In Part 1 of this series, I discussed the differences between installation consultants and technical consultants. In Part 2, let's dig into the details of technical landscape design.
For a successful Enterprise Performance Management (EPM) implementation, you need to know a lot more than the number of users and what you'd like to do with it. You need to define more granular information such as the network bandwidth required for each office user, record size, index, and statistic requirements.
But here's the key: You can only address those issues after the functional design is complete. Sizing, for example, depends on the product design plan; Auditing, as another example, requires doing a lot of calculations such as currency translation or allocation logic. At this point, we can give server approximations for a development environment, but production and sometimes quality assurance must be based on the landscape design.
Here is a sample Technical Landscape Design Agenda that Column5 recently compiled for a large multinational customer:
Application Design Overview
This section discusses:
- Overview of the models: User facing and BPC supporting
- Users
- Location
- Amount
- User tasks
- Data feeds
- Models: A model (referred to as an Application in prior versions of BPC) represents an InfoProvider within BW or more commonly referred to as a cube in MS Analysis Services
- Description of each model
- Lists the dimensions within each model
Data Questions Overview
This section gives an overview in relation to the amount of data eventually winding up in the system:
- Security Configuration
- BPC model records counts
- Import process
- Internal data flow
- Input templates
- Trigger calculation (the amount of record will be created from the trigger
- Non-trigger calculation
- Dimensions
- Highlights dimension with members over a limit that will affect sizing
- Fact data counts
- Performance pain points
- Performance issues are discussed here
- Items affecting performance are also described
- Calculations
- Overview of how models will use allocations and calculations
Product Version
This section describes:
- The product and version
- The environment description: This is referring to the system environments, not to be confused with a SAP BPC Environment (referred to as Application Sets in prior versions of BPC), which represents an InfoArea within BW or database in MS Analysis Services.
- The amount of environments
- Names of the environments
- System metrics of each environment
- Server utilization of each environment
- Communication protocols between environments and client machine
- Production system requirements
- Service level agreement (SLA)
- Change management
- Corporate Standards: This can affect how the software service packs are applied and SLA
- Hardware recommendations and sizing
- Software requirements
- Hardware recommendations and sizing
- Software requirements
- Support Resource Model
- Report detailing required client resources to support the BPC systems
- Details regarding BPC specific support
- BPC maintenance tasks
- Report detailing required client resources to support the BPC systems
Next Steps
With the landscape design complete, we move on to installation of a quality assurance and/or a production environment, including SIT/UAT performance tuning and looking at specific processes to improve the performance. This service is a technical audit of the systems integration testing, ensuring the EPM application(s) are functioning as expected following data/meta integration:
- Ensure end-to-end testing from data/meta source system through to EPM suite
- System monitoring to determine optimal performance of SIT processes
- Build an effective testing methodology to deliver meaningful results
- Execute testing, capture results, and measure performance
- Tune system configuration based on testing results
- Test application upper limits to effectively communicate maximum user load
- Onsite/remote support starting 2-4 days prior to go-live
- Technical environment review of infrastructure to verify best practices and address any regular maintenance (windows, database, application, etc.)
- Onsite/remote technical support to address issues as they arise in real-time
- Post go-live, onsite/remote support between 2-4 business days based on client request