Geoff Armstrong sums up what he gleaned from an insightful podcast on HANA from three SAP mentors.
If you have an hour to understand what is happening with HANA from the perspective of three SAP Mentors, then have a listen to this podcast from Jon Reed.
I found this very useful, and a summary of the podcast is below:
- The vision of in-memory looks good. Even though they were not the first to do it, SAP seems to have created a high level of visibility for this technology.
- There have been some changes in the naming of the product set which probably reflects a changing position on how widely it can be applied.
- At the moment we are back to the world of tables and joins. So having a strong understanding of the ERP processes and data models is essential.
- HANA has the ability to streamline the table structures required (e.g. FI is supported by 25-30 tables in SAP, but in HANA perhaps we only need 2-3 to deliver the analytics). This could potentially simplify the data modelling.
- Once SAP delivers some basic content (e.g. data model for procurement) and a rapid deployment approach, clients will be able to implement much more quickly.
- Applications designed specifically for HANA e.g. CO PA will deliver flow-on benefits for clients by speeding up the ERP processing.
- SAP needs to assist the community to get hands-on experience and support this with appropriate training courses as there is significant up-skilling required. It would be useful for early adopters if SAP develops ‘how-to guides’ on SDN.
- HANA in the cloud will work but the question is what type of application would work effectively in the cloud. For example, sales and operations planning seems more like an on-premise scenario due to network limitations, security and legal restrictions
- Currently the marketing has been very successful with the business PR. We now need to see how successful the technical PR is. Additionally, licensing may be a PR issue depending on the take-up and value achieved by clients.
- At the moment the scenarios attempted should be fairly standard. Still waiting to see when HANA will be able to deal with complex modelling scenarios.
- The latest service pack stacks seem to be coming out rapidly. Does this suggest that it was rushed to market? Implication is you need to plan carefully the timing of a HANA implementation.
- HANA made it to GA before BOBJ 4.0, reflecting new product vs. an established product rollout. Expectations of what can be achieved in the early versions need to be realistic.
Geoff Armstrong is National Solution Lead – BI and EPM at PLAUT IT Australia.
