Skip to Content
Author's profile photo Tobias Trapp

A Quick View on different SPs of SAP NW 7.31 regarding Quality & Disruptiveness

For me as software architect the last time has been really difficult – I was uncomfortable about the quality of AS ABAP software components. I spend too much time analyzing errors and in for some of them there working but not satisfying solutions. One source of the problems are disruptions which I mentioned on SCN and I told you that fortunately SAP could help in many cases.

In the meantime I learned that some of my calculations are in favor of SAP because I encountered a kind of “morass” in software components that was deleted and didn’t occur in my statistics because it consists of elements that are not assigned to a software component and therefore don’t appear in my statistics.

Then I had the feeling that some SPs of NW 7.31 had a terrible quality – especially SP4 and SP5 because in two strategic frameworks disruptive errors have been shipped and in another strategic frameworks the number of OSS messages much too high. I had mixed feelings when looking at different SPs of NW 7.31 and I am asking what happened here. Of course I can only guess but I don’t trust my impression so I tried to find out what was going on.

Nearly no Deletions from SP 5 to SP 8 🙂

My first question was whether SAP continued to delete the platform and I was surprised. At first some good news: from SP5 to SP8 there have been no deletions in SAP_ABA – congratulations!  The number of TADIR objects grew from 47.871 objects to 48.347 objects which is under 1% new development. The same is true for SAP_BASIS: I found only 385 deletions, but if you look at the deletions then you will find many classes like CL_EPM_PD_UPLOAD_FORM_FEEDER which belong to NetWeaver demo application. The software component grew from 251.069 to 254.040 which is roughly 1%.

What happened from SP5 to SP8?

It seems to me that from SP5 to 8 SAP fixed bugs but didn’t neither ship new functionality nor deleted anything. It seems to me that in these SPs only bugs have been fixed which is the purpose of an SP.

This is quite in line with my personal experience. My first test of NW 7.31 software started with SP3 and I didn’t found any problems. In SP4 and SP5 I encountered some severe problems. Thanks to SAP they have been solved in OSS notes and in SP6, 7 and 8. From my experience upgrading to SP8 was a good decision.

My Personal Conclusion

I experienced the following:

  • some SPs (especially 4 and 5) had a very poor quality
  • at the moment I can’t predict the costs of a SP upgrade in terms of test effort
  • upgrading to some SPs has been more problematic than implementation of some Ehps

At the moment I am a bit helpless. My former strategy of estimating disruptivity of Ehps early (and at a low SP level is failing) because everything can change with a new SP. As consequence at the moment some SPs require so much test effort that I’m thinking of additional investments in automated software tests.

The reasons is simple: I expect that new frameworks can be unstable but bug in strategic frameworks that are foundations of business solutions are fatal.

My Advice to SAP: Improve the Quality of SPs and keep an Eye especially on old Code

I hope that SAP will continue the strategy to prevent deletions.

I think customers will embrace new SAP NetWeaver features in SPs as long as the platform remains stable. But achieving high software quality and stability should be highest priority. And from experience there is much room for improvement – especially the quality of SP4 to 7 was poor from my experience.

For me it is especially annoying that some changes in old code had drastic and negative effects on stability. In my opinion SAP should change old code only if the following preconditions are met:

  • the change is compatible and so no disruption will occur and
  • the change has been thoroughly tested.

Assigned Tags

      You must be Logged on to comment or reply to a post.
      Author's profile photo Alexander Sperling
      Alexander Sperling

      Thanks Tobias, very interesting article, especially if currently upgrading to NW 7.31. 😉

      Could you give some details on the frameworks you mentioned and which one should have a closer look into?

      Author's profile photo Tobias Trapp
      Tobias Trapp
      Blog Post Author

      There are many frameworks having bugs. Even if I would give you the name of the frameworks this won't help you because it is possible that in your scenario these problems won't happen. So I recommend to check the OSS whether the number of OSS notes increase in the applications you are using and there are many high priority notes especially for code that you considered as mature because it kept stable since years.

      Moreover lots of SAP's changes (and so the bugs) have been ported down to earlier SPs so in fact some of the problems waited for us in SP 13 of NW 7.02, for example. So even staying at 7.02 wasn't a good solution for us - and not implementing SPs would be a bad choice as well.

      If you read my previous blogs then you will recognize that many objects have been deleted from 7.0 to 7.31 - so please be warned that the implementation of Ehps and SPs may require more test effort. For us the upgrade to SP8 of NW 7.31 was the best solution.

      So my recommendation is:

      • try to implement the latest SP,
      • perform directly after SPAU/SPDD an SGEN of your codes as well as SLIN to do static syntax checks of your custom code,
      • try to test as much as possible and
      • try to plan that there is additional time if problems should occur.

      Best Regards,


      Author's profile photo Jelena Perfiljeva
      Jelena Perfiljeva

      Tobias, thank you for sharing, You've mentioned that:

      at the moment some SPs require so much test effort

      What do you use to determine/measure the possible test effort? We usually just end up doing the same end-to-end testing including as much as possible, but I wonder is there is a good alternative (other than some expensive software marketed for this purpose, allegedly 🙂 ).

      Thank you.

      Author's profile photo Tobias Trapp
      Tobias Trapp
      Blog Post Author

      The cost is easy measurable: just look at the number of people who do test and how long they work. In some cases we use automated tests since we have many eCATT tests and can run batch jobs and control the results. But this is topic for another blog.

      Best Regards,