Life cycle of a Planning Server application
Updated: 2009-04-30
PerformancePoint Planning software is enterprise-oriented software designed for your business needs in the area of budgeting, forecasting, consolidation, and other business purposes. Planning Server allows you to customize the system to meet your business requirements.
Due to the complexity of the system, we recommend a three- to five-phase implementation of Planning. The complete life cycle of the implementation of a Planning Server application includes a Proof-of-Concept (POC) phase, Design phase, Testing phase, Pre-production phase, and Production phase. Some of these phases can be combined depending on your environment. For example, you could either skip the POC phase or combine the POC and Design phases. You may also choose to run Testing and Pre-production individually or combined into one phase.
The following is an example of a five-phase implementation:
| Phase | Topology recommendation |
|---|---|
Proof of concept |
One computer |
Design |
Three servers. One computer is also acceptable. |
Test |
Three servers* |
Pre-production |
Three servers* |
Production |
Three servers* |
The following is an example of a three-phase implementation:
| Phase | Topology recommendation |
|---|---|
Proof of Concept/Design |
One computer |
Test/Pre-production |
Three servers* |
Production |
Three servers* |
Note
See the PerformancePoint Server 2007 Deployment Guide for instructions on installing single or distributed environments.
* While the minimum recommended topology for a testing or pre-production or production environment is a three-server configuration, we highly recommend that you analyze your business requirements, goals, and loads, and utilize our scalability potential to pick an optimal topology for your production system.
Once you have completed the Proof of Concept/Design phase and are satisfied with the results, use the migration process to migrate the settings you have customized at the Design phase to your pre-production environment.
The following Planning application life-cycle process diagram illustrates the relationships between each phase and details the steps needed to implement a Planning system.
.gif "application process diagram")
Proof-of-concept and design phases
Step 1: Install and configure
The first step is to install and configure the software to match the needs of your environment. After initially running the installation and configuration wizard, use Planning Server Configuration Manager to finish the installation step. Refer to the PerformancePoint Server 2007 Deployment Guide for instructions on installing single or distributed environments.
Step 2: Planning Server configuration
Using Planning Administration Console, perform the following tasks:
Create a new Microsoft Office PerformancePoint Server 2007 Planning application
Set up system-level security data (set up Global Administrator (GA), User Administrator (UA) Data Administrator (DA) Modeler system roles)
Set up data sources, which will be used by Planning Business Modeler for loading data
Set the data destination location for the outbound feature
Setup or change configuration settings such as timeouts.
Refer to Planning Administration Console online help for details on these tasks.
Step 3: Content Design
Use Planning Business Modeler to design your content. Based on your design plan, define your business structure definitions, model sites, models, dimensions, hierarchies, dimension properties, global assumptions, and model dependencies (linked models). Once you have saved these metadata into the Planning Application Database, use Planning Business Modeler to load reference data from the data sources you defined in the earlier step from Planning Administration Console.
Step 4: ETL Process
After the Content Design step, you can work on the extract, transform, and load (ETL) process to load reference data and fact data into the Planning Application Database. The ETL work can proceed in parallel between the Content Design step and the Business Process Design step.
The ETL process includes the ETL design and ETL execution steps. Optionally, the ETL design step includes:
Identifying source data
Creating your schema mappings
Loading dimension and fact data from sources to the Planning Staging Database.
You can either create SQL Server 2005 Integration Services (SSIS) packages or SQL Server scripts to extract and transform the data into the appropriate format in the Planning Staging Database.
The ETL execution step consists of moving data from the Planning Staging Database to the Planning Application Database. Use either the migration process, using PPSCmd.exe, or Planning Business Modeler to load reference data and fact data from the Planning Staging Database to the Planning Application Database. See the "PerformancePoint Server 2007 Operations Guide" under "Data Integration" for details on the ETL process.
After you have finished the structural definition design, use Planning Business Modeler to do the initial model site deployment. This step creates a new Analysis Services database and OLAP cubes for the new model site that you are working on. This step is required before you work on the next step, Business Process Design.
While you are working on these designs, you can periodically save the metadata and reference data to the Planning Application Database to avoid any data loss. We recommend that you do backups of your system's databases periodically, including the Planning Application and Planning Staging Databases and the PPSPlanningSystem and PPSPlanningService databases.
Step 5: Business Process Design
In this step you:
Design business rules
Design forms
Enter data security
Define workflow cycles, assignments, and jobs
Define model associations mappings
Step 6: Deploy the Model Site
This step performs incremental updates to the model site's Analysis Services database and cubes. All the metadata, including structural metadata, security data, calculation rules data, reference data, and fact data is saved to the Planning Application Database and incrementally processed into the Analysis Services database and cubes.
Step 7: Testing
Testing is the final step. Use PerformancePoint Add-in for Excel, SQL Server Management Studio, SQL Server Business Intelligence Development Studio, SQL Server 2005 Reporting Services, PerformancePoint Monitoring and any other tools your business needs to do the following testing and verifications:
Data accuracy
Performance
Data security
Workflow submissions
Business and financial jobs execution results
The testing process needs to be iterative, where you continually test from the Configuration or Content Design phase through the official Testing phase until you are satisfied with your results. For example, start with creating a model site and one model, repeating this process until you are satisfied with the results. Then expand this process to multiple models within one model site and multiple model sites with one application. During the Configuration phase, test with each configuration setting until you are satisfied. Note that this process does not apply to initial deployment of a model site. See the PerformancePoint Planning Business Modeler help topic, "Model Site Deployment," for complete details.
Note
While you are working on the Design/POC environment, you should perform complete Planning system backup frequently to avoid any data loss. See the "PerformancePoint Server 2007 Operations Guide" for detailed information on how to back up your system.
Migration from the Design phase to the Pre-production and Production phases
After you have completed the Proof of Concept/Design phase and have fully tested the system, you are ready to move on to the Pre-production or Production phase. You can use the Planning Server migration process to migrate the metadata from the source system to the destination system. For reference data and fact data, load data from the Planning Staging Database into the destination system directly. It is likely that the fact data that you used in the design environment is a small fraction of your production system. The migration process only migrates the metadata to the target system. You still are required to design ETL to load data into the target system. You must do the Business Process Design step on this pre-production/production system.
The following is the overview of the PerformancePoint Planning migration. The application migration does not copy the complete source system data to the destination system.
The following information will be migrated during application migration and is shown in the Migration step in the earlier diagram that depicts the application process:
Security roles
Structural metadata: all definitions for model sites, models, dimensions, and member sets
Calculation rules and any associated job templates
Form templates
Calendars
The following will be migrated during data migration and is shown in the ETL process in the "application process" diagram.
Dimension and hierarchy data
Fact data and annotations
Associations
The following will not be migrated during application migration and is shown in the business process design in the "application process" diagram:
Reports
Workflow: assignments, jobs, cycle definition and instances
Users and their security role assignments
Note
All objects that are not migrated must be re-created manually or imported using .csv files.
The major steps to migrate PerformancePoint Planning system to a new destination are as follows:
Export from a source system.
Import to a destination system.
Migrate data: export data from the source system, prepare the staging area, and load the prepared data.
Load workflow, security, and report data.
Deploy all model sites in the destination system.
Bring the migrated system online.
Verify application migration.
See the PerformancePoint Server 2007 Operations Guide for more information on migrating data.
The steps in the Pre-production and Production phases are the same. You follow the same migration process in these two phases.
Data flow within Planning Server
The following diagram shows the high-level data flow throughout Planning Server. It breaks PerformancePoint Planning into three tiers: Client, Server, and Storage. The following section shows the data flow by scenarios from each main client component shown in the diagram. The numbers (such as #1, #3, and so on) refer to the data-flow numbers in the diagram.
.gif "data flow diagram")
Planning Business Modeler:
Content design: #3, #9.1, #18
Business process design: #3, #9.1
Deploy model site: #3, #9.1, #9.2, #13
PerformancePoint Add-In for Excel:
Data entry: #5, #9.1, #9.2
Reports design: #12, #5, #9.1
Publish form templates: #5, #9.1
Run reports: #10
Publish reports to Reporting Services: #19
Run-time scenarios: #5, #10, #12
Planning Administration Console: #2, #8, #9.1
Reporting: #1, #6, #7
PPSCmd: #4, #9.1
ETL:
Sync application database to Planning Staging Database: #14
Loading data from Planning Staging Database to Planning Application Database: #15
Loading data from data warehouse to Planning Staging Database: #16
Data Integration
Before data analysis and planning is available in any business application, Planning Server must get data from one or more combinations of business data sources. Data can be moved from any one of the data sources to the Planning Staging Database. The data can then be validated and moved from the Planning Staging Database to the Planning Application Database.
PerformancePoint Server provides tools and methodologies to move data from your company's internal operational data sources to a staging database. It is important to note that PerformancePoint Server does not provide tools to support extract, transform, and load (ETL) processes.
Careful planning of data loading-processes is important, as failure to properly plan can result in data that does not reconcile between models, dimensions, and related member sets.
Key data-loading considerations include how often you will load data, whether you do a full or incremental data load, the amount of data involved, and performance.
Consider the timing and frequency of data loading and which users will be responsible for these processes. This can become significant with multi-site, multi-model implementations.
In Planning Business Modeler you can load dimension data and related hierarchy data from the Planning Staging Database to the Planning Application Database. When you do, all changes to dimension data in the Planning Staging Database are integrated into the Planning Application Database. This includes new dimension member rows, updated dimension member rows, and deleted dimension member rows. Dimension members that are not changed in the Planning Staging Database remain unchanged in the Planning Application Database. See Planning Business Modeler's online help for detailed information about the data-loading feature.
See the PerformancePoint Server 2007 Operations Guide under "Data Integration" for detailed steps for integrating data.
Download this book
This topic is included in the following downloadable book for easier reading and printing:
See the full list of available books at Downloadable content for PerformancePoint Planning Server.