Upgrading any enterprise software application that is used heavily in an organization can be a frightening undertaking. Not only do you have to acquire all of the hardware, licenses, time, and people to make the upgrade happen, you also have to ensure that nothing breaks in the process. When it comes to upgrading SAP BusinessObjects – whether it be a minor patch or a major version change – there are many complexities that can be worked out beforehand to ensure that the upgrade is a success and does not negatively impact your users or business processes.
Timeframe and Anatomy of an Upgrade
Let’s assume that we’re considering an upgrade from SAP BusinessObjects XI 3.1 to SAP BusinessObjects BI4. A typical upgrade from planning through to execution takes anywhere between 3 months to a year – depending upon the size of the deployment and the new technology or features that are being adapted as part of the upgrade process. This total process of an upgrade takes into consideration: planning, preparation, hardware deployment, software deployment, software configuration, testing, and training. The bulk of the time with an upgrade of SAP BusinessObjects, and arguably any enterprise software, is spent in the planning, preparation, and testing phases.
- Planning – determining required scope, budget, skills, non-people resources, licenses, phases for the upgrade
- Preparation – determining affected content, business processes, other applications, and users of the deployment
- Hardware and Software Deployment – getting servers and software acquired, installed, and in place
- Software Configuration – ensuring the software will meet the demands of the users
- Testing – ensuring that the software and servers can do what you think they can do
- Training – teaching your administrators to maintain and your users to take advantage of the shiny new features
Given the work that needs to be done leading up to, executing upon, and cleaning up after an upgrade, a typical organization will spend somewhere between 450 to 1,750 man hours getting themselves to BI4. The goal is to reduce this number so that you can move faster and with fewer headaches.
How Can Sherlock Help?
Prior to executing up on an upgrade with a customer, we run Sherlock against their existing deployment. This gives data that we need so that we can eliminate unnecessary steps in the planning preparation, testing, and training phases of the upgrade. In addition, it gives us better insights into specific sizing that needs to be taken into consideration during the hardware deployment, software deployment, and software configuration phases. Finally, it helps us with preparing your team for maintenance tasks that will need to occur after your deployment is live and stable.
Planning and Preparation
A big part of the planning and preparation process is determining the users, applications, and content that will be impacted by such a big change to a key system like your business analytics deployment. During a Sherlock assessment, we highlight users who do not login, content that isn’t being used, rights and limits that aren’t being leveraged, and problematic content (e.g., Desktop Intelligence, Web Intelligence documents affected by calculation engine changes, or error prone content).
The images below some charts from real customer assessments that highlight inactive users, new Desktop Intelligence created in each month, long running schedules, and error prone content.
The goal is to eliminate all users and content that are non-essential or to fix users and content that are unnecessarily complex or error prone. This reduces the number of hours required to move content, test content, and support content after the upgrade has been completed. We have seen reductions in the amount of content by as large as 90% using this process. Most customers see at least a 25% reduction in the amount of content, number of users, or the complexity of rights as they start their upgrade.
One of the big things to highlight as part of this upgrade process, which was mentioned quickly above, is the amount of Web Intelligence content impacted by calculation engine changes. This can be a big problem if you are heavily reliant upon Web Intelligence and it’s features related to variables and functions. We’ve used Sherlock during upgrade preparation to highlight every Web Intelligence report impacted by each calculation engine between one version and another. This has helped us to ensure not only that those impacted reports are tested after the move to the new version (BI4 in our scenario), but also that we know exactly which users are using that content so that we can warn them ahead of time that there may be issues.
Hardware and Software
When it comes to hardware and software related deployments, there are a few things to consider. All of them stem from an understanding of how many users you intend to service with your business analytics deployment and whether those are users are taking full advantage of the content and services available to them. As shown above, we can tell how many users are active versus inactive; however, we can also point out what percentage of the users are using Web Intelligence content versus Dashboards versus Analysis content. We can tell how many users are primarily using schedules as compared to those who are interacting with content via BI LaunchPad. We can also tell what percentage of users are putting the greatest strain on the system by refreshing large reports, complex reports, or exporting reporting content to another format.
All of these details feed into several ways to analyze this data for the purposes of determining if:
- The right number of licenses have been acquired for each type of user within the BI deployment (i.e., viewers, editors, creators, and schedulers)
- There will be a positive or negative impact on the budget for the upgrade by needing to switch from the current licensing structure to the new CSBL license structure
- The current security model that is being used is too complex for how users are accessing the system and leveraging content on a day-to-day basis
- New licenses or types of licenses are required based upon the use of new tools within the BI4 landscape
- The hardware that has been set aside or acquired is big enough to support the proposed BI4 landscape given the number of users, types of content, and how users are leveraging that content
- This same hardware is able to grow with the BI deployment based upon how quickly the user community and amount of content has grown in the past and may grow in the future
The two charts below show some examples from real license assessment detailing the number of creators, editors, and schedulers in each month and, specifically, the number of new documents created each month. These stats can be used as an indicator of the number of licenses required for viewers versus creators versus schedulers.
In addition, this analysis will allow us to determine how the BI4 deployment should be configured and sized to support the required actions on the part of the user community.
All of the details discussed above will allow us to better define the amount of hardware that is required for a BI4 deployment, the amount of time required to install and configure BI4 and required client tools, and the number of licenses that will be required to support the user activity. This will be a large part of the costs associated with deploying the BI4 system. As such, it is important to ensure we are working with a fact-based analysis, whether than guessing and what we think is correct.
As part of the Sherlock assessment process, we always create a test plan for once the upgrade has been completed. This gives us the list of stakeholders who should be involved in testing, power users that are actively creating or viewing content, the content that is used most frequently, universes and data sources that are used most frequently, the list of reports affected by calculation engine changes, and a list of the most complex reports, dashboards, and Universes. All of this leads us to a very defined scope for testing efforts and the users who should be involved as part of the testing process. By focusing on only testing the critical content, the content that is used heavily, or the content that is most impacted by the upgrade, we can help to reduce the costs associated with testing after the upgrade has been completed.
Training is a vital part of the upgrade process. Not only do administrators need to be trained on supporting and maintaining the new system, users need to be trained on using new features and new tools. We can use Sherlock to determine which users are using each tool – properly and improperly. This allows us to determine how many Desktop Intelligence users may need to be trained on Web Intelligence, how many users may need to be trained on Analysis (if this is being deployed as part of the new BI4 deployment), how many users are downloading large amounts of data from reports to Excel, and many other interesting facts that help to shape a training plan. Similar to focusing our efforts for testing, we can use our Sherlock assessment as a way to focus training efforts as well. This ensures that we are not using resources to training users who do not need it.
The move to SAP BusinessObjects BI4 doesn’t end with the upgrade. The team of administrators must continue patching the deployment, handling user requests, deploying new tools, allocating new users, and ensuring that the deployment is running and stable. Using Sherlock, we can help with designing a maintenance plan that is based upon facts about how many users are using the deployment, how they are using the deployment, and how the deployment needs to change as new users are added and new functionality is being included as part of business analytics within an organization. By using Sherlock as a means of tracking the system on an on-going basis, we can help reduce the number of resources, the amount of time, and the amount of money used to manually monitor the deployment.
Where Do You Go From Here?
Interested in learning about how we use Sherlock to help our customers simplify the upgrade to BI4? You can join us on an upcoming webcast to hear how we plan for upgrades and ask us questions about the process.
Thanks for reading!