The BW - HANA Relationship
转自:http://weblogs.sdn.sap.com/cs/blank/view/wlg/21575
Thomas Zurek
Business Card
Company: SAP AG
Posted on Oct. 15, 2010 01:11 AM in In-Memory Business Data Management, ERP, Enterprise Data Warehousing/Business Warehouse, BI Accelerator
With the announcement of HANA*, some customers, analysts and others have raised the question on how HANA relates to BW with a few of them even adding their own, home made answer in the sense that they speculate that HANA would succeed BW. In this blog, I like to throw in some food for thought on this.
Currently, HANA's predominant value propositions are
· (i) extremely good performance for any type of workload
· (ii) a real-time replication mechanism between an operational system (like SAP ERP) and HANA
Let's match those for a moment with the original motivation for building up a decision-support system (DSS) or data warehouse (DW). In the 1990s, a typical list of arguments in favour of such a system looked like this:
- Take load off operational systems.
- Provide data models that are more suitable and efficient for analysis and reporting.
- Integrate and harmonize various data sources.
- Historize - store a longer history of data (e.g. for compliance reasons), thereby relieving OLTPs from that task and the related data volumes.
- Perform data quality mechanisms.
- Secure OLTPs from unauthorized access.
Installing a DW is typically motivated by a subset or all of those reasons. There is a particular sweet spot in that area, namely a DW (e.g. an SAP BW) set up for reasons 1. and 2., but with all the other arguments not being relevant as it is connected to basically one** operational system (like SAP ERP). Here, no data has to be integrated and harmonized, meaning that the "T-portion" in ETL or ELT is void and thus that we are down to extraction-load (EL) which, in turn, is ideally done real-time. So value proposition (ii) comes along very handy. Critics will argue that such systems are no real data warehouses ... and I agree. But this is merely academic as such systems do exist and are a fact. So, in summary, there is a good case for a certain subset of "data warehouses" (or reporting systems) that can be now built based on HANA with (i) and (ii) as excelling properties - see the top left scenario in figure 1 below.
Now, will this replace some BWs? Yes, certainly. In the light of HANA, a BW with a 1:1 connection to an ERP might not be the best way anymore. However, will this make BW obsolete in general? No, of course not. As indicated above: there is a huge case out there for data warehouses that integrate data from many heterogenous sources. Even if those sources are all SAP - e.g. a system landscape of multiple unharmonized ERPs, e.g. originating from regional structures, mergers and acquisitions - then this still requires conceptual layers that integrate, harmonize and consolidate huge volumes of data in a controlled fashion. See Jürgen Haupt's blog on LSA for a more comprehensive discussion of such an approach. I sometimes compare the data delivered to a data warehouse with timber delivered to a furniture factory: it is raw, basic material that needs to get refined in various stages depending on the type of furniture you want to produce - shelves might require less steps (i.e. "layers") than a cupboard.
Finally, I believe that there is an excellent case for building a BW on top of HANA, i.e. to combine both - see the bottom right scenario in the figure below. HANA can be seen as an evolution of BWA and, as such, this combination has already proven to be extremely successful: BW 7.0 and BWA 7.0 have been in the market for about 5 years, BW 7.30 and BWA 7.20 have pushed the topic even further albeit mainly focusing on the analytic layer of BW (in contrast to the DW layer). When you continue this line of thought and when you assume that HANA is not only BWA but also able to comply with primary storage requirements (ACID etc.) then the huge potential opens up to support, for example,
· integrated planning (BW-IP): atomic planning operators (used in planning functions) can be implemented natively inside HANA, thereby benefitting from the scalability and performance as seen with BWA and OLAP and also from avoiding to transport huge volumes of data from a DB server to an application server,
· data store objects (DSOs): one can think of implementing such an object natively (maybe as a special type of table) in HANA, thereby accelerating performance critical operations such as the data activation.
This is just a flavour of what is possible. So, overall, there is 4 potential and interesting HANA-based scenarios that I see and that are summarized in figure 1. I believe that HANA is great technology that will only come to shine if the apps exploit it properly. SAP, as the business application company, has a huge opportunity to create those apps. BW (as a DW app) is one example which has started quite some time ago on this path. So the question on the BW-HANA relationship has an obvious answer.
* High Performance ANalytic Appliance
** The case remains valid even if there are a few supporting data feeds, e.g. from small complementary sources.
Figure 1: How HANA can potentially evolve in existing SAP system landscapes ... in a non-disruptive way.
· HANA - What is different from BWA
2010-12-17 20:52:52 Sanjay Sinha Business Card [Reply]
- Hana is an appliance but in all the above figures I see that HANA is going to eventually become the underlying database ( In Memory ).
- In BWA technology failover is DB what will be failover in HANA
- How long will it take to make HANA stable and robust enough like an RDBMS today.
- Will the in memory technology help even if the data model is bad.
- Data Life cycle management - How will the archiving work?( will it require another HANA for backup or archiving )
- Disaster Recovery -> Is it ready.
So far whatever has been advertised about Hana, it does not seem to do any other function other than accelerating data.
What is so different about HANA and BWA ( apart from the blades and software and OLAP + OLTP capability )
o HANA - What is different from BWA
2010-12-17 23:31:14 Vitaliy Rudnytskiy Business Card [Reply]
Sanjay, ideally questions should be posted at the forum: http://forums.sdn.sap.com/forum.jspa?forumID=491
Some of the answers where already covered, so I put just links there.
- Hana is an appliance but in all the above figures I see that HANA is going to eventually become the underlying database ( In Memory ).
VR: Please see http://bit.ly/eUkzrO for distinction between HANA and its database
- In BWA technology failover is DB what will be failover in HANA
VR: In BWA failover is to another blade; same is going to be with HANA - fail-over to another node (rack or blade - depending on the used)
- How long will it take to make HANA stable and robust enough like an RDBMS today.
VR: HANA is built on existing technologies: MaxDB, TREX, P*Time, Sybase and inherit many of the existing features. Nevertheless it is a new product, so let's see.
- Will the in memory technology help even if the data model is bad.
VR: Please have a look at point a) at http://bit.ly/fAWeYh
- Data Life cycle management - How will the archiving work?( will it require another HANA for backup or archiving )
VR: Future release of HANA should include data aging functionality: hot data - in RAM, cold data - on disk
- Disaster Recovery -> Is it ready.
VR: You mean ready for 20 Dec 2012? ;-))) http://www.nasa.gov/topics/earth/features/2012.html
§ HANA - What is different from BWA
2010-12-17 23:44:29 Sanjay Sinha Business Card [Reply]
Hi Vitality,
You are Guru. I was not expecting the crisp and solid answers. Thank you very much. I am a newbi so far HANA is concerened. But will keep a watch. I read your other blog and so nicely written from context to content. Hats off to you.
Sanjay
§ HANA - What is different from BWA
2010-12-17 23:57:18 Vitaliy Rudnytskiy Business Card [Reply]
Hi Sanjay. If I am Guru, then the end of the world is closer than I expected ;-))
Thanks anyhow for reading my blogs and sharing your positive experience.
Good night. -Vitaliy
· HANA makes OLAP and OLTP relationship stronger
2010-11-11 11:03:02 Vitaliy Rudnytskiy Business Card [Reply]
Hi Raghav. Sorry, I did not quite get that. Could you please elaborate a bit more on your thoughts what specific scenarios, specific capabilities and technologies will support this statement? Greatly appreciate. -Vitaliy
o HANA makes OLAP and OLTP relationship stronger
2010-11-12 00:38:50 Thomas Zurek Business Card [Reply]
Hi Raghav,
just one comment regarding unstructured data: the search engine will not be part of the HANA 1.0 delivery. But similar to BWA, most of the infrastructure is there and HANA will be enhanced accordingly in future versions to fill out a role as a search engine (e.g. as infrastructure for our Enterprise Search product) or as a BWA. This is, by the way, also the reason to look at HANA not only as a DB but as a multi-purpose technology piece. However and as previously stated, it is necessary that you have the appropriate application that leverages that piece. BW is one example, ES will be another one. Alternatively, you can build your own, simple ones with the given tools. See the ERP-HANA side-by-side scenario.
Regards. Thomas.
PS: Interestingly, I cannot see Rghav's comments via my browser but have received them via email.
§ HANA makes OLAP and OLTP relationship stronger
2010-11-12 08:57:24 Raghav Jandhyala Business Card [Reply]
<My original message. I had to pull back to add details.>
BI will still be a platform for both
unstructured and structured data and tools like data services (text
analysis) along with HANA 's (TRex seach capabilities) will provide
unstructured data extraction and search.Unstructured data is where we can see more and more customers spend thier dollars.
It is not suitable to store mergers data, historical data and
unstructured data in your ECC system.
Also with HANA being a platform where analytic and transactional
applications will be able to run in a single environment, we would
likely see scenarios in which Operational and analytics data being part of a business process, example in planning. With HANA,
real time operational and analytic reports can be embedded into your applications (embedded analytics).
· Great read...some questions..
2010-11-04 09:30:15 Michael Bestvina Business Card [Reply]
Thomas,
Thanks for this. A couple of comments/questions:
1. "I believe that HANA is great technology that will only come to shine if the apps exploit it properly." - Absolutely agree, just hope SAP sales really understand this ;) How is it possible then to separate out what business scenario does HANA supply and which does BW supply? If "(i) extremely good performance for any type of workload" is true, then why doesn't it replace every scenario which BW is used? I do believe value prop (ii) is a very valid case of which no current SAP solutions exists (RTA and hourly loads aren't real-time).
Maybe a value-prop matrix might make sense: (correct me if I'm wrong here)
Value prop (i): Evaluate whether HANA or BW is needed based on business logic requirements.
Value prop (i+ii): Use HANA
Value prop (ii): Use HANA (but restrictive in the sense that you only see the individual business tables..i.e. no extra logic)
2. I'm trying to understand where HANA might replace BW ... Let's say you have a current BW system and one of your business areas is FI-CO (vanilla FI-CO data extracted from ERP and using BI CONT to bring that data to a reporting layer). Would a HANA appliance be a suitable replacement for such a scenario? If so, what are you trying to offer as a lowered TCO? Since this scenario is very vanilla (read-> relatively easy to implement) then wouldn't installing a BWA in the existing structure be much easier and satisfy all requirements (functional and performance). Plus gives you the flexibility in the future to add some additional business logic you want later...OR is this what you mean by putting BW on top of HANA, so that if you wanted to extend something later you have the ability to do so via BW?
3. Since BWA is not ACID compliant (AFAIK), what is different in HANA that makes it ACID compliant? Maybe too techie of a question for a simple answer ;)
4. I'm not sure why this hasn't been mentioned in discussion yet, but what happens if the appliance turns off? Isn't data stored in RAM?
danke!
-Michael Bestvina
o Great read...some questions..
2010-12-16 08:26:54 Tansu Aksu Business Card [Reply]
All,
About HANA replacing BW - I think people are not worrying about losing BW, but rather losing its capabilities that can't be found in ERP systems.
Maybe we should list them for HANA architects to consider developing them for future releases.
There are some obvious reasons that we prefer having BW such as
* extracting and consolidating data from different source systems
* one master data/one truth
* be able to cleans and/or manipulate data while extracting data in user exits, etc.
But what else? Here are some minor but important things to be considered
1) Time-dependent master data reports
2) Non-cumulative key figures/cubes-now ERP systems have to handle this
3) Information broadcasting/data bursting reports?
Tansu
o Great read...some questions..
2010-11-05 06:34:29 Thomas Zurek Business Card [Reply]
Hi Michael,
I try a brief answer - you have raised many good points and they are all worth a comprehensive discussion.
Ad 1.: I think Tansu has provided some good points here.
Ad 2.: You are already speculating in the right direction.
Ad 3.+4.: In-mem databases store the data also on disk. That provides the "D" (durability) in "ACID". The idea is that RAM is leading and holds the data in its original version while disk keeps snapshots + logs (to recreate the recent state from a snapshot). In traditional RDBMS it's the other way round: RAM caches "blocks" ("pages") of data, basically as a 1:1 copy from disk. So the disk structure is leading.
The term "in-memory" has been heavily used by marketing and sometimes oversimplifies things. One group argue that the disk-based RDBMS use huge RAM caches (so they are in-memory too) while the other group says that in-memory DBMS have disks which comes sometimes as a surprise. Wikipedia has a good introduction on this.
I hope this helps.
Thomas
§ Great read...some questions..
2010-11-05 06:44:25 Michael Bestvina Business Card [Reply]
Thomas/Tansu,
Thanks for the information. I'll carry the discussion on with some of your dev colleagues in the next comings weeks. It will be important to completely understand this from the perspective of sales, value prop, support and consulting.
Ad3+4: Thats what I speculated, but wasn't completely sure. This is where MaxDB fits in then right?
@Tansu - Long time no see ;)
-Michael
o Great read...some questions..
2010-11-04 10:06:42 Tansu Aksu Business Card [Reply]
Micheal,
Based on the limited information available about HANA today, I tend to agree with you that "BW+BWA" is more flexible solution for performance problems unless you have to have real time data.
Let's use your FI-CO example; as we know almost nobody uses SAP delivered content 100%, so we make some custom changes in user exits, start routines, DTPs, FMs etc. Personally I prefer to customize the reporting system(BW) then the transactional system(ERP).
Another reason to keep the "BW + BWA solution" is, it is easier to bring data from all other non-SAP ERP environments to BW then to ERP.
For these 'practical' reasonons I believe the best solution will be "BW 7.3 + BWA 7.2". With BW 7.3 OLAP will be accelerated, DSOs, master data reporting will be supported by BWA etc. But I am very optimistic and excited to see what HANA 1.5 (for BW) will offer.
Lastly, when it comes to creating "practical" business cases for HANA 1.0 (except providing real-time data), wouldn't it be unbelievable to have ad-hoc reports on top of CDPOS and CDHDR tables(ERP change logs)?
Cheers
Tansu
· So how does it work with Business Suite?
2010-11-02 19:57:09 John D Business Card [Reply]
This is probably covered somewhere amongst the recent blogs about In-Memory and HANA, but I couldn't clearly pick it out.
I don't have much experience with the newer flavours of BI/BW yet, so please forgive my noob questions :)
If I was a customer just running ERP Financials/HR and maybe SRM could I use HANA as the platform for my analytical reporting, instead of BW?
In addition to the appliance, what are my reporting toolset options? Business Objects only? If so, which ones will work with HANA?
And lastly, when will all this actually be available to customers?
o So how does it work with Business Suite?
2010-11-03 00:32:30 Vitaliy Rudnytskiy Business Card [Reply]
Hi John. I just have published another blog "Important Questions with Answers about SAP's High-performance Analytic Appliance HANA", as a summary of my last weeks discussions.
http://www.sdn.sap.com/irj/scn/weblogs?blog=/pub/wlg/21589
Hope you find something useful there as well. Cheers, -Vitaily
§ So how does it work with Business Suite?
2010-11-03 15:23:23 John D Business Card [Reply]
Thanks Vitaliy, I just came across your latest blog while trying to find this one again. Yes, it does make things clearer, so thanks for taking the time to write it up.
I guess I will have to wait and see what analytical tools eventually become available for HANA, probably a bit early yet for SAP to release that info.
· an opinion about HANA replacing BW?
2010-11-02 14:04:15 Tansu Aksu Business Card [Reply]
I found the concerns/thoughts that "HANA will replace BW" similar to "creating a reporting system on top of PSA only". It's technically possible, but would you do it?
Tansu
· cost of innovation
2010-10-18 14:20:35 Vijay Vijayasankar Business Card [Reply]
Blades are not cheap - so I wonder if a significant portion of the SAP instalbase can make use of HANA. Sure the top 100 or top 500 might have the CAPEX budget, but what about the rest?
Also, why would I need to maintain BW and HANA in future. Why shouldn't I expect HANA to do everything that BW does (and then some), so that I don't need to keep multiple systems for similar purposes? If BW cannot be phased out some time in mid term, I wonder if HANA will delver on its full potential.
o cost of innovation
2010-10-19 00:21:52 Thomas Zurek Business Card [Reply]
Vijay,
w.r.t. comment on phasing out BW via HANA: HANA is basically a DB, i.e. pure technology. BW is a DW management application sitting on top of a DB. So it's apples and oranges. Likewise, a pure Teradata installation is not a DW as you have to complement it with programs and tools that manage semantics, data containers, mappings, processes, consistency mechanisms etc.
There is a good use case for replacing BW when most of this constitutes an overkill - see Vitaliy's comment or the paragraph in the blog on operational data marts.
Regards
Thomas
o cost of innovation
2010-10-18 14:33:10 Vitaliy Rudnytskiy Business Card [Reply]
Hi Vijay. SAP HANA in the first version is not targeted to BW customers, but to ERP customers without BW (left top quadrant on Thomas's chart) as an "real-time" "fast-performing"
analytics with "agile modeling" for business users. As such the value proposition is really appealing, and the question is exactly the one you raised - what would be the ROI of this solution. -VR
· On "agile data marts"
2010-10-18 01:27:02 Mark Foerster Business Card [Reply]
Hello Thomas,
thanks for the blog, it helps understanding what HANA can do or in the near future will be able to do. However, I am really wondering who on earth
would implement this "agile data marts" scenario.
I know many BW customers who are in desperate need of a BWA (because of horrible BW query response times), but refrain from buying a BWA. And of course there is only one reason not to
buy a BWA. So I wonder who on earth has the money to buy a BWA and HANA?
Maybe time will fix this. Once SSDs are getting cheaper, the reasonable way to proceed might be
to put the "X DB" on SSD until SAP makes the BWA (or HANA) affordable. Not my preferred solution,
but economics might dictate this way.
Regards,
Mark
o On "agile data marts"
2010-10-18 04:05:24 Thomas Zurek Business Card [Reply]
Hi Mark,
putting price considerations aside: the "agile data mart" scenario does exist today: replace HANA by any commercial RDBMS. Actually, they exist all over the place. The motivation frequently is that such an "agile data mart" is controlled by a business department rather than IT. The latter needs to run the BW (EDW) along rules like SLAs, compliance, company-wide consistency etc. and are therefore less flexible to implement or change models in an agile way. However, a marketing department (that decides to initiate a new campaign today) might want to create a cube to track that campaign and that might be 3 weeks down the road.
So, in summary, an "agile data mart" is a place that technically might run a similar workload to the EDW but that is managed in a significantly different way.
Also, you can look at HANA in that particular context as a competitor to an arbitrary RDBMS.
You are certainly right that there is always a price component to this as TCO increases. Obviously, the benefits (of the gained agility) need to outweigh that additional TCO.
Regards
Thomas
· Time context
2010-10-15 18:05:42 Vitaliy Rudnytskiy Business Card [Reply]
Hello, Thomas.
Would be great to add time dimension to all these HANA deployment options: when which is planned to become available.
Regards,
-Vitaliy
o Time context
2010-10-16 01:29:00 Thomas Zurek Business Card [Reply]
Hi Vitaliy,
the side-by-side scenarios will work with v1 of HANA, due end of this year. The other two are further ahead but have already been publicly mentioned at SAPPHIRE and Teched.
The above is no official roadmap but simply intending to put stuff into a logical context independent of time.
I hope this helps at least a bit.
Thomas
· One more AND
2010-10-15 05:08:25 Bala Prabahar Business Card [Reply]
This blog certainly answers the scope of HANA and BW. And thank you for this blog. Your blog supports there is a time and a place for everything.
What are your thoughts on stories that either HANA or HANA type DB(in-memory) replacing OLTP systems thereby replacing BW systems? I even heard stories that-in order to deliver this solution in non-disruptive way(buzz word these days)-SAP will come up with some kind of an abstraction layer which could distinguish a SQL statement written for traditional RDBMS from one written for columnar in-memory db. And this abstraction layer would-according to those stories-automagically convert RDBMS-SQL statements to columnar in-memory DB statements so no migration to columnar in-memory DB would be required(great vision;since I'm a little bit cynical type, would this be cost-effective?, I'm wondering. Those stories also suggest in-memory DB would eventually replace BW systems. Your thoughts?
Thanks once again for writing this blog validating my thoughts on the scope of HANA and BW.
Bala
o One more AND
2010-10-15 07:17:54 Thomas Zurek Business Card [Reply]
Hi Bala,
I find it hard to follow some of the rumours you cite above. Do they make sense to you? Maybe I'm simply lacking the wider context in which they were sifted.
Thomas
§ One more AND
2010-10-15 11:22:23 Bala Prabahar Business Card [Reply]
Hi Thomas,
No, they don't make sense to me. I read those rumors in sdn.sap.com(blogs, comments to blogs etc). They appeared sometime after Sapphire '10. Since I didn't attend Sapphire '10, I didn't know if they were based on SAP's statements or not. At any rate, you've answered my question. Thank you very much.
Best regards,
Bala