In my Unwrapping SuccessFactors’ Cloud Architecture – An unconventional treasure hunt, I examined SuccessFactors’ technical architecture. This time, I’d like to continue my explorations and look at how this architecture corresponds with SAP’s current and future cloud strategy.
Currently, SAP’s cloud architecture is based on a division between Core and Edge (These designations are probably going to change soon).
Let’s reduce this architecture to the building blocks that are most relevant for our discussion and then start rearranging the blocks to see what possibilities emerge. This simplification allows us to make general statements about how the two platforms will work together in the future.
Note: In this blog, I’m not going to examine the SAP OnDemand offerings in detail. I’m more interested in high level architectures. If you’d like to get details of these offerings, then I suggest that you look at my previous blogs on the topic.
Breaking it down to a bare bones diagram, here is what we have to start out with:
Notes for the diagram:
- SAP’s OnDemand offerings are largely based on two separate PaaSs.
- Although many SuccessFactors applications share a common heritage in Java / J2EE (for example Plateau and SuccessFactors), there is really no PaaS in use. It looks like JBoss is the core of the environment but this is an Application Server not a PaaS. Before someone comes back and says that Jam is Ruby on Rails-based, I assume that Jam is also running on the JBOSS environment.
- SAP’s Edge Platform / PaaS (Neo) is currently starting its Public Beta, so the lines to StreamWork and Carbon Impact are dotted – meaning that they are planned for the future.
- ByDesign and the other ‘OnDemand’ applications that are associated with the SAP Core PaaS are often called Line Of Business (LoB) OnDemand applications.
Adding Cloud Foundry
Recently, SuccessFactors announced an agreement with Cloud Foundry to allow partners and customers to create custom applications:
The need to customize, extend and integrate enterprise applications is a common one. Doing this for SaaS applications brings particular challenges, as those applications run remotely and historically have been limited in terms of programmability. By bringing together Cloud Foundry’s robust runtime capabilities with SuccessFactors’ business processes and data, developers get the best of both worlds.
We have to change our diagram accordingly:
With this foundation in place, let’s examine what technical scenarios are possible.
Reshuffle the cards: Cloud Foundry and SAP’s Edge Platform
I assume that in the mid-term SuccessFactors will work closely with SAP’s Edge Platform team to allow SuccessFactors customers and partners to use the functionality of the platform to create lightweight applications based on SuccessFactors’ APIs. Whether SuccessFactors will support both SAP’s Edge PaaS and the Cloud Foundry to meet this business requirement in the future is unknown.
Since the agreement with Cloud Foundry is recent (just 2 months ago), I’m sure that many SuccessFactors partners and SIs who were planning to use Cloud Foundry are now revaluating their choices. For those who are planning to develop or have already developed such SuccessFactors applications, I’d enquire in the ability to port from Cloud Foundry to SAP’s Edge Platform.
Reshuffle the cards: SAP’s Edge Platform and SuccessFactors
Based on the fact that SuccessFactors is Java-based, there seems to be a perfect match between SAP’s cloud strategy and that of SuccessFactors.
We have two applications platforms. For the Core/Suite Application, BusinessByDesign-based applications. For the Edge / the Java –based Platform as a Service, and SuccessFactors fits perfectly into that. So over time, we expect to stay perfectly loyal to this roadmap – that we have articulated over the year and bring the power of SuccessFactors and of Mobility and of HANA to every application from our customers whether it is OnDemand or OnPremise.
-SAP CTO Vishal Sikka from the SAP / SuccessFactors call on December 5, 2011 [SOURCE]
In a Twitter exchange with Zia Yusuf (Former EVP Global Ecosystem & Partner Group SAP AG), Vishal associates SuccessFactors with the Edge category.
This statement is, however, a very general one. What exactly are the architectural options that are available?
One idea would be that the SuccessFactors applications run on the Edge Platform.
Initially, I was very excited by this option. If set in motion, it would be a major boost to the importance of SAP’s new platform. But after researching the technical architecture of SuccessFactors, I feel that this option might be unrealistic.
I assume there are some internal investigations within SAP examing the ability of porting JBoss applications to the Edge PaaS. Inasmuch as the majority of SuccessFactors applications are J2EE-based, a port to the Edge PaaS should be possible. If the SuccessFactors applications were ported to the SAP’s Edge PaaS, however, then there would be a functional overlap between the JBoss Enterprise Application Platform functionality (Hibernate, Caching, Clustering, and High Availability, Integration and Messaging Services, etc) already used by SuccessFactors’ applications and that offered by SAP’s Edge Platform. I can assume that SuccessFactors will be reluctant to start replacing components in its successful unified platform just to support SAP’s PaaS.
Before its acquisition by SAP, SuccessFactors had the agreement with Cloud Foundry. There are a variety of similarities between SAP’s Edge platform and Cloud Foundry. If SuccessFactors really was interested in moving away from JBoss, then they would have started examining a port to Cloud Foundry. I have yet to find any indication that this option is even being considered. As the announcement with Cloud Foundry emphasizes, this cooperation focuses on new lightweight extensions rather the core applications.
As I mentioned in my last blog, SAP plans to acquire SuccessFactors for its success in the Cloud.
In fact, in the same conference call, [Bill] McDermott underscored SuccessFactors’ unmatched ability to scale across big numbers of users and indicated that the company has 5 times more subscribed seats than any other cloud-based business-software competitor:
“Together, with SuccessFactors’ widely respected team and its deep experience in building and bringing to market scalable cloud solutions, we see tremendous upside to create new business opportunities,” McDermott said. [SOURCE]
This success is largely partially on the existing architecture, the processes and associated experience that SuccessFactors has built up over the last decade. This platform is based on a SaaS-model rather than on PaaS. Why would SuccessFactors trade these material and immaterial assets for a port to the new platform?
A more likely scenario is that SAP Edge PaaS-based applications and the JBoss-based SuccessFactors applications initially exist side by side and are both branded as ‘Edge’ applications – apps that can run Java-based environments.
For our analysis, there are two useful aspects when examining the difference between Edge and Core characteristics as described in SAP’s cloud strategy: technology and application focus. In terms of Edge referring to a Java-based environment, then SuccessFactors apps are Edge apps. In terms of application focus, the SuccessFactors applications are more Core applications. This LoB-associated focus is also apparent in the the SAP press conference with analysts on Dec. 3:
In addition much of the cloud market today is a line of business market. These solutions focus on targeted functionalities and very specific business-users. While companies appreciate the simplicity of focused cloud solutions, they increasingly require integration with their core on premise applications. They realize that in a world of hybrid solutions, they will get most value when they integrate on premise, cloud and mobile.
That is where SAP comes in with the acquisition of SuccessFactors. We will be the only company that can offer end-to-end processes integrating world-class line of business cloud offerings with the most respected on premise solutions available. [SOURCE]
Thus, you have discrepancies between SuccessFactors’ technical foundation, its application focus and how both fit into SAP’s existing cloud strategy.
Reshuffle the cards: A new dealer sits at the table
While reading various blogs about the SAP SuccessFactors deal, I read one paragraph that made me spill my coffee.
SAP is the company that has the most intellectual property and the most applications for the last 40 years since Hasso Plattner started this company. We can take all of that IP and put it on our platform and we can compete in every market—and even more exciting, compete in markets that never existed before! Create incredibly lightweight, beautiful applications—more beautiful than the iPad—for all users, everywhere in the world.
-SuccessFactors CEO Lars Dalgaard [SOURCE] [Emphasis is mine]
What exactly does Lars mean? He is suggesting that SAP process knowledge / experience (“that IP”) be moved to the SuccessFactor’s platform. I interpret this as a new line of LineOfBusiness applications that you might assume to be based on SAP’s Core PaaS but which will be based on SuccessFactor’s SaaS platform. There might also be a set of new Edge applications based on this SuccessFactors platform.
We should also recall that Lars is going will lead SAP’s cloud efforts as a member of the executive board [SOURCE], so his intentions aren’t to be ignored.
If this new type of application emerges, it will be curious to see if they will be marketed under the SAP or the SuccessFactors brand. It will also be interesting to see if these new applications threaten functional areas that are now currently covered by Business ByDesign. As ByDesign is a complete integrated suite, it will definitely be a challenge for Lars to select those use cases currently covered or planned in ByDesign that might be ideal for the SuccessFactors platform. A complete port of the entire ByDesign suite to the SuccessFactors platform is a technical possibility (remember all that great IP) but is probably not an option at this point due to the costs (in terms of time to market delays and development resources) involved in such an endeavor.
Porting the existing LoB OnDemand applications (SalesOnDemand, Sourcing OnDemand, etc) might be an option. The tight integration of SalesOnDemand with the existing SAP Core PaaS would make a port to the SuccessFactors platform, however, a challenge. I hoping that the features in these applications that have emerged over the last year based on SAP’s co-innovation with customers and design thinking won’t be lost in any future transition.
I see these changes occurring in the mid-term as the cloud strategies of the two companies co-mingle and the synergies between the two platforms are explored.
Long-term developments are more difficult to judge. Obviously, some consolidation is necessary between the offerings / products of the two companies. Currently, SAP has a broader approach focusing beyond SaaS to the creation of two PaaSs that enable the involvement of a wider circle of developers dealing with a much greater variety of applications than that currently faced by SuccessFactors. Although very successful with their focus on single SaaS offerings, SuccessFactors has had the ability to concentrate on niche products. It will be interesting to see how the involved technology, experience and processes deal with the new broader scope. Although Lars heads SAP’s Cloud efforts and can rightfully brag about his successes in the marketplace, he would be wise to consider the impact of his expanded mandate and listen to his new SAP colleagues and their motivations and experience which are the foundation of SAP’s PaaS platforms.