I/0 performance is ritical for the operation of the FRS ‘Web inteligence (Webl) Processing Service You can scale the Web Intelligence Processing Service up or eut. For every 200 active concurrent users, ivis suggested you create a Webl Processing Service instance. This amount is suggested given the I/O processing that Webl usually requiras, You can adjust the number of active concurrent users allowed per server by adjusting the Maximum Connections setting, In-order to ensure the load on Webl serversis spread evenly among multiple instances, ensure the sum of all Maximum Connections settings is greater than or equal to the number of active concurrent users expected. For example, iFyou had 400 active concurrent users and twa Webl Processing Service instances, each should be set to at least 200 for Maximum Connections, for a total of 400. When adding additional Web! instances, it is recommended that each instance run on a separate maching in order to maximize the I/O capacity aveilable to the servers. If you need to accommodate more users ona single machine and it has the I/O capabilities, processing and memory to handle more Users, you can just increase the Maximum Connections setting {in addition to increasing the memory allocation), 39 © SAP AG — Business Intelligence 4 Siting Guide =The following shows how the Maximum Connections setting is configured. Web Inteligence Core Service Use Configuration Template Timeout Before Recycing (seconds): Idle Document Timeaut (seconds): Server Poling Interval (seconds): Maximum Documents per User: Maximum Dacuments Before Recycling: FF Allow Document Map Maximum Size Errars Ile Connection Timeout (minutes): FF Enable Memory Analysis Memory Lowee Threshold (MB): Memory Upper Threshold (Ma): Memory Mascmum Threshold (MB): FF enable APS Service Monitoring When tuning your Webl server instances, be sure te review all the cattings shawn here to besure they make sense for your intended usage. For example, if you intend to scale Webl up to take advantage of a very capable machine but don’t adjust the Memory Maximum Threshold setting to match, you may find the Webl service recycles itself unnecessarily. See the Bl4 Admin Guide for more details on each of these settings, Data Source Considerations When sizing Webl, it is important to consider the date source(s) being used. Ifyou are using UNX Universe or SAP BW data sources, the DSL Bridge service is employed Ry Webl, which hosts the data connectivity. See the 40 © SAP AG ~ Business Intelligence 4 Sizing Guide aWeb Intelligence section above ragarding sizing of the DSL Bridge in addition to the next section. DSL Bridge Service The DSL Bridge service is used by Webl when it connects to SAP BW and UNX-format Universes, It hosts the data connectivity to these data sources. Memory is calculated based on the number of expected active concurrent users as follows: * 0.25 GB peractive concurrent user © 868 minimum © 2063 maximum The recommended maximum number of active concurrent users per DSL Bridge service is therefore 120. H you require support for more active concurrent users, yau can create additional DSL Bridge service instances, You can create DSL Bridge service instances with smaller maximum heap sizes if that suite your landscape better. Connection Server For data connectivity scenarios that use the Connectian Server (commanly Webl and Crystal Reports), the recommended database connectivity is ODBC or native middleware, where available. For Java-based {JD8C) middleware, including JDBC drivers for HANA, additions performance configuration may be required for the Java Virtual Machine (JVM) that is created by the Connection Server that hosts the JDBC driver. For larger user loads, the default VM memory settings may be inadequate, The memory allocated te the JVM created by the Connection Server can be increased by editing the esc file, located in ...instelidic/sap_bobi/enterprise_xiQ0/dateAccess/connectionServer on Unicand Linux and here on Windows: C:\Program Files (x86)\SAP BusinessObjects\SAP BusinessObjects Enterprise XI 4.O\dateAccess\connectionServer. Sot the “Kms and =Xmx settings as shown below. Here they ara being set to 168 at startup and 268 maximum mamery usage. €l-- The default JVM configuration can be overriden here ¢Libraryhame INIVersion --> soptéens> * en ssn INI_VERSION_4_4" »ABSOLUTE_PATH/Sve.dL1/sap_bobj/tomeat /bin/setenv. sh Invthat file, change the JAVA_OPTS parameter to include -4X:+UseParallelOldGc. tt will lack similarto this: Minimum Memory Settings For Java-based processes such as the APS-based servers in a deployment, setting a nominal minimum memary setting aids Java in its management of memory. The default for Java process such as the APS is 32M, which is wary small. For most APSes, the minimum suggestion is $12 MB, depending on the expected load of the system. How te Change the Initlal (minimum) Memory Setting of an APS # Inthe CMC, go to the Services listing Right-click on the server (APS) * Choose Properties # Locate the Command Line Parameters © Scroll tothe right until you locate -Xms parameter, if present, * Change or add the parameter, setting the number after the -ms w h ne space, For example to set the APS to start with $12. MB, set the command line te read -Xms512M ) SAP AG — Business Intelligence 4 Sizing GuideWeb Intelligence Tuning DSt Bridge You should ensure there are one or more dedicated APS instances running only the DSL Bridge Service, The recommended heap size is 0.25 GB per active concurrent user, 868 minimum, 30 GB maximum. More DSL Bridge service instances can be created if necessary to suppert the number of active concurrent users required. Maximum Connections Under heavy load, the Webl Processing Server can benefit fram having @ maximum connection setting that is higher than the maximum number of expected concurrent connections. This is because under heavy load, the time te close a connection can increase and Web| may not be ready to accept the next connection in time. See the section Web Intelligence (Webi) Processing Service Web Intelligence (Wel) Processing Service for mare information on setting the maximum connections. Multiple Processing Instances It isnot recommendad that yau run multiple Weal Processing Server instances on the same machine, The two services can degrade the perfermance of each other as they make heavy use of I/O bandwidth, etc. If you wish to run more than ane server instance, they should be run on separate machines. SAP BW If you are using SAP BW, the recommended DSL Bridge Service paramaters are: + BICSResultGhunkSize < of rows of the biggest rowset> / 50 + BICSLOVChunkSize <# of items of the biggest LOV> / 20 Thase are both registry settings that are set differently on Windows and UNIX. (On Windewe: Go to [HKEY_LOCAL MACHINE\SOFTWARE\SAP BusinessObjecte\Suite X1 4.0\default\Webintelligence\Calculator] and create the above as DWORDs (decimal) (On UNIX: Go te INSTALL_BASE_DIR/sap_bobj/data/.bobi/resistry/64/softwrare/sap businessobjects/suite xi d.0/detault/webintelligence/calculator and add two keys using parameter names as above to the .registry file. Visualization Service The Visualization Service (CVO) generates graphics and charts for Webl. This service should run in one or more dedicated APS instances. 26B of memory is recommended for this service. Web Application Server For deployments that involve sizable numbers of users, mobile users or remote users, itis recommenced that the Web tier be split to deliver static content by a separate Web application server, Static content (HTML pages, images, documents, JavaScript and Cascading Style Sheets) - content that doesn’t change—can be delivered by a Web server dedicated to that tack. The Apache Web server is: typically used for this 56 © SAP AG — Business Intelligence 4 Siting Guide =You can learn more-about splitting the Web tier by consulting the Web Agolication Deployment Guide found on htto://nele.sep.com, Additional help en doing a split Web tier can be found in the SAP Community Network document, “Improving the User Experience in SAP Bl Platform 4.0 with Apache and WDeploy.” An update to this document is available that covers SAP BI 4.1 and the latest version of Apache httpd, version 2.4, You can find that documant here: http://sen.sap.com/docs/DOC-50325 ‘A simplified guide has been created that walks you through the steps of the Apache Split for Windows deployments: Quick Stea-t je to Ag lit Deployment with BIS, Te learn mare about deplaying multiple instances of Tameat, see the SAP Business Intelligence Platform Pattern Books, Crystal Reports for Enterprise Tuning Controlling Job Creation To ensure the number of job processing child pracesses created by the Crystal Reports Processing Service doesn’t exceed your needs, itis recommended that you configure the Maximum Concurrent Jobs entry to match the number of users you expect to support. Open the Properties of the CrystalReportsPracessingServer and locate the Crystal Reports Processing Service section 3s shown here: Crystal Reports Processing Serie Use Configuratian Template DLL Name: Temporary Directory: ‘SeDotaultDtaDiry CrystalReportsP Manamum Concurrent Jobs (0 for automatic): ° Maximum number of Pre-started Chidren: i ~«Y Ile Connection Timeout (minutes): 20 Tle 3ob Timeout (minutes): 20 FF Share Report Data Between Chents Viewer Refresh Always Yields Current Oata (Oldest On-Demand Data Given ta Clients (seconds): (g Java Class Path: ‘S¢CommondavalibDir94/ proc jar Java Child VM Arguments: (Dousinassabjects.connectuty. direct TT Restore System Oefauits [Set Configuration Template 87 © SAP AG ~ Business Intelligence 4 Sizing Guide aIncreasing Memory Allocation for Improved Response Time Depending on your reporting needs and the design of your reports, it may be possible to imorove the processing time of reports by allocating more memory for Crystal Reports to work with. This is done by increasing the amount of memory allocated to each Crystal Reports child process. You may want to do this in the case where you have extra processing and memery headroom on s machine beyond your ‘essential siting requirements. Allowing Crystal Reports to more quickly process reports has obvious responsiveness benefits and also allows the system to be better abla to handle spikes in demand. Each Crystal Reports processing child process has a defautt memory allocation of approximately 268. You can increase this by setting the Java memory parameter in the Java Child VM Argumentlist, as shown below in the settings for the CrystalReportsProcessingServer in the CMC, In this example, the memary is changed to.8GB by specifying 8192 MB using the parameter setting Xmx81920. Frapertios node Crystakeportt rac Sereer 7 Tide eaten = ae Use Confira Vera Hace Nees ang Server croup) | Temporary Braco: ‘isfauiuatial yeaah annum Concur be (0 for autora: a aur lene beach Maxroum Wariber of Presa Chien: 1 be Coremton Harm Grten)= = Ba Tet (ma) a \& StareRegor ata Between Cents 1 nt Raton sy Yk Cet ats (lies or-bemar Data Go Chere (secs) Cas Path en Oa egies: fesore Sytem befits set coneurason Terai. You may want to increase the number of concurrant jobs if your goal is to increase the number af reports that each child process can deliver. As showin here, the number of cancurrent jobs has been set 1080. 58 © SAP AG — Business Intelligence 4 Siting Guide aDashboards Tuning To ensure the number of job processing child processes created by the Dashboards Processing Service doesn't exceed your needs, tis recommended that you configure the Maximum Concurrent sobs entry to match the number of users that you expect to support. Dashboards Processing Sendce 7 Use Configuration Template Manammurn Concurrant sobs (0 for automate): |g Maximum Lifetime 2obs Per Child: 10000 Mazimurn Number of Prestarted Children: L dle Connection Timeout (minutes): 15 |e Job Timeout (minutes): 15 Java Child vet Arguments: ‘xrsS6M, Dswlinjecton.lang,drect 7 Restore system Defaults Set Configuration Template 59 © SAP AG ~ Business Intelligence 4 Sizing Guide aAnalysis OLAP Tuning [Analysis OLAP (MultisDimensional Analysis Service or MDAS) works the best when run in its own APS. ‘See the section Step 7: APS Configuration, “Step 7: APS Configuration”, Memory (Once in its own APS, it is recommended that MDAS be configured with a minimum of 4GB of memory, The section Step 7: APS Configuration outlines the steps for configuring the memory settings af an APS, Sessions The MDAS "Maximum Client Session” setting is often set to 8 low number by default. It is recommended that it be set in a range of 45 to 100 depending on the expected load. Each MDAS can support up to 100 client sessions. Te support more sessions, create additional APS instances that hast an MDAS service. It is recommended that additional MOAS APSes be deployed on separate machines, Muli Qimensional analysss Service Use Configuration Template OF cells returned by a query: 100000 Maximum number of members natured when filtering: “roo000 T Restore System Defaults TO Set Configuration Template 60 © SAP AG ~ Business Intelligence 4 Sizing Guide aSAP BW Considerations and Recommendations When using BIS to report on SAP BW data, the following recommendations are made to ensure the best performance possible: BW Configuration ‘+ Apply the latest fies available from SAP by applying the latest suppert packs to your BWW system ‘+ Itis strongly suggested thet your BW be sized for the anticipated query load expected fram the 81d system. * Have @ BW expert tune the system generally and BEx queries specifically for your Bld usage Limiting Query Result Size Consider the BW Safety Belt feature, which helps to ensure report developers don't create reports with result sets that are larger than desired. The Safety Belt faaturs is described here: https://sen.sap.cam/does/D0C-31900 BEx Query Performance Check the performance of BEx queries. See the SAP nate: BiLCuery Runtime Statistics which is found on the SAP Support Web site at httes/Jhelp.sap.com/saphelo_nw70/helpdata/EN/43/e29fd25 #50278 10000000815 5247/frat htm You can determina how much time the execution of certain user actions requira in the front-end and in the SAP BW analytic engine. Navigational Attributes A large number of navigational attributes defined in the underlying InfoPravider may impact the overall performance (please refer to SAP best practices for data modeling). This can lead the Bid tools to generate lot of "Crossjoin" operators, causing extra unused data to be returned in qui BExqueries to be sure they include only the infarmation necessary for Bl. ies, Revisit your Characteristic attributes can be converted into navigational attributes. They can be selected in the query inexactly the same way as the characteristics for an InfoCube. In this case, 2 new edge/dimension is added to the InfoGube. During the data selection for the query, the data manager connects the InfoProvider and the master data table join’) in arder to fill the Query. From @ pure performance point of view, yau shauld model an abject on a characteristic rather than ona navigational attribute. 6 © SAP AG — Business Intelligence 4 Siting Guide =8W Specific Tuning and Configuration ‘Set Specific Properties '* Check the option: Use Selection of Structure Elements ‘+ Choose the Read Mode option: Query to Read When You Navigate or Expand Chacking the Read Made property helpsimprave performance in situations where there are many Restricted Key Figures and other calculations in the Query Definition. analysis of the query performance indicates very high EVENTID 3200 times and/or the FEMS number is very high, try enabling this property and check performance. Checking the Use Selection of Structure Elements option ensures the structure elements are sent to the database for processing You can see these settings shown below: ‘Query Monitor (Qesccue [Oeste Debug B] Genera Repo | [[EPrpetes [OBmessaces [Grice roxs [SFeenermonton | |Hirerormance im [[ETeceicaliromnaton |Ecaens Nornor [SEcueyvarants [108 Test [Roetvaracn [Scomata vant | [Bhscar vee [Ebsarawven | use oS TPS auey tear i B] rarer O Peremer2 ead Mode | neyo Fd man You Haig EEA Tinto Faq Status [GAIPeaPrniders Uo Ralevted Chava PROT intvPr Cache Made {Nain Mero Cache Pict Swapsing _|E h/uPnier Seting TUdate ache Geiecs Delia Process (2rvoProier Sting (lee Season af Sacre are oowtetsteweracind Murer (Dropmate orcessing (Cloceneratonog Optintaton Mose [Diaueer Wt Be Oririaed ar Oaneraion Distasi bey Figures input Ready Disagarepenon an Tatas | Daa a Seatetc et [Terror EnaiCaewlaton Layer ata] a © SAP AG ~ Business Intelligence 4 Sizing Guide a‘Activate OLAP Cache and Delta Cache Anew cache mode is delivered with Support Package 16 for 817.0 and currently, the new OLAP cache in B2P is not active yet (SPL7 for BI7.0 is implemented in 82P), The Delta cache, which is expecially important to uncompressed InfoCubes, is not in use for some queries at this moment. This means that the OLAP has te fetch the data from database again ence new requests are sent to InfoCubes, while delta cache is not enabled. This can result in an increase of OLAP time as well as Database time and ‘eventually the entire query runtime. It is recommended to activate the new OLAP Cache mode by running the report SAP_RSADMIN_MAINTAIN and setting "Update cache objects in delta process". Additional Content More BW performance tuning information can be found here: © Performance Ootimizing SAP BusinessObjects Reports Based Upon SAP BW Using BICS Connectivity http://wurw.sdn.sap.com /iri/scn/g0/ portal/prtroot/docs/library/uuid/d0s3c552-e4.19-3010-1788- b3226210058d?overridelayout=trus © Business Intelligence Performance Tuning https://inws_sdn.sap com iri/sdn/bicperformancectunin © Performance & Scalability Best Practices and Tuning Recommendations for SAPBO XI 3.1 on BW. hitp://werwscn.sap.com/iri/scn/go/ portal/prtroot/ docs/library/uuid/005b1374-8F91-2d10-Fe9c- #Bfa1 25259570 uickLink=indexBoverridelayout=true 4874787963301 1 cy © SAP AG ~ Business Intelligence 4 Sizing Guide =CMS Database Tuning The CMS database performance has a significant impact on the performance of the system, The following Sybase database tuning tips are offered. These tips may apply to other database servers as 1 64 Data Cache [buffer pool] Data cache sllows efficiently reducing IO. Having a sufficient size of data cache can speed up queries. Procedure Cache Configure the Procedure Cache so frequently used database procedures are cached. This allows the database to not need to reload procedures from disk as often. Lock Granularity The CMS does:a lot of reading and writing from/to the database. 8y default, the granularity is st to page level, which is toc high. Page level lacking promotes contention by the CMS when, for example, several records are stored together in a same page. One of them may be being updated but another one is being read. One of these twa sessions has then te wait until other has finished. Setting Lock Granularity te use row level locking can aveid contention and deadlork issues, Parallel Processing: This is important te set so Sybase can fully utilize all the CPU cores available on the machine, © SAP AG ~ Business Intelligence 4 Sizing Guide =Virtualization: Service Level and Performance B14 Works very well in virtualized environments, Testing has shown it performs equally well on virtualized hardware as it does on dedicated hardware. However, in order to achieve equal performance, the hardware it runs on must be dedicated to the virtual machines that the BI services run SAP strongly recommends that Bl virtual machines have reservations far the memory and CPUs assigned to them. Bl is a bursty application workload. it can be very 1/0 intensive when heavily loaded by users. When ‘experiencing maximum load as defined by your sizing, all parts of the system can go from moderately loaded to very busy momentarily. Fthe BI hardware is being shared with other workloads in a virtualized environment, the response to bursts in activity can result in slow-downs between dependent: Bl services and ultimately and user response time. It is recommended that you build your system to target 65% utilization so that bursts in activity can be handled. Performance of both physical and virtual systems degrades when system utilization is ereater than 80%, Hf you are unable to secure memory and CPU reservations for your virtualized deployment, itis recommended that additional scale-aut be done to provide adequate Bl processing resources for peak periods. The amount of additional scaling out will depend on your IT infrastructure and performance policies. The goal af Sizing Bid is to have a deployment that can hendle the workload needed and provide good response time for end users. If you deploy to a shared virtualized environment, extra cares needed to ensure your be maintained. resultsin a system which has a Service Level and responsiveness that can For more in-depth information on | and Virtualization, see the SAP Bld Virtualization site at mb alization 6 © SAP AG — Business Intelligence 4 Siting Guide =Desktop Virtualization When running BI4 desktop applications sre deployed to virtualized desktop environments such as Citrix ‘and Windouss Server Remote Desktop Services, sizing is required to determine how many concurrent users can be supported far commenly used Bl applieatians, The following table shows the memory footprint of applications that you may run in a virtualized desktop environment. You can use the memory footprint infarmation to calculate the amount of memory required to service the number of users required. “Product =——SS~*~<~*~s*~

You might also like

• SAP BusinessObjects BI4 Sizing Guide PDF

  

  No ratings yet

  SAP BusinessObjects BI4 Sizing Guide PDF

  76 pages
• Sizing Guide

  

  No ratings yet

  Sizing Guide

  20 pages
• Determine Your Sizing Requirements With Quick Sizer

  

  No ratings yet

  Determine Your Sizing Requirements With Quick Sizer

  5 pages
• Business Solutions by IBM and SAP: Send Completed Questionnaire To Either Your Local IBM Contact or

  

  No ratings yet

  Business Solutions by IBM and SAP: Send Completed Questionnaire To Either Your Local IBM Contact or

  24 pages
• Hardwae Sizing Formulas

  

  No ratings yet

  Hardwae Sizing Formulas

  41 pages
• SAP BPC 7.0 NW Configuration Guide

  

  No ratings yet

  SAP BPC 7.0 NW Configuration Guide

  11 pages
• SAP Solution Manager 7.0 As of SP15: Sizing Guide

  

  No ratings yet

  SAP Solution Manager 7.0 As of SP15: Sizing Guide

  19 pages
• Principles of SAP HANA Sizing - On Premise and Cloud

  

  No ratings yet

  Principles of SAP HANA Sizing - On Premise and Cloud

  47 pages
• BPC Administration Guide

  

  No ratings yet

  BPC Administration Guide

  60 pages
• SAP Virtualization Sizing Guide

  

  No ratings yet

  SAP Virtualization Sizing Guide

  5 pages
• SAP Solution On IBM

  

  No ratings yet

  SAP Solution On IBM

  100 pages
• Quick Sizer Factsheet

  

  No ratings yet

  Quick Sizer Factsheet

  2 pages
• SAP HANA Sizing Guide

  

  No ratings yet

  SAP HANA Sizing Guide

  47 pages
• SAP Sizing Guide for Beginners

  

  No ratings yet

  SAP Sizing Guide for Beginners

  14 pages
• GRC - PC - 10 0 - PreliminarySizingGuideline - V5

  

  No ratings yet

  GRC - PC - 10 0 - PreliminarySizingGuideline - V5

  16 pages
• Bi4 SG Feb2014b

  

  No ratings yet

  Bi4 SG Feb2014b

  75 pages
• MDM Sizing Guide

  

  No ratings yet

  MDM Sizing Guide

  32 pages
• SAP Questionnaire 061510

  

  No ratings yet

  SAP Questionnaire 061510

  55 pages
• Sap Hardware Sizing

  

  100% (1)

  Sap Hardware Sizing

  3 pages
• SizingAM 2017

  

  No ratings yet

  SizingAM 2017

  16 pages
• Sap Implementation Questions

  

  No ratings yet

  Sap Implementation Questions

  36 pages
• SSM10 Installation Guide

  

  No ratings yet

  SSM10 Installation Guide

  90 pages
• Business Solutions Ibm and Sap

  

  No ratings yet

  Business Solutions Ibm and Sap

  54 pages
• Kpro March 2006

  

  No ratings yet

  Kpro March 2006

  6 pages
• Installation Guide Sap Businessobjects Access Control™ 10.0, Process Control™ 10.0, and Risk Management™ 10.0

  

  No ratings yet

  Installation Guide Sap Businessobjects Access Control™ 10.0, Process Control™ 10.0, and Risk Management™ 10.0

  52 pages
• SAP BPA Sizing Guide V9

  

  No ratings yet

  SAP BPA Sizing Guide V9

  11 pages
• CRM Billing V2

  

  No ratings yet

  CRM Billing V2

  9 pages
• How To BPF BPC 7.5 NW

  

  No ratings yet

  How To BPF BPC 7.5 NW

  37 pages
• Installation Guide SAP Enhancement Package Installation

  

  100% (1)

  Installation Guide SAP Enhancement Package Installation

  88 pages
• 3609 Variant Configuration and Make-To-Stock Production - A Wish or Reality

  

  No ratings yet

  3609 Variant Configuration and Make-To-Stock Production - A Wish or Reality

  16 pages
• SAP Sizing Request Form

  

  No ratings yet

  SAP Sizing Request Form

  25 pages
• ADM115 01 01 MethodsTools PDF

  

  No ratings yet

  ADM115 01 01 MethodsTools PDF

  32 pages
• BPC 75 NW Operation Guide

  

  No ratings yet

  BPC 75 NW Operation Guide

  68 pages
• BPC Office Guide2011

  

  No ratings yet

  BPC Office Guide2011

  159 pages
• SAP Configurations with HP Servers

  

  No ratings yet

  SAP Configurations with HP Servers

  46 pages
• Guide de Sap Netweaver hp1

  

  No ratings yet

  Guide de Sap Netweaver hp1

  130 pages
• Upgrade Guide SAP Enhancement Package 4 For SAP ERP 6.0 ABAP

  

  No ratings yet

  Upgrade Guide SAP Enhancement Package 4 For SAP ERP 6.0 ABAP

  144 pages
• Sbo41 Bip Deploytemplates en

  

  No ratings yet

  Sbo41 Bip Deploytemplates en

  1 page
• SAP Sizing Methods & Tools Guide

  

  No ratings yet

  SAP Sizing Methods & Tools Guide

  31 pages
• Introduction Into Post Go-Live Sizings

  

  No ratings yet

  Introduction Into Post Go-Live Sizings

  26 pages
• BPC-Custom Process Chain

  

  No ratings yet

  BPC-Custom Process Chain

  45 pages
• SAP Cloud Sizing & Tuning Guide

  

  No ratings yet

  SAP Cloud Sizing & Tuning Guide

  10 pages
• SSM10 Server Upgrade Guide From 75

  

  No ratings yet

  SSM10 Server Upgrade Guide From 75

  38 pages
• Business Intelligence 4 Adaptive Server Processing

  

  No ratings yet

  Business Intelligence 4 Adaptive Server Processing

  6 pages
• FPM Slides

  

  No ratings yet

  FPM Slides

  16 pages
• Lumira Sizing

  

  No ratings yet

  Lumira Sizing

  27 pages
• Better Business Intelligence

  

  No ratings yet

  Better Business Intelligence

  14 pages
• Xi4 Install Unix en

  

  No ratings yet

  Xi4 Install Unix en

  76 pages
• Unit 1: Vision and Customer Value: Implementing SAP S/4HANA Cloud With SAP Central Business Configuration

  

  No ratings yet

  Unit 1: Vision and Customer Value: Implementing SAP S/4HANA Cloud With SAP Central Business Configuration

  8 pages
• SPM Operation Guide

  

  No ratings yet

  SPM Operation Guide

  54 pages
• SAP - BPC - 70 - SP01 - NW - Sec PDF

  

  No ratings yet

  SAP - BPC - 70 - SP01 - NW - Sec PDF

  40 pages
• SAPS - SAP Sizing Benchmark Https WWW - Sap.com About Benchmark

  

  No ratings yet

  SAPS - SAP Sizing Benchmark Https WWW - Sap.com About Benchmark

  4 pages
• How To Set Up Standard Roles in SAP HANA

  

  100% (1)

  How To Set Up Standard Roles in SAP HANA

  26 pages
• Introduction Into Post Go-Live Sizings

  

  No ratings yet

  Introduction Into Post Go-Live Sizings

  22 pages
• Sizing SAP HANA

  

  No ratings yet

  Sizing SAP HANA

  44 pages
• Databricks Genai Slide

  

  No ratings yet

  Databricks Genai Slide

  1 page
• BW Stock Dataload Communication

  

  No ratings yet

  BW Stock Dataload Communication

  22 pages
• Arabian Oryx: UAE's National Animal

  

  No ratings yet

  Arabian Oryx: UAE's National Animal

  1 page
• SAP Analytics Cloud and SAP Digital Boardroom

  

  No ratings yet

  SAP Analytics Cloud and SAP Digital Boardroom

  21 pages
• How To Report On Cross Company Stock in Transit PDF

  

  No ratings yet

  How To Report On Cross Company Stock in Transit PDF

  30 pages
• Eng PDF

  

  No ratings yet

  Eng PDF

  302 pages
• SAP Yard Logistics SAP Yard Logistics SAP Yard Logistics SAP Yard Logistics

  

  No ratings yet

  SAP Yard Logistics SAP Yard Logistics SAP Yard Logistics SAP Yard Logistics

  1 page
• Unit 1: End-User Perspective

  

  No ratings yet

  Unit 1: End-User Perspective

  1 page
• SAP BI Road Map Overview

  

  No ratings yet

  SAP BI Road Map Overview

  5 pages
• Connecting To SAP BW With Microsoft Excel 2007 PivotTables and ODBO

  

  No ratings yet

  Connecting To SAP BW With Microsoft Excel 2007 PivotTables and ODBO

  28 pages
• 2017 Spring Season Variant List

  

  No ratings yet

  2017 Spring Season Variant List

  572 pages
• Dashboards For Quality Management

  

  No ratings yet

  Dashboards For Quality Management

  22 pages
• Connecting To SAP BW With Microsoft Excel 2007 PivotTables and ODBO

  

  No ratings yet

  Connecting To SAP BW With Microsoft Excel 2007 PivotTables and ODBO

  28 pages