With HANA, SAP looks like it is walking on yet another path of failed strategy. Customers really don't care about which database their cloud enterprise application is runs on. Yes, HANA is able to crunch operations in milliseconds what could take days. And yes, in-memory technology is here to stay. Yes, customers care about such performance. But what SAP is missing is that in-memory technology is already the norm in the cloud and several cloud vendors are already working on adopting such start using in-memory technologies. Customers don't want to hear about what database is being used underneath their cloud subscriptions. The competitive cloud vendor will offer an application where data layers, operations are abstracted and the application is able to perform operations using in-memory technologies when appropriate. What SAP is offering is a technology that is a non-virtualized, non-multitenant, non-scalable, and far from being cost effective solution. An application will run either on Hana or traditional db even if your operation is transactional in nature. SAP provides a non-multitenant instance dedicated for a customer. Unlike Hana, newer in-memory technologies such as Amazon's red-shift, Platfora etc are in-memory tech are multi-tenant that scale based on the needs, re-distribute data to nodes as needed and have concepts of multi-node parallelism. Ultimately, customers will pick cloud vendors who not only meet their application use cases but also perform the fastest at the lowest cost without caring about which technology or products are used underneath. SalesForce, for example, talks little about how they use Oracle underneath to provide a solution to their customers. Salesforce can change their database from Oracle to PostgreSQL tomorrow. Or Salesforce can use Terradata's in-memory solution or combination of both. The point is, the customer will go after the cloud application that meets their usage and performance needs without having to deal with licensing, extra costs. Vendors like SAP, Salesforce, Netsuite have to worry about how they can provide all of this in a cost-effective manner.
SAP HANA Enterprise Cloud: early thoughts
Vishal Sikka, executive board member SAP’s infectious enthusiasm for SAP HANA – his little girl – is palpable. Even when he’s suffering from sleep deprivation as he was the other night in a pre-briefing session, conversations are peppered with ‘incredible’ ‘unbelievable’ and ‘amazing.’ I sometimes wonder if even he is astounded by what HANA can do. That kind of enthusiasm and passion is essential for any business. It would not surprise me if Sikka turns out to be SAP’s most important sales asset going forward. This from his post on the topic:
I have the privilege to work with an unbelievable team that has worked *tirelessly* to get us to this point. They are the ones who’ve made this happen, the harbingers of change, the bringers of sunshine, on a cloudy day. J And as my little girl gets her wings and soars to the clouds and beyond, I want to take a moment to thank them, you know who you are, and to the incomparable HANA team, for making all this happen! A heartfelt thank you.
So what’s all the excitement about?
In short, the announcement is the start of a journey that that will see SAP eventually offering a series of deployment options for its whole applications portfolio. Today, the announcement concentrates on SAP’s BusinessSuite (ERP and CRM), custom HANA apps and BusinessWarehouse on NetWeaver in a private cloud environment. This is good news for customers who have been wooed by Oracle’s promise of Fusion applications deployed any way they choose.
This addresses SAP’s need to get its applications operating in real time rather than the batch processes upon which its applications rely. In remarks, Hasso Plattner, co-founder SAP said that SAP HANA Enterprise Cloud allows SAP to get from underneath the tyranny of batch and address the user interface (UI) topic. This is a telling admission and very good news for users.
Anyone who has worked with SAP applications already knows that UIs have long been a sore point. In some deals, modern UIs win deals, even when functionality is inferior to SAP’s offering. How SAP addresses this will be telling. My hope is that SAP drafts in designers who ‘get’ what modern apps need to look like.
This new offering should mean customers enjoy a lower cost of ownership and much faster time to benefit. Having said that, details are a bit thin on the ground and there are a few disappointments. Despite I attended a private briefing, tapped into the virtual event and also attended another virtual meeting on the day of the announcement, I was left with important unanswered questions.
Four big questions
These were the questions I put to SAP at the time of the event:
- If SAP doesn’t own the data centers then what assurances around the topics of security and governance will SAP provide to customers? If SAP does own data centers then what is the forward investment program?
SAP talked about having seven data centers available to it around the world. It is still not clear who owns what but we have to assume that at an absolute minimum, SAP will be using facilities it acquired when it bought Ariba and SuccessFactors. Even so, customers will want assurances on this point because SAP is effectively asking customers to further lock themselves into the SAP brand.
- If Amazon remains a complementary partner then how will pricing work out? Redshift is going up market and Infor already has plans in place.
SAP offers Amazon Web Services as a platform upon which customers and developers can spin up test and development instances. Cusomters and developers I have spoken with like this model. My guess is that SAP is thinking its largest customers will want something SAP assures as part of its contract and so opt for a private cloud option. That has a cost. The question then is to what extent will SAP share its economies of scale with customers?
- While the vision is clearly expansive, how will SAP solve the integration problems that will allow for frictionless business in and around an SAP world where as much as 60% of apps will be non SAP?
SAP is talking about a truly expansive vision where any apps you care to name will run on the HANA Cloud and database. Sikka went as far as to say he would welcome competitors running their apps on HANA. Magnanimous and audacious for sure. I don’t see too many takers any time soon when those same competitors already have long term commitments to existing databases and in-memory technology. Having said that, it will be fun to get reactions from CEOs at NetSuite, Salesforce.com and Workday. The more important question – what will SAP do about integration? Past efforts for some of its own applications have hardly proven successful. If the vision is to be realized then SAP must address this in a way that doesn’t impact the performance improvements that SAP is promising.
- SAP mentioned a 180 hours transition time. How will SAP persuade partners that this is a fast track solution instead of feeding their addiction to billable hours?
Here, I was thinking about the long standing systems integrator problem that SAP implementation push onto customers. In on-premise deployments, customers are effectively writing blank checks to their implementer. SI’s argue that projects start out one way but often end up as something very different. It is no surprise then that, depending on who you believe, time and budget over-runs occur in 50-70 percent of implementations. And guess what? Customers are always at the bad end of it.
Fast track transitions of the kind SAP is offering will make customers sit up but how will SIs react? Looking at the announcement on its face, SAP will take a lucrative part of data center operator/SI revenue. If successful, then IBM and Accenture could easily end up the biggest losers. Bet that these co-opetition partners will be whispering in customers’ ears. That’s a pinch point that injects disruption into the SAP partner ecosystem that also triggers negotiation.
The multi-tenancy topic
Right now, SAP is not saying too much about how they will manage the cloud environment or whether they will offer some form of multi-tenancy. They are clear however that each instance will be separate. For those not familiar, most SaaS vendors consider multi-tenancy as table stakes. This is because this kind of deployment achieves two important benefits:
- Vendors are able to keep operational costs very low.
- All customers are upgraded at the same time.
I have heard Plattner make impassioned statements about myths surrounding multi-tenancy but it has never been entirely clear to me what SAP’s alternative position looks like and how that stacks up with competitive offerings.
All this might be moot. SAP’s real problem is that all its large customers customize their applications rather than configure. In other words they make very deep changes that impact the running code. Attempting to put those customers into a multi-tenant environment is nigh on impossible. It therefore remains an open question as to what happens next.
Expectations are high and I can understand if Sikka feels as though the gloss is being tarnished by relentless questions. But when SAP makes these kinds of announcement, there are always many more questions. For example,
- When will SAP get its HCM apps into the cloud such that they can co-exist with SuccessFactors?
- What are SAP’s expectations around numbers of customer moving to SAP HANA Enterprise Cloud?
- Beyond the high level announcement, how will SAP articulate benefit to customers – will it only be the hosting element or will it include apps patching (for instance)?
- More broadly, what will the economics look like? SAP has said that the more customers use HANA, the more cost effective it becomes. OK – so what does that mean in a managed service center?
- Does SAP expect customers to sign up for long term subscriptions or will it offer flexible plans?
SAPPHIRE Now starts next week and it is a sure fire thing that this topic will dominate parts of the conversation among colleagues. Hopefully, SAP will address these and many more questions.
There is a comprehensive FAQ available here (PDF view/download)
Just a couple of clarifications, Den 1. SAP data center strategy - it will be a mix of SAP owned and co-located ones. The intelligence of managing the cloud is in the cloudframe software that Wes demonstrated . 2. Amazon and pricing - there are no changes to the Hana One strategy - that stays a valid option. In the SAP Hana Enterprise Cloud, SAP will host the systems in a managed service model which is different from AWS model. 3. Integration - this will be a key strength of the SAP Hana Cloud Platform, the PaaS aspect of the cloud. It is an open, Standards based platform and supports Native, integrated and Open development experiences. The idea of the platform is to use it for new apps and extensions for existing apps. 4. Disrupting the partner ecosystem - SAP is offering an additional deployment option to customers which we believe has high value. Partners are free to do the same with their own value additions. End of the day, customers make the choice - not vendors. And as Vishal clearly said - we are very supportive of customers being the decision makers. Ecosystem will most probably face some disruption - but the scale will be known only in due course. 5. Multi-tenancy is a complex topic - and what people often forget is that application is not the only level at which multiple tenants can exist. For hosting ERP and BW type systems, hardware level separation is usually needed - and is smartly done using the cloudframe software.
Very good article and congratulations to you and your partners on the launch as outside of what I know will be top notch content the website is very clean and user friendly. A few things jumped out at me regarding the announcement: 1. Why announce it one week prior to Sapphire? Should we be expecting even bigger news there or did SAP want to try to control the EnSw news feeds for two weeks. 2. Interesting that Vishal mentioned Workday and SFDC could be future customers when they havent even managed to get any functions of SuccessFactors running on HANA 15 months after the acquisition. They announced LAST Sapphire that SF Workforce Planning would be on HANA by Q3/2012 and now that is looking like July 2013 and no formal roadmap for the other areas of SF. Putting offerings on HANA is one thing but I think SAP is finding that getting substantial improvements is not always that easy. 3. I like the 10K and 100K club but always amazing to me how awful a process must have been to be able to be improved 10K :-) 4. The SI concern is a major one as one of my clients currently has 300 consultants of one of the firms you mentioned above and lets just say 180 billable hours got eaten up in the time I wrote this comment :-) 5. Will SAP be able to properly communicate this offering to customers as that has not been one of their strengths with the cloud offerings and with everything seeming to now carry the "cloud" and "HANA" names it can be very confusing. All in all I love the passion of Vishal and team but as with many areas of SAP execution after the launch will ultimately play a key role in the overall success.
The reason to announce this early is mostly because it will give customers and partners enough time to understand the options and come with questions and clarifications to SAPPHIRE. I share your thoughts on how awful some of the processes were before being rewritten, but now have a better grip on why this is important. Even if it was perfectly written in old technology - the improvement would have been only like 2X or 3X . That is not enough to make a dramatic process improvement. If a nightly job finishes in 2 hours instead of 6 hours - it would probably remain as a nightly job. However, if it becomes 10 seconds - then it might not be a batch process at all, and the overall business model can be disrupted where it makes sense. A good example is MRP runs in manufacturing companies
Hi Dennis, my first #Disqus post has disappeared somewhere, but i still like the platform. in any event, you pose very good questions around the enterprise software ecosystem, even without mentioning Microsoft and letting Oracle be present by proxies. so much expectation is built into HANA's architecture (shouldn't Intel have something to say here?) that many prior ERP practices may no longer be relevant like deep customizations, SI's billable hour maximization business model, or threats from competitors like NetSuite or Workday. it doesn't mean that history won't repeat itself, but it will not repeat in the exact same familiar way. overall, it seems like you are carving out a niche for your new venture from Gartner, SAP, and The Economist. i wish you all the best and many commenters and followers (and not just passive readers) of diginomica.com whose launch was more successful that this other on-line event to which many of us could not connect yesterday. rgds, greg
Hi Greg and thanks for dropping by. That's a very interesting observation and sparked a thought. I have seen new consultancies like Appirio arise in the Salesforce.com space where they fill in the need for customizations that get coded on the Force.com platform. In their model, customers benefit from fast track POCs based upon competitions created for their community. Now take a small step....SAP is desperate to get developers onside with HANA. 400 startups is a good start but nothing like enough. What if they took a leaf out of Appirio's book? Just sayin' Yes - we are trying to do something different. The model and reasoning behind it are explained in the About page.
The idea of SAP Hana Cloud Platform is to allow new applications and extensions to be built on it - with native, integrated and open developer experience. Crowd sourced development etc are all good ways for customers to get faster ROI - but that needs serious disruption of ecosystem. I do think that is bound to happen one day soon - it already happens in small pockets, just needs to scale.