It’s come around again, as has happened so many times in the past, SAP has released (or even just announced) a new data warehousing product which has simultaneously excited and scared customers about its possibilities and implications. No doubt you’ve already heard of SAP Data Warehouse Cloud, since its shiny unveiling at Sapphire back in May; the internet – especially SAP Business Warehouse (BW) communities, has been buzzing. SAP has been drip-feeding the hungry masses with tidbits of information, so hopefully I can bring you up to speed and share my thoughts on the next-gen of data warehousing while I’m at it.
The Down-Low on when to Down-Load
With enterprise data ecosystems relying more and more on information stored within the cloud, it makes sense to have a Data Warehouse (DW) which also exists there. What doesn’t make sense, is to download and re-persist data from the cloud which already lives there, only to expose it back to the cloud via a Business Intelligence (BI) tool like SAP’s current flagship BI product SAP Analytics Cloud (SAC). It’s just logical to keep everything “up there” so to speak and do your data modelling virtually without having to spend time on ETL (extract, transform and load) which is taking up precious time, among other precious resources (storage, processing, etc), which could otherwise be better spent on getting data and valuable insights to information consumers.
Whilst this is all very well and seems logical for cloud-based data sources, what about the vast amounts of on-premise data? In its raw or even normalised extracted form, this seems like a lot of data that will need to be ETL’d up to the cloud. Thankfully the SAP strategy, at least at this point, isn’t suggesting this is needed, at least not immediately. SAP Data Warehouse Cloud is going to complement customers with an already established on-premise Data Warehouse, especially when connecting to cloud-based data sources, although they do suggest considering it for customers embarking on a new BI journey. I’m very interested to see if any organizations take this plunge.
Sizing - only bite off what you can chew, but you’re welcome to have seconds
Traditional on-premise EDWs, or any application for that matter, have many overheads that we’ve come to accept as the norm. Initial hardware sizing and dealing with the challenge of either starting and staying too big to ensure you don’t run out of space or being conservative to begin and then tackling how to grow, scale-out or scale-up? Not to mention the ever-present threat of the need for patching and upgrades and the associated stress, when do you plan downtime and how do you appease your users / developers? Then there’s the factor that every CIO will ask about first – cost. The idea of paying a premium for a system which is barely used is the stuff of nightmares.
Thankfully, one of the unique selling propositions of SAP Data Warehouse Cloud addresses these factors directly via elasticity of sizing. This enables the scaling up or down of system resources as dictated by usage requirements. Likewise, there is a pay-as-you-go pricing model depending on storage and computation usage. By only paying for what is utilized, the correlation between cost, user adoption, and outputs derived should be easily on display and help drive further investment to continue the cycle.
Embedded Agility without the Scrum
A common complaint I’ve often heard from consumers of BI applications, is how there seems to be a great chasm between their understanding of their requirements from their data and what is understood by I/T. Compounded to this is the time it can take for IT to prepare and then ETL the data into the system so that it can be then modelled into structures which can only then be queried or turned into reports and dashboards. Even then and after many meetings, designs and reviews, chances are, the result is still not exactly what the business users want. Engaging users in the process or *breathes deep* following Agile can help, but there’s always going to be a gap. What if business had direct access via an aesthetic interface which would enable power users to slice, dice, pivot and model their data into information that could easily provide the insights they desperately seek? Well, that’s exactly what the new concept of Spaces purports to do in SAP Data Warehouse Cloud. I say new, but it’s not, well not really, Workspaces have been around for a long time across different versions of SAP BW and Agile Data Preparation (ADP) also professes to assist with the same, but both products need to be added to a larger platform to be utilized and to be honest, I’ve not seen them to be very popular with customers. Spaces, on the other hand, comes out-of-the-
box cloud with this functionality ready and waiting to be embraced by users of all levels. Data can be collected, categorized and configured based on user requirements through the so-called Semantic Layer which provides business users the context and relationships between business artefacts without being overwhelmed by the technical details.
IT is not completely excluded though, in order to optimize performance for one Space and overall system performance, administrators are able to control persistence, performance (through memory and compute power allocations) and authorizations. Not to be left out of the data modelling fun, IT can also be responsible for developing template or base models which the business can use as a starting point for their own analysis or they can take the traditional role of creating data models and reports for users directly.
What happened to SAP BW/4HANA Cloud?
Prior to Sapphire, there were rumblings that SAP would be soon offering DWaaS (Data Warehousing-as-a-Service) under the guise of Project Blueberry. This wasn’t so much shrouded in secrecy as there was just not much information made available, what was shared however, was that it was to be a private cloud version of SAP BW/4HANA. Specifically noted was that it was to utilise the same ABAP code-line as BW/4HANA and this meant that what one could do on-premise would be synonymous to that in the cloud. Whilst I’ve not had confirmation of this from the horse’s mouth, it appears that SAP DW Cloud is instead a newly developed product based on different technology to the SAP BW products of old. Now this makes sense, as it appears much more in line with the similarly public cloud available SAC and is likely built on the same web technologies.
New world, same goal
So, in terms of providing “a single source of truth”, in a world with organizations utilizing a mix of traditional on on-premise data warehouses like SAP BW and the new SAP Data Warehouse Cloud, it seems like it might be a while before this goal is truly realised. Instead as SAP calls it, customers will have the ability to operate in a hybrid environment where they can receive the best of both worlds. This means utilizing cutting edge data warehousing with mature business content and advanced modelling with SAP BW/4HANA or Native HANA SQL Data Warehousing and on-premise data while simultaneously reaping the benefits offered by SAP Data Warehouse Cloud with cloud hosted data sources. Data from these two platforms can then be merged with on-premise data being remotely accessed and subsequently extended if further data modelling is required or even used directly in reports and dashboards within SAC or similar analytics tools. This is sure to be a massive reassurance for customers who have recently completed (or planned) an implementation of BW/4HANA that they’re not going to be left behind for a newer product. Don’t worry, SAP is continuing to invest in SAP BW/4HANA and SAP SQL Data Warehousing which will continue to be very strong options for on-premise and private cloud requirements. I’m sure the focus for cloud is only going to increase but this hybrid capability will allow customers to gradually transition to the cloud at a pace which suits them.
SAP Data Warehouse Cloud is set for general release in Q4 2019, so while there are still many questions about functionality and future direction, having seen the product road map overview, it appears that SAP is planning to try and address these and more in the very near future. I, for one, welcome this new cloud based direction and look forward to getting my hands dirty with SAP Data Warehouse Cloud, well as much as the business users let me..
(Images taken from SAP presentations)