It's amazing to me how much information is coming out of SAP before Sapphire on the cloud side - I cannot remember any Sapphire conference of my 20+ or so editions, that this much information about very key products for the company have leaked out before Sapphire so widely and publicly.
I hope its an indication for a more open, social and sharing SAP - which would be good for SAP and the whole ecosystem. On the flip side I feel for the marketing professionals at SAP, because this is not (at least the classic and proven path) how to build up and execute momentum before your key user conference.
![]()
So in the last 24 hours a blog post by @aiazkazi (you can find it here) and @bgoerke (you can find it here) did a lot of cloud busting before Sapphire. And since Aiaz and Bjoern work right under the executive board, and are kind of partners in crime for cloud matters, it is reliable enough to chart the way forward into the cloud, as we can expect it to materialize next week at Sapphire.
My take was that HEC clearly was IaaS, which got confirmed with tweets with a few SAP mentors in the aftermath of my post. It was unclear earlier in the week what would make HEC to become a platform, a PaaS player and the foundation to build the game changing killer apps, that I am waiting for SAP to come up with - hopefully sooner than later. So Hana Cloud Platform (HCP) provides the platform support needed to create applications on HEC, so what are the features and products in HCP? Earlier this week the focus was only around the database with Hana - which was fine - but that didn't shed much light on how to build applications on HEC.
Neo was the codename for the initiative of bringing NetWeaver to the cloud - re-inventing (or better re-writing) the infrastructure pieces, that in the past had weighted SAP products down, a key component being the Landscape Virtualization, which is SAP's attempt to make their products more elastic. Neo's product incarnation was NetWeaver cloud platform (NCP). It looks to me that NCP components find themselves again in HCP and the latest NetWeaver releases - NetWeaver 7.4. Here it's actually still quite cloudy and it would be good to know from SAP, which components go where (but anyone can make an educated guess).
For customers this should be good news, SAP embracing more cloud will ultimately lead to a more modern application architecture and infrastructure. And as the cloud is all about lowering TCO, this trend may als happen to the SAP cloud offerings. But in the neartime, make very sure that you now what it will cost you and that you can benefit from hopefully sooner than later coming cost reductions - both on the licensing as well as the operating cost side. You can influence SAP by keeping the pressure up to keep supporting the Business Suite AMI packages on Amazon - and push and shove for SAP to commit to support production instances on Amazon as well. This will keep SAP honest with cost and keep a focus on value.
For partners it means much more to play, more to learn and more to advise SAP customer on. It's also not clear what customization strategy you should pursue, I hope SAP will clarify this next week much more. It's also an opportunity to create value added applications, filling the void of the killer app that SAP has left so far. SAP is doing well at partnering and fostering an ecosystem for Hana partners, true to the spirit of let many flowers grow. But if you build applicatons on HEC - you need to be aware that for good or bad - you are tied to SAP's ecosystem.
For SAP this means the closure of many separate development streams - just think of the consolidation that this means for the NetWeaver platform. But it also means that a new programming model (pardon the SAP lingo) - that will need to be trained, digested and executed. Building edge applications is only a short to medium term strategy, so at some point SAP will have to start building core automation, hopefully with 21st business practices in mind, on the HCP for HEC. That would be (pardon the Oracle lingo) a Fusion type event. But ABAP will not run cost effectively forever in a multi core world.
I hope its an indication for a more open, social and sharing SAP - which would be good for SAP and the whole ecosystem. On the flip side I feel for the marketing professionals at SAP, because this is not (at least the classic and proven path) how to build up and execute momentum before your key user conference.
So in the last 24 hours a blog post by @aiazkazi (you can find it here) and @bgoerke (you can find it here) did a lot of cloud busting before Sapphire. And since Aiaz and Bjoern work right under the executive board, and are kind of partners in crime for cloud matters, it is reliable enough to chart the way forward into the cloud, as we can expect it to materialize next week at Sapphire.
Hana Enterprise Cloud
This week we saw the - again with surprising timing - the launch of the Hana Enterprise Cloud (HEC). I called it a promising start, since it was the first time I saw senior SAP executives, in this case Hasso Plattner and Vishal Sikka, use cloud unique benefits to pitch homegrown SAP product. This is different than e.g. Lars Dalgaard preaching about the cloud, his whole background is cloud, he sells his past Successfactors products, so that's an easy pitch. But to see Hasso Plattner associating easy ramp-up, making hardware provisioning obsolete for customers, staying on the latest code with a SAP product while claiming to satisfy the much postulated sub 1 second response time - is a first and shows a good turn in thinking and thus in SAP products.From IaaS to PaaS - thanks to Hana Cloud Platform
My take was that HEC clearly was IaaS, which got confirmed with tweets with a few SAP mentors in the aftermath of my post. It was unclear earlier in the week what would make HEC to become a platform, a PaaS player and the foundation to build the game changing killer apps, that I am waiting for SAP to come up with - hopefully sooner than later. So Hana Cloud Platform (HCP) provides the platform support needed to create applications on HEC, so what are the features and products in HCP? Earlier this week the focus was only around the database with Hana - which was fine - but that didn't shed much light on how to build applications on HEC.
 |
| The HCP as described here |
Behind Hana Cloud Platform stands Neo & more
Neo was the codename for the initiative of bringing NetWeaver to the cloud - re-inventing (or better re-writing) the infrastructure pieces, that in the past had weighted SAP products down, a key component being the Landscape Virtualization, which is SAP's attempt to make their products more elastic. Neo's product incarnation was NetWeaver cloud platform (NCP). It looks to me that NCP components find themselves again in HCP and the latest NetWeaver releases - NetWeaver 7.4. Here it's actually still quite cloudy and it would be good to know from SAP, which components go where (but anyone can make an educated guess).
So what has been taken / moved from NCP into HCP:
- All the good Java / JVM work that was done for NCP is now part of HCP, including the Eclise based IDE
- Platform enablement services such as persistence (SAP is re-using the code for the Business Suite here, goo), identity management, document management, mail services, HTML 5 support
- The operations console - a key component for any cloud platform
I am also assuming the ABAP Server was somehow made part of HCP - no SAP product without ABAP support - but that's an assumption right now.
How to build HCP applications
Thanks to the re-use of NCP for HCP, developers now can create applications in an Eclipse environment, using Java - and JVM byte code compatible applications with e.g. Ruby, Python, Scala etc. And there is of course the existing Hana Studio with support of JavaScript / SQLscript. Good news is, that the River RAD platform with its RDL language seems to be coming to Hana too - but no date is set.
Developers than have access to a plethora of services
- Application Services
Mobile, Portal, Security and integration - courtesy of NCP
Analytics - courtesy of Hana and BW
Collaboration - courtesy of Jam (not sure how though that will integrate)
- Enablement Services
Applications & Systems Management - courtesy of NCP)
Administration & Monitoriing - courtesy of NCP and Hana
- Database Services
Transactions, Analytics (again?), Streaming, Predictive, Spatial, Text Mining - all courtesy of Hana (and all the way back to fabled TRex for text mining
That is a powerful and impressive collection of services to build applications, no doubt. And some of it - is free for trial, application building.
NetWeaver 7.4 - confusion stopped
At the same time SAP announced the GA of NetWeaver 7.4 - so NetWeaver is still around and doing well - as the foundation of the Business Suite. And it shares code lines with the NCP / HCP - e.g. in regards of database services, as SAP is using the same code to bring the Business Suite to use Hana as for the new built applications on top of HCP.
 |
| The original NetWeaver RoadMap - from here |
Good news for NetWeaver 7.4 is also, that it allows the older Business Suite applications will be able to call newly built HCP applications, of course something highly desirable and enabled by splitting the former NCP services across the NetWeaver line and the future HCP platform line.
Likewise the claim is there, that HCP built applications can call back to the Business Suite, a claim I am close to believe since again, same foundation and a huge necessity for SAP to build the next generation applications on HCP.
Still looking for the killer apps
I have criticized SAP many times for not showing the though leadership for businesses running their enterprise software in memory. All the examples and apps are just throwing a faster database to an old, well known, but not really solved performance problem of the fast, created by business process locked in the past. We all know - or at least feel, that 21st century best business practices, running in memory - need to be and will be different than what was created on an incredible more limited application infrastructure last century. Just remember that one key driver for relational databases was to... save disk space!
And while @vijaysankarv lays down the case, why there is no real killer app in his blog, my reply to him on Twitter was, that SAP needs to first create the next generation platform, before it can build the next generation killer app. And the recent clarifications bring SAP much closer to the next generation application platform.
What's missing
Well, first and foremost pricing, and I hope that SAP will shed some light at Sapphire next week. The BYOL (Bring your own license) is definitively not a cloud business practice.
The other key question is elasticity of the offering, the key benefit of the cloud. Requiring customers to buy capacity on a monthly based for HEC is not a good sign of easy administration and dynamic memory allocation etc. Which leads to the question of the capability of the landscape services.
Equally there is no clarity (yet) on how productive a developer will be in building brand new applications, and hybrid applications that enhance the existing business suite applications.
Equally there is no clarity (yet) on how productive a developer will be in building brand new applications, and hybrid applications that enhance the existing business suite applications.
The part that concerns me the most at this point is elasticity. And while SAP has done a good job with the NCP Landscape Management - it does not make elasticity much easier to handle - as you have a bunch of older, non elastic apps to administer. We all remember the TCO debacle around by-design - caused by too many layers in the application architecture. And while SAP got rid of the layers in the next generation application architecture, the overall system landscape is more complex than the one by design ran on. I would really like for SAP to abandon any internal, proprietary standard for virtualization / elasticity and follow the many other vendors that have endorsed Openstack standards.
It's ironic - the by design product stumbled over too many layers on a single system, HEC may stumble despite fewer application layers over too many systems in the landscape... I hope this is a concern about nothing...
It's ironic - the by design product stumbled over too many layers on a single system, HEC may stumble despite fewer application layers over too many systems in the landscape... I hope this is a concern about nothing...
What does this all mean?
Let's play this through for the constituents of the SAP ecosystem...For customers this should be good news, SAP embracing more cloud will ultimately lead to a more modern application architecture and infrastructure. And as the cloud is all about lowering TCO, this trend may als happen to the SAP cloud offerings. But in the neartime, make very sure that you now what it will cost you and that you can benefit from hopefully sooner than later coming cost reductions - both on the licensing as well as the operating cost side. You can influence SAP by keeping the pressure up to keep supporting the Business Suite AMI packages on Amazon - and push and shove for SAP to commit to support production instances on Amazon as well. This will keep SAP honest with cost and keep a focus on value.
For partners it means much more to play, more to learn and more to advise SAP customer on. It's also not clear what customization strategy you should pursue, I hope SAP will clarify this next week much more. It's also an opportunity to create value added applications, filling the void of the killer app that SAP has left so far. SAP is doing well at partnering and fostering an ecosystem for Hana partners, true to the spirit of let many flowers grow. But if you build applicatons on HEC - you need to be aware that for good or bad - you are tied to SAP's ecosystem.
For SAP this means the closure of many separate development streams - just think of the consolidation that this means for the NetWeaver platform. But it also means that a new programming model (pardon the SAP lingo) - that will need to be trained, digested and executed. Building edge applications is only a short to medium term strategy, so at some point SAP will have to start building core automation, hopefully with 21st business practices in mind, on the HCP for HEC. That would be (pardon the Oracle lingo) a Fusion type event. But ABAP will not run cost effectively forever in a multi core world.
MyPOV
6 days before Sapphire SAP has unleashed a lot of information around their future cloud offerings. Still don't know why that early and not at Sapphire. But it was very good to see senior SAP executives use cloud benefits to justify the new HEC offering. On the flip side - a lot of key cloud DNA is still missing, the good news is - if SAP wants to succeed (and I am sure they want) then the rest will have to fall in place. SAP may like it (faster modern applications) or not (lower license sales, teaching the elephant to be elastic).
With every announcement there comes confusion, but thanks to blogs and twitter and open information sharing by Aiaz, Bjoern and a relentless Vijay (@vijayasankarv) - much more clarity is there by now... and it's still 48 hours (or so) before Sapphire 2013 starts.