The shift from IBM Planning Analytics v11 on-prem to IBM Planning Analytics as a Service (IBM PA SaaS) promises modernization. It also requires careful planning. Here’s what you need to know before making the move.
Why PA SaaS?
IBM PA SaaS is a fully managed cloud solution deployed on Amazon Web Services (AWS) or MSFT Azure, supporting organizations in improved planning, budgeting and forecasting across the enterprise. At first glance, IBM PA SaaS provides some long-awaited capabilities:
-
Native high availability, automatic backups and isolated upgrades ensure minimal disruption
-
Direct API access from TurboIntegrator (TI)
-
Hosting in standard cloud data centers (AWS/Azure)
Nevertheless, migration to IBM PA SaaS also brings significant changes, including:
-
TI process adjustments
-
No command-line access (see our technical note on the deprecation of ExecuteCommand in IBM PA SaaS)
-
New ODBC gateway technology
-
New methods for file transfer and management (learn more about our solutions, including ACG SaaS Uploader and Google Cloud Storage integration)
-
New methods for deployments
With lots of modern amenities and some major adjustments, migrating to PA SaaS might make you feel like you're about to move house. To help you make an informed decision, the following article explores the potential move in more detail.
The 4 Steps to Moving to PA SaaS:
1. Learn the Neighbourhood
1.1 Understand the new architecture
Before you begin, make sure you are familiar with the IBM PA SaaS environment. In your IBM cloud “instance”, there is a bucket of resources that you can manage independently. This allows you to:
- Create new environments and databases
- Upgrade different databases at different times
- Enable high availability on key databases or at key cycles
- Create backup schedules and on demand backups
1.2 ODBC Data Connector (ODBCIS):
- The ODBC Data Connector is configured on a machine within your environment and acts as a "translator" between ODBC calls to your database and ODATA used by Planning Analytics as a Service.
- You will need to install and configure ODBCIS with TLS certificates.
- Replacing cloud-to-cloud connections, it is required for all ODBC connections, providing a centralized location for all ODBC connections
- The PA Server no longer communicates with ODBC drivers; ODBCSIS does the work to query and convert. You can develop and test queries in TI or using Postman. Queries will continue to work – just bulk update your ODBC settings in a Workbench.
ODBC Data Connector – Using Postman
2. Plan the Move
2.1 Decide your methodology for object deployments. Migration utilities include either:
- Git-based deployments (GitHub, AWS CodeCommit, Azure DevOps), or
- Custom REST API or TM1Py scripts.
2.2 Data Migration via Git. Various considerations must be made:
- Data migration doesn’t follow traditional GitHub pull request i.e. PR approval is not needed.
- Network security exceptions may be needed for locally hosted repository with Cloud-based PA (i.e. pinholes may be needed in the firewall)
- With GitHub, you can only move object definition (i.e. code and structure) but no data or attributes.
So how do we move select data between environments? Below is a detailed workflow outlining data migration via Git.
Step 1: Create system cubes in the source to control if data is getting exported.
Step 2: Create Generic Export Processes in Source Database
Step 3: Create Generic Import Processes and sub-processes to deploy to target database
Step 4: Include the Execution of the Export, Import and Directory in your push/branch
Step 5: Include the Execution of the Export, Import and Directory in your push and push to a Git branch
Step 6: Execute the “Pull” into the target environment
2.3 Other changes
Plan for changes involving TM1Web (no longer available), deprecated TI functions (SaveDataAll, ExecuteCommand, etc.), and Control cubes (Stats, Application Security, Capabilities and others)
3. Move-In Day: Migration to IBM PA SaaS
- Purge unwanted objects and data in the current environment
- Download and run the migration tool. This requires a tm1s.cfg with a matching data directory
- Review migration report for required code changes
- Upload and restore the generated tar.gz backup
- Apply fixes and validate functionality
4. Settle In (Fix, Improve and Customize)
There will be a variety of post-migration adjustments you will have to make, including adjustments to:
4.1 Command line utilities
- Given native backup and archival capabilities in IBM PA SaaS, command-line utilities are no longer required for this purpose.
- All command-line utilities must be replaced with REST API and HTTP endpoints
4.2 Email Integration
- In IBM PA SaaS, email integration now uses Microsoft Graph API and requires HTTP endpoints and ExecuteHTTPRequest.
- This replaces the method in IBM PA V12, which used a combination of PowerShell connecting to Office 365 and ExecuteCommand. ExecuteCommand is deprecated in IBM PA V12 (IBM PA SaaS). You can read more about the deprecation here.
4.3 RunTI Replacement
- Use improved RunProcess for intra-instance parallel processing.
- Use Orchestrate with GetJobStatus and CancelJobs
- Use HTTPExecuteRequest with tm1.ExecuteWithReturn for cross-instance orchestration
4.4 Data Movement
- Data can be moved to and from IBM PA SaaS with ODBC (see above), flat files or APIs.
- SFTP is no longer available. We can use REST endpoints to create folders and files on the server.
- For third-party integrations, we can implement ExecuteHTTPRequest, replacing ExecuteCommand to compiled Python or PowerShell
Conclusion
IBM PA SaaS is the future of Planning Analytics. While the migration requires significant technical adjustments, and some elements of the platform still need some work, its scalability, high availability and advanced analytics make the transition worthwhile. As IBM continues to enhance the platform, organizations can begin evaluating their environments and familiarizing themselves with the roadmap. Ultimately, moving to PA SaaS is not a question of if but when.
Reach out to us directly at solutions@acgi.com if you have any comments or questions, or would like to discuss in more detail.