Skip to Content

This year, as an SAP Mentor, I was lucky to attend both the US and European TechEd and found a consistent message about the SAP HANA Cloud Platform (HCP). I’ve been following it for a long time as I know some of the developers of the platform, but I had not seen such a prominent set of messaging about it. This blog will outline why I, as an SAP Mentor, think HCP is so important for customers and how it may evolve.

So let’s get one thing straight – HCP is for custom developed applications it is not for transactional systems like ABAP stacks (eg ECC or Solution Manager) – there is no ABAP engine in HCP and there are currently no plans to offer it. HCP is like Heroku, Engine Yard, Google Compute and Azure, technically it should be possible to take a Java application and deploy it on any of the above platforms and it will just run – try doing that with a transactional application :-). Although this is an over-simplification of the PaaS vision, there are of course challenges to doing this and these have been identified in Matthias Steiner’s blog “Spring Cloud Connectors”.


Below is an example architecture for a simple custom application running in HCP and connected to a back end system.


HCP as I see it is a great opportunity for customers in the following ways:

1.   Application testing – in HCP you are developing using technologies like Java, Javascript, UI5, HTML, CSS etc…Basically, Web technologies which have a strong heritage in lean and agile development, which creates a much more open playing field than traditional ABAP.

A good example is the ability to use quality/testing framework components like Cucumber, Git, Jenkins, Gerritt and Selenium which can be combined into a solid testing platform. Imagine this, a developer pushes a commit to Github, this triggers Cucumber to run a suite of tests in Selenium against the applications WebUI. The diagram below shows an example of a development of code in an environment with automated testing capability,


As a result the developer is able to run a battery of tests against their application – perhaps executing as much as 80% of the testing before a human tester had even seen the application. What is even better is that those tests are now stored and able to be executed again and again for that code to ensure it still functions. There are pre-built images in Azure and Amazon Web Services with these testing architectures, all the user has to do is point the testing components at the application to be tested and populate with the tests.


2. Regression testing – one of the biggest pain points for customers which reduces the number of deployments they do in year is regression testing. The impact on the business for SAP heavy customers is massive – I know a retail company who spends 6 weeks every quarter doing business and regression testing for a release.

If you develop your custom application in HCP with most of the business logic inside it (instead of in the back end) and use OData services exposed through NW Gateway to push data back into the on-premise transactional engine, then you will be minimising the customization in the back end.


As a result, the burden of testing reduces as well, simply because you are changing very little in the back end – as we saw above – the testing HCP applications is much more flexible than traditional ABAP testing using tools which also do not carry as high a license cost against them. This is not a hard and fast statement but I believe that it is possible to reduce regression testing effort by pushing more of the customisations into the HCP layer and into automated quality testing frameworks.

3. Unused/obsolete applications – in HCP when you no longer need applications you just turn them off. If you had a custom application in ABAP, that code would have to be backed out using transports (how often does that happen?).

Application lifespans are much shorter now than they used to be, as a result applications get replaced more often as business processes change.



The great thing about this is that if you are building customisations in HCP and not in your transactional engine then patching and upgrades are much easier as it reduces the amount of new customisations to be dealt with during software updates.


4.     Cadence of change – above I discussed a retail customer who had quarterly release cycles due to testing effort by the business. If the previous points hold true and overall testing effort is reduced in the ABAP stack due to moving customisations out of the back end and the ability to use automated testing; then it follows that customers can deploy new applications, changes and updates faster than previously possible. It should be possible to have an HCP application which can be updated every two weeks whilst just using OData services to connect to the back-end, whilst the main back end system still uses quarterly release cycles. As most of the customisations and/or business logic is in the HCP application then and required changes can be made in the ABAP stack as small change with limited testing. This would be a massive cultural shift for a lot of organisations who are used to slower change life cycles – but it will enable them to drive faster projects to unlock better business value, provided they are targeting the right opportunities and using the right development methodologies. We have seen these opportunities with companies like Etsy or Netflix being able to deploy code on both IaaS and PaaS platforms with ease and high frequency. This will also drive changes in Ops on how they deal with change, for example,approval processes will need to accelerate. Developers will have to take greater responsibility for their handovers and the life cycle of their applications in helping the wider team resolve issues.


5.     Consistent UI – I love the promise of UI5 and the mobile application services in HCP, it really provides the opportunity of “build once, run anywhere”. This has the ability to vastly reduce the workload on development teams to create a great application quickly, effectively and across multiple platforms to a large section of the user population. It really becomes a force multiplier for application deployment of quick and lightweight apps.

These 5 points I feel combine to make HCP a compelling platform for customers to start taking advantage of in order to accelerate their development plans.

As SAP offers more SaaS solutions to customers (Concur, Successfactors, Ariba) which have limited customisations, customers will have to use a combination of their SaaS application and HCP to create customised/extended applications. This will hopefully encourage developers to develop in HCP, thus centralising customisations for the SaaS products and also enabling customers to develop applications connected to their back-end systems rather than directly creating custom applications in their backend system. This also benefits from an operations point of view in terms of change management, security, monitoring points of view – but that is another blog.

To report this post you need to login first.

5 Comments

You must be Logged on to comment or reply to a post.

  1. Matthias Steiner

    Hi Chris,

    happy to see you chime in on the topic. Given your Ops background it’s great to hear your perspective on things! Looking forward to more 🙂

    Cheers,

    matthias

    (0) 
  2. James Wood

    Hi Chris,

    I echo Matthias’ sentiments – it’s very interesting to get your perspective on this. I did have one question for you though. In reading through the various usage scenarios you described, it occurred to me that (and I know you know this) it would be possible to implement these scenarios on-premise, too. Which brings me to my question: in your opinion, what is it about the HCP that will make customers finally bite on setting up scenarios like this?

    I ask this question because I’ve seen customers dismiss these kinds of setups for years and I’m not sure why since they’re fairly commonplace outside of the SAP world. Now, I am starting to see a gradual shift in the direction of distributed architectures with Fiori applications (i.e. they’re at least standing up Gateway as a frontend server, etc.), but so far the scenarios have been pretty vanilla and based almost exclusively off of pre-delivered content from SAP.

    Anyway, curious to get your insight. Great work as always.

    Thanks,

    James

    (0) 
    1. Chris Kernaghan Post author

      James,

      Glad you like it, I suppose there is no barrier to having these architectures on-premise, but as you say we, in the SAP world, are not used to operating them and lack a compelling reason to change.

      In my head the compelling reason to change and become the centre part of the custom application landscape will be because SAP will offer their SaaS products more heavily. For example Successfactors in place of traditional HR and Talent Management, Ariba in place of SRM and Concur for Travel and Expenses. As these SaaS products have limited customisability, it pushes customers to HCP for extension and customisation. As customers become more aware of the ability to innovate, develop and deploy faster in HCP then it will hopefully encourage them to go onto HCP for more than SaaS extensions.

      Although in order for customers to do these things, SAP and partners need to invest in skills and use cases. This has the potential to change fundamentally how change and customisation is delivered in SAP landscapes. It opens the door to the use of non-SAP technologies which can break them&us barriers which exist in Enterprise IT structures and create common languages and understanding.

      Possibly – dare I say it – World peace 🙂

      In all seriousness though, if it allows SAP landscapes to provide better and faster business value then it can better align itself, IT and the business to create that elusive Digital transformation we all hate hearing about.

      Chris

      (0) 
  3. Fenton Morris-Winmill

    All very good, adding new technologies for development (although I prefere to keep my developments in the same place as my application, rather than extending it outwards).

    However what is missing to me from an SAP strategy point of view, is the traditional SAP message.

    Applications, Applications, Applications.

    Customers don’t want to develop any more. They want new ways of doing business, they don’t have the all the answers.

    SAP have a platform in HANA that could revolutionise the way we do business in true real time (time for R/4 ;0) ), but so far SAP has only come up with simple finance.Nothing really game changing.

    The CEO/CFO/CIO needs to see a true compelling business reason to adopt the new Platform. Services such as Successfarctors,Ariba, Travel, are parts of the business that in the past have been overheads, so it makes sense to consume a cloud service and to leave well alone (why bother with further extentions, if another provider offers further functionality, switch to them), they hold no real strategic value to the core business.

    Now give me an Application with true Business value that helps me to become more competitive, money will flow.

    (0) 
    1. Chris Kernaghan Post author

      Fenton,

      Like you and customers I do not have all the answers 🙂

      I cannot comment on the traditional SAP strategy, although I would say that Applications, Applications, Applications has not gone away – as you look at the SaaS products, these reflect that Applications strategy excellently.

      SAP’s provided customers the ability to customise their experience – that is what people like Hasso keep talking about providing to our customers. The HCP PaaS empowers customers to centralise their extensions and customisations in a single location, which is scaleable and provides a great UI/X capability.

      As regards customers not wanting to do development anymore, I disagree and would counter that with SAP customers don’t want to do SAP customisation which clogs up their systems, screwing up patching and upgrades which raises the CTO of the platform. They are more than happy to develop new lightweight and exciting applications which empower their staff and make their lives easier and which can be delivered in more Agile means. I have seen more customers wanting to leverage the building blocks of technology to create exciting applications and not depending on the vendor to give them a standard solution.

      As regards the business value, there is a massive amount of business case related documentation for HANA on the SAPHANA.com website. We live in a world of more informed customers, as far as I have seen, SAP and partners have done massive amounts of work in providing this documentation for customers to allow them to build business cases. As S4HANA becomes clearer in it’s capability, around add-ons, cleaner data model, external application interface/connectivity – you will find a lot more messaging about the applications and business value which takes into account a hybrid world of where SAP is not the only solution a customer has access to.

      Thanks

      Chris

      (0) 

Leave a Reply