In case you missed it, Comerit held our first webinar focused on SAP BW/4HANA this month, which you can check out here. Jesper took attendees through an introduction to this powerful new technology, concluding in a Q&A, re-capped in part below.
Is it possible to convert existing BEx queries?
The Bex queries that you have today will open up in the new query tool in Eclipse. There’s no conversion process for that. The queries are available as soon as you open the object up in the new query builder.
Can we access HANA calculation views using external ETL tools like informatica, SSIS without any special licensing requirement?
Yes, that is technically possible. I think from a licensing requirement perspective, that’s still something to work out with SAP. From the latest discussions I’ve had with SAP on this, there’s no requirement to have an additional license beyond the BW powered by HANA license to access calculation views directly from external sources. As long as it’s not an online user interaction with the HANA database, you can load the data to other applications. Again, this is something you’ll need to check in with your SAP account exec about to get the details of your license agreement.
Currently our dashboards are based on Bex queries. So, if we decide to go for BW/4 and convert out backend objects, then will we need to either convert existing BEx report (if possible) or build from scratch?
If you convert from HANA and use the transfer toolbox, it allows you to take the existing data flows. One of the things that’s being developed is a tool that moves the queries. We at Comerit have a tool that does that; there’s no need to rebuild everything, including your dashboards, with this tool. It takes an existing query and moves it between infoproviders.
Does the star schema still exist – fact tables, dimension tables, etc.?
It doesn’t really exist the way you have it with infocubes today. With these infocubes, if you’re not on BW powered by HANA, you have an extended star schema with dimension tables and SID tables linking out to the master data etc. Already with BW powered by HANA, that has been simplified. Essentially you are linking directly from a fact table to SID tables. That’s going away with SAP BW/4HANA as well. Now you’re linking from your DSO directly to the SID tables. Similar concept, but again, it is from an object type perspective. Since it’s a star schema, you can still have master data. You have the option of using info objects or not using them. If you have data you still want to use across models, using info objects is still best practice with SAP BW/4HANA for utilizing that same customer master data.
What is the difference in perspective of data modeling for current snapshot scenarios? What are the top changes regarding standard extractors? How do we define complex variables (BEx) in BW/4HANA? What standard extractors had issues in BW/4HANA?
If you want to do a snapshot scenario, you can still do that in SAP BW/4HANA. The modeling objects you’ll use will be slightly different, though. You can do that using an SDSO and then build a snapshot based on a query or based on a snapshot that you generate with the data loop, so it hasn’t really changed; you still have to build that into your model.
There’s not a lot of changes to standard extractors. Some of the new HANA or BW/4HANA optimizations do have some new extractors that they rely on, but the extractor development is more related to the version you have in your ECC system. So, if you are on SAP S/4HANA, most of the extractors are set that you have and have been using up until now. They haven’t changed a lot. What they are doing with those extractors is that the underline tables have changed, so SAP is kind of shielding you from that by developing CS views that are being used by the extractors instead, but the extractor itself will still operate the same.
There’s a list of a few extractors that are not supported – mainly the old LIS extractors are not supported with BW/4HANA; it’s more tied in with the ODP framework of what is supported with BW/4HANA and what is not supported as all the extraction processes from SAP systems is going by ODP and not just straight to the service API.
Is BCS supported?
BCS is not supported by BW4HANA and it’s not in the roadmap. If you’re relying on that, you will not be able to convert your system anytime soon.
In your experience, do you know how much performance degradation is going to happen if you plan to use what you call the data temperature management modeling?
We have done that with a few clients, and what we’re seeing is that when comparing to MMS, it’s about 2-3x’s faster. It depends on how over-subscribed you are on your system. You can oversubscribe your extension by 200%, which means that it’s likely that you won’t have the data in memory at any point in time, and it’s likely that it has to load it from this. But that process has been improved in the latest support packages quite rapidly and we’re seeing that it’s much faster now than with SAP IQ. You also don’t have to manage any secondary database, which has been a problem with a lot of companies.