on 10-04-2013 4:41 PM
We recently got to know that MBO is no more supported from SMP 3.0. What does that mean for the offline app strategy?
Will like to open this discussion to get insights from the experts here.
Thanks,
Shabz
Interesting question and good discussion. I think most people agree SAP's "strategy" around SUP and SMP has been interesting to say the least. It's making it almost impossible to go into customers with a convincing and solid business case around SUP or SMP, as there are too many unanswered questions.
In my eyes, the lucky ones are those who have already invested in 2.2 or 2.3 SUP and have got working, MBO based applications up and running and are in a position to not worry (too much) about what will come with 3.0. Those customers who are just beginning to think about SUP 2.3 or SMP 3.0 have a much more difficult decision to make.
Anyway, here are some facts/fictions/wishes I know (or at least think I know) in this space. How true and/or accurate they actually are is open to debate...
In summary, the missing link in so many mobile discussions is off-line support of the oData (i.e Gateway) protocol. I personally think you can then rationalise all of this technology down, so that you can model your business objects in and publish them as oData/REST type objects. You could then consume them in mobile apps, BPM processes, SAPUI5/Fiori style app's, etc. If you need them to be off-line, you let the SAP Highlander platform take care of it with a simple checkbox. Simples*.
The topic of off-line oData came up a lot at the recent SITMAN event where many agreed this would get us closer to developer nirvana, as put it. Something else that came out of that day was the rate of change of technology in the mobile/UX world - typically "stuff" gets old within a year or two, which is great for us techies but a nightmare for SAP and it's customers who have to invest lots of time and money and typically aren't agile enough to keep up.
Just my thoughts/ramblings...
Gareth.
*It always is when you say it quick, however in practice, as we all know, it often really isn't it.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
>>>>
this has been something i have always considered to be quite a tricky one. consolidation of the integration platform is something else i do hope SAP thinks about. At a highlevel, investing onto PO for a seamless integration platform makes sense to me. But i know a lot of people disagree
This is what I make of it at this point in time - based on what I read in SAP TechEd presentations and Sybase InfoCenter documentation :
1. MBOs are not supported in SMP 3.0 - Some SAP Notes suggest it may be re-introduced as part of a SP for backward compatibility with existing customer solutions, but the in general the conclusions of the initial poster are correct : MBOs are dead. OData is the way to go.
2. SDK Documentation for migrating existing MBO-based apps simply states that they have to be redeveloped using the OData SDK. The word 'migration' I think is rather euphemistic, I think, as migration in this case basically means : re-develop your data access layer from scratch and update your native UI development projects to use that new OData layer. Not a quick fix.
3. Data integration of heterogeneous back-ends (JDBC, SOAP, REST, OData) is deferred to the SAP's incarnations of OData provider platforms (read : SAP NW Gateway and now also the new Integration Gateway Java delivered as a feature on the SMP 3.0 OSGi platform. So, the deprecation of MBOs also has to do with the more general idea taking shape that data integration as such is not a role for the SMP platform. Data Integration is for the Gateway products.
4. Offline OData is basically a combination of an extension of the client-side API in the OData consumption toolkits (NWGPA) on the one hand and new capabilities built into the SAP NW Gateway (ABAP) - whether the Integration Gateway JAVA also supports it is uncertain at this point in time (documentation only mentions the ABAP incarnation of the SAP's OData provider). One of the key features for Offline OData is the support for DELTA queries which allows a reduction of the payloads between devices and backends as it will allow the client to only download and merge the changes applied to a collection since the previous request. On the client-side, the caching capability is extended with support for offline transactions against cached OData collections (adding, updating and deleting entries). Change records are kept in a separate client-side cache - the so called 'local cache' as opposed to the 'server-cache' - which acts as a write-back queue for pushing updates from the device back to the backend. With the additional support for batch-updates on the SAP NW Gateway side, offline changes can thus be submitted for replay in the backend with a single request.
5. One has to be aware of the limitations of Offline OData at this point in time :
Offline OData is NOT an ORM implementation - it does not auto-magically create an SQL database on device (as the proprietary solutions did - MBO or Agentry MDO). The serialization is XML document based. If you want a device side database, the native developer will have to code it manually.Because of this limitation, Offline OData may not be suitable for more complex, larger data sets because of its limited query options : you either get an entry by ID or you have to load an entire collection into memory before querying it. That's not exactly scalable.
Message was edited by: Theo Paesen
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Offline OData is NOT an ORM implementation - it does not auto-magically create an SQL database on device (as the proprietary solutions did - MBO or Agentry MDO). The serialization is XML document based. If you want a device side database, the native developer will have to code it manually.Because of this limitation, Offline OData may not be suitable for more complex, larger data sets because of its limited query options : you either get an entry by ID or you have to load an entire collection into memory before querying it. That's not exactly scalable.
And that's why I won't bother with these over expensive solutions that are actually little better than custom development on a open platform with no licensing. There is little to gain by using SMP, I will just use NW Gateway which comes with Netweaver, and that's all.
I wonder how SAP has so many customers actually paying for SMP.
Hi Joao,
I completely agree with you.Its getting very difficult and frustrating for developers to learn new technology every six months.Lots of time and effort are being invested in creating MBO based apps.And now SAP is discouraging MBO.Few months later they might come with something else and again we have to struggle and build the same apps from the scratch over and over again.Clients are even more confused to invest in SMP version.
Hi Shabarish,
My understanding is MBOs won't disappear in SMP 3.0 but will not be enhanced going forward. I've had my fair share of struggles with MBOs as well. Once you get them working, they're a charm but can be a frustrating journey. I just hope they don't go the Sylco exchange data model route which is far too complex considering synchronization with MBOs comes out of the box with minimal effort in configuration.
Raj
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Last week in Amsterdam I heard various statements - read : contradictions - on the topic of MBOs in the upcoming versions of SMP.
One relevant fact is SAP Note 1936078 which states that SMP 3.0 will not be fully backward compatible with SMP 2.3. Two items are listed :
- old HWC apps will no longer be supported (obvious, cause KAPSEL will take over)
but also this statement :
Offline applications implemented via Mobile Business Objects (MBOs) must be migrated to the offline feature implemented using OData
So, this again raises the question : does SAP intend to re-introduce support for MBO based apps in the next service packs of SMP 3.0 and does this restriction only apply to the initial release of SMP 3.0. ?
or :
- as the initial poster of this thread already suspected : are MBOs indeed R.I.P for good ?
Customers with an investment in MBO based apps better wait a while before moving to SMP 3.0, I would say...
and the second question is : what of this OFFLINE ODATA is already a reality today ? - does OFFLINE ODATA automatically also mean : full and clever support on the SAP Netweaver Gateway for data synchronization between device and backend ?
An answer to your second question, NO, offline ODATA is not a reality today and no date has been communicated when it will be available. If you need it now then you will need to code it yourself Customer with a big investment in MBOs are stuck using the SUP 2.3 for the moment but in reality I dont believe there are so many customers that actually have made that investment. Which means there is very little reason for SAP to invest in supporting them in upcoming service packs of SMP 3, so yes MBOs are dead IMO.
I guess the reason they are not widely used is due to the fact that SUP has not been widely deployed, sadly. The reasons for that are many but the fact that SAP themselves moved away from MBO based development towards Gateway apps was an early hint that investing too much into MBOs would not have been the smartest move going forward.
Hi Shabarish,
This has been a real concern right from the time SMP 3.0 was announced. There was a confusion around discontinuing MBO support in this version and we've been trying hard to make sense of it. The concern can be split across two avenues:
So coming to the point, for building new apps, if we broadly categorise them into online and offline apps (which is your original question),
So there is a definite push towards using the Agentry platform for complex offline apps until Offline OData will be made available.
Here's a recent webinar on SMP which also throws a lot of light on these architectural options [http://scn.sap.com/thread/3426503]
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
The real problem is communication and a clear roadmap. At the time of the Syclo aquisition everyone was confused and I remember asking a SAP representative what was their roadmap for integrating Syclo and how Sybase and Syclo would coexist.
The answer at the time, was that Syclo was bought for the apps, but the development platform for new apps would still be SUP. That's the exact opposite of what happened, Syclo is now being pushed for offline development.
That's the real problem, both consultancy firm and clients invest in a "new" technology that becomes obsolete after few months. Next time they will turn to competitors like Oracle.
I guess we will get answers to most of these questions at Teched in 2 weeks time. I believe we will see some really good UI improvements on the Syclo Agentry side in the coming months, someone mentioned OpenUI as the new "technology" but currently there is close to zero info generally available on that. There are other issues that SAP has to solve regarding Agentry, only having the option of running one agentry app on a device, multiple apps running on the same agentry server (currently one app pr. server) etc.
From what I have heard from SAP key people this year is that the Syclo agentry and the MDO (mobile data objects) really does handle offline and synchronization far better than Sybase/MBOs are capable of.
Going forward it is clear that the NW gateway has a major role to play in SAP Mobility and it is already a mature product with the latest release covering most requirements for online apps.
I agree that there is a communication gap from SAP to the developers, architects that are trying to understand, sell and support these solutions and this gap is really big when it comes to the more complex apps and technologies supporting them. Getting information about Syclo has been extra hard to get while the lighter apps and technologies like the Gateway, SAPUI5 etc. has been really well communicated.
To the original question, Yes MBO´s are dying that seems 100% certain at this point and MDO´s short term and Odata/Gateway long term are replacing them.
I am very new to the mobility space and glad that i started this discussion
primarily the challenge i believe is been with the multiple acquisitions of the recent past and a desperate attempt to consolidate onto a mobile platform.
Thanks to everyone who have shared their comments. I will be at TechEd LV , carrying with me tons of questions i would hope SAP answers.
I'm about 99% sure MBO's won't disappear with SMP 3.0 and that off-line OData will come in a later SP of 3.0 (I've seen SP7 mentioned so far.)
I see the BPM API also has an OData component with the latest EHP1 SP9 release, so SAP are definitely following up on this becoming their interface protocol of choice (which is great in my eyes.)
Have a look at this article I tweeted the other day for some alternatives for off-line storage.
Hope this helps,
Gareth.
Yup, my recommendation would be to go for Netweaver Gateway to expose your backend functions using OData or RESTful Web services. On the mobile app front, a Cordova-based app (assuming for multiple platforms) using some local data persistence mechanism built into it would do the trick. But be wary, for complex operations, that will require some serious effort.
Btw, thanks for this discussion!
Hi Shabarish,
Interesting discussion. Personally I have never been a big fan of MBO's and hearing this doesn't surprise me that they're dying.
I started using a different approach, which I wrote a blog about a few months ago, see here: Building offline native apps with Gateway and LBO’s
As I said, I'm not a big fan of MBO's but I am a fan of LBO's (Local Business Objects). I have designed my entire app which I recently completed using LBO's for persisting the data. I did look at the SQL Lite approach, however that really would have required a massive amount of manual coding.
The approach of using LBO's has really worked great, you still get all the benefits of the Object API generated code that you would get from MBO's, obviously excluding the synchronization to the backend, as that is done using the OData SDK. It did however require a bit of manually coding of mapping the gateway service proxy objects to the LBO's. However that really wasn't too much.
I also had to handle the delta updates programatically.
I think the greatest benefit comes in now that I am building an iOS version of the app. I can just generated the Object API code for iOS without needing to redesign the data model for each operating system. This whole approach has worked really well for me.
I just posted the blog to the completed app last night Mobile Workflow using SAP Netweaver Gateway–Part 1
Hope this is helpful.
Cheers,
Brad
Thanks everyone for plugging into this discussion.
If I may summarize;
1. MBOs are dead from 3.0 so either stick with 2.3 or think about alternatives
2. In 3.0, we dont really expect much of an OData based offline support
3. Agentry is an option but seems quite a pain and overkill in comparison to what MBO used to provide.
At a high level, i will conclude that the strategy from SAP right now isnt very compelling. We hear about OData based offline capabilities being developed from SAP and perhaps Gateway might even support non SAP integration. As of now, we might have to wait and watch and for those going onto 3.0 to be cautious about what features would be missed.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
I agree that the strategy from SAP is not very compelling, and I've been pretty critical of it since the Sybase aquisition, which they have now effectively killed with the end of MBO (the defining characteristic of SUP).
What is left of Sybase's SUP anyway? Nothing really, OData and Gateway were introduced after the aquisition.
You just can't trust your investment into something that changes completly with every version, so I'll stick to Gateway which is now "free" for named users, while providing offline with custom developed code (Core Data in iOS for example).
PS: Now I remember I actually payed for SUP training focused on MBOs, what a waste of money, no way I'm paying for Syclo Agentry only for it to be killed of in 6 months.
, not to derail the topic but the architecture and prerequisites for SAP standard apps differ for various apps. The ReX app is definitely a native MBO based app and does not require NetWeaver Gateway as a prerequisite. Have a look at this blog from to begin with http://scn.sap.com/community/mobile/blog/2012/07/06/sap-rex-which-skills-are-required-to-implement-s...
The ReX Administrators guide further details this http://service.sap.com/~form/sapnet?_SHORTKEY=01100035870000760929&_SCENARIO=01100035870000000202&_O...
Cheers
Phillip
My mistake Philip. The app you are referring to (REX) is based on MBOs. I had configured a lot of SAP standard apps as a part of RDS (rapid deployment solutions), and all of them were based on Odata. Ex: SAP ERP Customer and Contacts,SAP ERP Customer Financial Fact Sheet, SAP ERP Order Status, SAP HCM Employee Lookup, SAP HCM Leave Request etc...there is a big list. After this I thought that all the apps from SAP are based on Odata. Good to know that SAP released apps based on MBOs too.
In that case SAP may come up with a different version of the that runs of Odata.
- Midhun VP
Learning a new mobile platform is a huge investment for both the customers and the consultants that implement them. It's just not pratical to invest in a wide range of solutions.
Getting proficient on SUP and then seeing it go obsolete in a short period of time really hurts your bottomline. Add to that the fact that you recommended a stable and scalable platform that has now gone obsolete.
I won't be caught recommending Syclo in the next few months.
PS: Playing with the toys is fun, the problem is once you want to get down to business.
Hi Joao,
It does not sound like you will be recommending any particular technology for mobile applications based on not knowing a firm roadmap from SAP. Am I understanding you correctly?
I am not convinced that there is a one size fits all answer or even two or three sizes. I do believe that the Agentry suite of apps and its platform will not go away anytime soon, mainly due to the large number of productive installations. I know that SMP 2.3 manages Agentry apps using the same technology as it uses (mostly) for managing other apps, but does not provide a unified development platform where you can develop Agentry apps along with, for example, HWC, MBO, SAPUI5, etc. Not sure when a unified development platform will come along, nor am I convinced that it is necessarily a useful effort to pursue.
I do understand the frustration people feel for MBOs being decremented after many have implemented solutions based on that technology. I even more agree that having a good clear and stable idea of where the technology is going within SAP is necessary to make good investments in training and resources. I have asked some of the invested management and key technology personnel to take a look at this discussion in the hopes that someone will clarify the future for you and others.
Regards, Mike
Yes you're correct, I won't be recommending a particular solution for the time being. There was a time I was pushing SUP as a platform for the future, and to be honest I'm glad none of my prospective clientes actually bought it.
I've wasted more then 1000 euros on MBO training, so that also hurts, especially since SAP was pushing it as "the future of enterprise mobility".
For the time being, until SAP stabilizes it's roadmap, I'll recommend Gateway + custom development. It's cheaper, and the maintenance cost is offset by the fact that SAP doesn't provide safety for your investment in an expensive platform.
Nothing official, or any calendar dates, but don't put MBO's on the dead cart quite yet(http://www.youtube.com/watch?v=Sh8mNjeuyV4). Stay tuned, and hope for solution management comments on this topic and timing for SMP 3.0.
You are right Jeremy. Officially SAP didn't mention that they would stop MBOs but SAP is discouraging the MBO approach and the recommended way of development is Odata. This is mentioned in recent webinars conducted by SAP. A good example is that the Windows 8 is not gonna support MBO but supports Odata.
-Midhun VP
According to SAP they haven't planned to support Windows 8 with MBO API approach. That means we can't create Windows 8 mobile apps based on MBOs. If we want a Windows 8 mobile app we have to go for Odata.
The above information is based on the webinar conducted by SAP on Windows 8.
-Midhun VP
Well according the blog already posted it seems to be true, but I'm not surprised. I've tried using SUP MBO but they provided little flexibility. There were times where a needed a dynamic data model, which doesn't work with MBOs.
For most scenarios Gateway + OData is enough, and with the Gateway updated licencing (named users get it free), for me at least it's the number 1 solution for mobility.
PS: To bad so few SAP shops have Gateway installed so I have to use SOAP....
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Syclo agentry will be the solution for offline and more complex apps at least short term.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
http://scn.sap.com/community/mobile/blog/2013/08/21/get-your-hands-on-sap-mobile-platform-30
Please read yhe comments of this blog on this.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Midhun -
We dont think MBOs will be supported going forward. So that leaves us with OData supporting offline capabilities (an assumption at this point).
Now when we talk OData, contextually we need to have NW Gateway. Gateway means exposing SAP backend. So Q. what happens to the non SAP backend?
Hope you get my drift here.
Hi Shabarish,
Good question. Some answers are in this thread http://scn.sap.com/thread/3386794
For non-SAP backend systems it is not unthinkable that SAP will build OData capability into SMP3, sort of like a gateway java, a mobile middleware that caters for numerous integration scenarios.
Fortunately, with TechEd and SMP3 ramp-up starting soon we should get clarity on these and other questions soon.
Cheers
Phillip
Hi Narendra,
Agentry is one platform among others where you can develop custom mobile applications. In a bit of a surprise to me, I have met a number of Partners who have gone ahead and developed custom apps in Agentry, although your comment is how most developers view Agentry. However, that does not mean it is not sometimes used for that purpose.
As far as offline focused applications, Agentry has the advantage that it is architected specifically for offline operational scenarios. Are there other options? Sure. I don't recommend using it without a good understanding of the other choices. Actually, that is usually a good recommendation regardless of the topic, but still...
Regards, Mike
SAP Rapid Innovation Group - RIG
Even it is possible to develop custom apps using Agentry it's really very difficult. When compared to SUP the custom app development is using Syclo is more complex and it takes more time. This could be a reason people use Syclo for standard apps. I haven't seen a custom app built using Syclo even.
- Midhun VP
Hi Midhun,
Since I am more familiar with doing apps in Syclo than SUP, it is hard for me to either agree or disagree. I don't think it is that difficult, but it does look more complicated on the surface.
One thing a partner who had done it told me, was that once they had finished doing the development, there was less headache with making it work in offline mode and fewer tweaks with synchronization.
Regards, Mike
You may be right from your experience Mike. Even I haven't worked on a development of new Syclo app I used to have regular discussions with Syclo developers. What I have seen is that in the case of SUP once we develop the MBOs there is an option for generating the APIs and that can be directly called while device coding, but in the case of Syclo we need to write the classes too (that includes complete logic ) from end to end ( I may be wrong) and that is a tedious task.
- Midhun VP
Hi Midhun and Narenda,
One of my co-RIG members, Alexandra Linde-Dunmire, just published http://scn.sap.com/docs/DOC-47550 so that people would have some help in building Agentry based apps from scratch. This is essentially the same design approach which was presented on Days 2 & 3 during the RIG SMP 2.3 Partner Enablement workshops (done globally starting in April and finishing in August). It is a longer version than several of us will be presenting in various TechEd events as MOB 262 (4 hour hands on workshop). It includes the installation, configuration of the development environment as well as the creation of the project. Please look it over, test it, and see if it is something that might be less difficult than you might expect. It provides the java code snippets, Agentry AddOn, and the ABAP custom code (function modules, etc.) necessary to build your own application as well as test it on your movile device.
Regards, Mike
SAP Rapid Innovation Group - RIG
"I don't believe Agentry is the right platform to develop custom mobile applications. It is meant for supporting only the packaged Syclo applications.
For all offline mobile applications development, developers will have to look for alternative strategies such oData and JSON."
Narendra, what makes you say you don't believe Agentry is the right platform for creating mobile apps from scratch? Specific examples?
It most certainly is a viable option and handles offline processing very well in complex applications with large data sets. There are MANY Agentry apps built from scratch used in production environments out there in the world. Whatever the marketing says, painting Agentry into a corner as just a way to extend the existing SAP portfolio apps is not doing it justice as the rapid-to-production mobile development tool that it is.
Jason Latko - Senior Product Developer at SAP
User | Count |
---|---|
84 | |
10 | |
10 | |
10 | |
7 | |
6 | |
6 | |
5 | |
4 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.