Gee, S/4HANA. Just what I always wanted: Customers are wary of what's in SAP's sack The lure of shiny new things is particularly irrepressible in December as Christmas approaches, but SAP customers seem to be able to resist it. Rob Moore, business intelligence systems manager at industrial engineering firm Edwards, for example, says that he would love to unwrap S/4HANA, the major upgrade to SAP's ERP platform …
I find it astounding that companies are willing to write off decades of SAP ERP investment to start over just so that they can make calculations in columns. I mean, they've been managing so far, what's the problem ?
Honestly, if my company was big enough to need SAP and I had a working system, I'd hate to budget for an entirely new system just because of a new calculation method. I know IT is all about redoing stuff, but this is pushing things a bit far.
I hope there's some other advantage that justifies spending all that money all over again.
"less DB licensing costs (vs. Oracle)"
Yeah, I don't think that is true. I have been involved in a situation where a customer looked at HANA. It was a fairly sizable environment but SAP asked for $20 million in HANA licenses, which many times their existing DB costs. I don't think HANA is less expensive than Oracle at any point, but certainly not less expensive than Oracle when you factor in that you have to buy new HANA licenses and everyone already owns the Oracle/DB2 licenses. You also have to buy servers/cloud services with a boat load of memory for HANA, also millions of dollars/pounds.
You need to read up on what column store databases do. I'm a BI consultant so I'm mostly concerned about getting data out. Column store is astonishingly fast - think 30 minute queries down to 5 seconds.
That said, MS SQL Server has column store that's 99% as good, so why pay more?
I quite like MS SQL column store but HANA moves the game on further. In my experience that equates to better performance (although start-up times suffer).
Whether or not that gain justifies changing is an exercise for the reader.
I think the other gain for existing SAP customers with BI is that BI and online can run against the same DB with HANA so you no longer need an extra copy for BI.
True, Oracle has columnar, MS SQL has columnar, everything has columnar. Likewise with in memory, Oracle has that now too. There is no real advantage and HANA is missing many of the HA/DR features of an Oracle DB.
There is only one reason to do it - because SAP says so. I think they will back off on the end of support date as I don't seem them wanting to lose billions of dollars in highly profitable support and pushing people to Oracle ERP, but who knows.
I suppose a more realistic use case, based on personal experience would be a 3 second query down to 8ms, which is nice to have, but probably not worth the cost / hassle.
Also, if you are migrating to a new system, you consider all options, as staying with SAP is no longer the easy choice.
You don't have to upgrade (in reality, migrate/reimplement) on to S/4HANA in order to use the HANA in-memory column store database on old-school SAP ERP. We are running it right now: all the many years of legacy ERP enhancements and bespoke code running many, many times faster with the same code base. Not for everything though as not all latencies are due to database IO.
As I understand it, S/4HANA is the in-memory column store database with redesigned ERP code on top to make the most of it.
Many companies hack SAP standard and make it work like Oracle, or some other ancient legacy system. To them, even upgrading ECC becomes a challenge, moving to S4/Hana seems impossible.
It is true that SAP have thrown out some poor design choices with new design of S4 relative to ECC, and some things don't have a clean migration path or even work yet, but at the end of the day ECC had reached a complexity point that there was no way forward for it.. It was just too twisted and ugly on the inside and if you look at SAP's own code you want to go blind, it's shockingly bad in places. Starting a fresh with a newer faster design is bold, don't know if it's the right thing to do, but they had to do something...
I write SAP code for a living and companies would do themselves a lot of favours if they stop buying it and then re-writing it..SAP allow you to tweak it, but far too many rewrite core bits and then wonder what to do next! If you want a bespoke system, write one, if you want an off-the-shelf solution then buy one, don't buy a dog and bark...!
Where I work, we have heavily customized one of our ERP tools. We're now finding that this was a bad thing to do, as every tiny upgrade needs a lot of work to re-implement our hacks customisations. (I'm sure you can imagine the pain we have for larger version upgrades)
Senior management are taking a surprisingly sane view. We're going out to market to look for a new system and insisting that we won't customise it to death.
And if a company spent $/£/€10m implementing SAP, it probably took 10+% of that to keep it running and up-to-date with ECC. Most companies didn't bother or only did it for a few years before firing those with the knowledge.
S4/Hana will require a similar investment in SAP/business re-engineering and most businesses aren't prepared to do that. Those companies that are prepared to invest tend to have significant political battles for what S4 and the business will look like post-implementation. External factors around long term company strategies (i.e. international trade/relations) don't help.
A big problem for SAP is that the initial SAP implementation many years ago was a failed project... or at least millions over budget. Companies were past the point of no return so they just had to deal with it, but many probably would have just stayed on whatever they had before SAP if they could do it all over again.... Now SAP wants them to completely reimplement SAP and pay for SAP again with little benefit to the customer.
I used to implement SAP. That problem can be laid directly at SAP's feet by promising that SAP would be everything to everyone and that customization is easy... or that their SI partner had templates which would make things quick and easy. That never happened and then people had to customize it to get it to do what SAP said it would do, or at least a portion of what they said it would do.
ECC is convuluted, but it seems odd that SAP made a mess of their existing architecture so they need everyone to reimplement, and pay SAP for new licenses.... and they still don't have SaaS, but will eventually need to have SaaS if they want to remain relevant, so that will probably be another reimplementation in a few years if you implement S/4 on prem or on IaaS now.
I think these comments are missing the forest for the trees. Yes, we over customized the original implementation, but that was required to make it work for the business in 2000. ECC had deficiencies and yes, the tech underneath it has changed. But that is not what's justifying the S4/HANA migration, at least where I work.
The way the business runs has changed since 2000 and there is too much legacy business process debt built into ECC for it to keep up and keeps the lights on, as there would be in any 20 year old ERP. Digital supply chain, Enterprise PLM driving shop floor operations, and Day One acquisition financials integration are what the business wants now: Each of which would be hard to incorporate into a 20 year old monolithic system, and together justify the sea-change.
And it doesn't have to be that disruptive, at least for the Tech folks if you crawl, walk, run. We started with BW to get our feet wet with S4/Hana. Then went to CFIN with a Enterprise side-car. Now planning the build out on that, while ECC keeps chugging. Doing it in the cloud just accelerated the machine builds because there we no PARs to sign, but the details didn't change too much. The cultural change in digital transformation is MUCH harder to achieve and the real challenge here rather than the underlying tech.
Just an opinion, and YMMV, but the business is changing radically around us and we are going to have to change with it.
I am fairly sure, after having managed a number of different ERP solutions that they are bristling with interfaces, batch processing, and other e-bondings that moving them to the cloud - even to the same version - would be an incalculable exercise in its own right - with zero added business value - or potentially negative if you have a good team running the environment already.
The upgrade itself, once you are on the cloud is "hidden" therefore what would have been assessed under internal change controls happens under cloud "magic" and therefore can massively increase the test complexity, at best, and at worst, be devastating to your business (TSB style)…
Have SAP lost sight of the sort of businesses they supply to? If you are going for a full system replacement type project, I would also be asking... should the solution be SAP again, or are there alternatives?
Ack, TDI (tailored datacentre infrastructure) is the search term. For quite a few things you´ll still want to have a cloud connector.
And, a free advice: scale up as long as you can, scale out is a real pain. hardware is cheap.
They are not forcing you to the cloud now... because their SaaS strategy is all over the board. They will be forcing you to their S/4 SaaS version as soon as you complete the S/4 upgrade on prem or IaaS.
"zero added business value - or potentially negative if you have a good team running the environment already"
But watch as your bosses look at the balance sheets, and realise how much money they (think they) will save by firing that "good team" of your, and shoving everything in the cloud.
Sure, it'll cost more in the long run, but by then you'll have been sacked, and those responsible will have got a bonus.
If you force a business to consider spending 7 to 8 digits redesigning their end to end processes in a significant way most large organisations will also start considering whether SAP or a competitor is the best platform to support their business going forward.
Yes, I think if Oracle picks off a handful of blue chip SAP customers, then SAP will drop this ECC end of support deadline.... SAP should be careful not to completely overplay their hand. They think that they can do whatever they want because switching to a new ERP is a huge project... but SAP is forcing a huge project on people regardless. Not like moving from ECC to S4 HANA is any easier or less costly than moving from ECC to Oracle Cloud.
To be totally blunt S/4 does not need HANA DB, other vendors sell products which can handle the in memory technology required to run S/4
This is a commercial decision by SAP to push out competition by using dirty tactics
S/4 is also not a full rewrite, it carries forward some of the poor choices and problems from it's ECC and R/3 days
Frustratingly some of the key things which are needed by ECC customers will never be downported because SAP want more money and so wish to sell (force) their customers to pay for something 'new'
Other vendors are behaving like it is 2020 whereas SAP are treating their customers like it is 1990
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
