Pros and Cons of Sap HANA vs MS SQL Server as the DBMS for a DWH for SAP BO

[Former Member]

Hey guys

I work for a mid sized company (170 employees). We currently have some basic reporting going on (WebI, Crystal Reports) straight on our ERP database (MS Navision). We now want to start dashboarding using Lumira 2.0, but since we currently don't have a dedicated DWH, we figured implementing one is the first step.

We originally had some plans for a DWH based on SAP HANA because that's what our consultants told us would be the best fit. Problem for us is: We live in a world of Microsoft. We have no experience with Linux and the only dbms we actually have experience with is MS SQL Server.

What I know is: HANA is an column based in-memory dbms.

What I also know: MS SQL Server also supports a column based design and in-memory technology.

For someone without any deeper knowledge, that basically sounds like the same thing and the obvious decision would be to just stick with MS SQL Server.

So, what I want to know: Why would we prefer HANA as the dbms for a DWH instead of MS SQL Server? How does HANA work technically and performance wise compared to MS SQL Server? I need some background knowledge and arguments to discuss and convice our management to invest in a HANA based DWH.

Best regards and a big thanks in advance,

Daniel

[lbreddemann]

Ok, you have no in-depth technical know-how around DBMS but want a comparative (ideally independent) discussion about differences in two specific products with a tie back to your actual use cases? Honestly, that's not what you get for free. Either you make yourself familiar with these products on a much deeper level than powerpoint and documentation or get somebody to walk you through these topics.

Trying out the product that you don't know can be a great way to check how you go with it and if you can find added value for your organisation using that tool. With HANA Express (https://www.sap.com/developer/topics/sap-hana-express.html) you can have a free HANA instance to work with.

SAP HANA's main benefit is not a single feature that other DBMS don't have. Instead, the combination of many different tools and functions, that are commonly separate (build in machine learning, graph engine, data federation, text analytics, spatial processing, ...) and a general design that has been made with current hardware in mind (lots of processors, huge main memory, NUMA architecture, NVM access, multi-level cache structure with high speed memory access) delivers the great performance that this current hardware promises.

Specific to data warehousing I think, one major benefit of HANA is that it allows building virtual data models (based on views) that don't require storing data (unlike snapshot/materialised views) and allow to model different DWH layers without copying data around. This reduces the time until data is actually available in reports, sometimes drastically. This approach also opens the option to integrate "live"/"real-time" data and mix it with already stored data in a single consistent view, that can be easily consumed by clients like Lumira.

Finally, for this answer, I'd say that since many things you do with HANA are just really fast, without any additional effort in tuning/indexing/etc., as a developer you get much quicker to try out things. Adding trend-analysis to your report? Doing cluster analysis based on GIS input? Allowing 3rd party data subscribers via ODATA? All this is really easy and straightforward to implement with HANA. This community, the SAP HANA Academy (youtube channel) and the openSAP.com HANA development courses (free moocs) give many templates and show in detail how things can be done.

In short, I think HANA can enable your team to deliver more and better solutions to your customers than what you do today.

If that's not relevant for your company, then I'd say, HANA probably won't add much value.