Another year and another teched arrives and then passes us by, as an end customer representative I consider myself lucky and very privileged to be able to attend once every 3 to 4 years or so, I find Teched is always inspiring in a myriad of ways, and always worth the investment my company provides, my last visit was 2015 and so here in 2018 things certainly appear to have moved on.
Teched always seems to start off with a bang, and long may this trend continue. This year Björn Goerke (SAP’s CTO) once again led the presentation, at my previous teched he was trying to escape Mars (in the Martian themed keynote) so it came as a bit of a shock to see that he is now considering training as an astronaut, that seems somewhat back to front there Björn (no wonder you got stuck on Mars back in 2015 Björn ? ).
It was an entertaining presentation from all involved, that perhaps, considering the main audience, I felt, could have dealt with the topics in a little more technical depth. It would be great to see the inner workings of the intelligence suite, and Leonardo, which I think to many of us still appears as a wonderfully described magicians black box of tricks.
The keynote successfully set the clear themes for this years teched, Cloud based services featuring a key part fully supported by the somewhat frustratingly mysterious “intelligent suite”. The clear message is now the move towards a cloud platform really opens up SAPs API’s to connect with the outside world, already over 150 external APIs to other major cloud services that can be intelligently integrated with SAPs cloud platform to SAPs own backend, and can all be combined and enabled with CDS featuring a major role in the cloud platforms open programming model, with a business domain focus, introducing new implementation possibilities. Things do seem to be all coming together to mean that we end customers can soon just plug in the components that we need together with the tools and APIs that we have on our own backend systems to really produce innovative solutions, relevant to our own solutions, with relatively little development effort. So from a developer centric perspective it is an exciting time to be around that is for sure, and that is the main take home message I took with me from the keynote. SAP is truly rapidly becoming a modern and very agile development toolset.
We are finally being encouraged by SAP to “Keep the core clean”, which is a message I like to try to ensure our developers understand, and which is perhaps a message that would have been more useful about 20 years ago, but still better late than never I suppose. Now with many of us about to make the long and difficult journey to an S/4 Hana (possibly cloudy) world, through stormy seas of hundreds of user exits, enhancements and modifications, it definitely seems like the time to reemphasize this point, that world is a world we must leave well behind us, and move to the newer cleaner worlds of S/4.
The intelligent robotic process automation was the one SAP-LABs sneak preview we got of the intelligent suite, and it appears that this will enable us to record activities in FIORI almost as if it is some super modern SM35 transaction, but instead of backend GUI recordings, it is recording our behaviour within the FIORI front end and then able to generate automated scripts that can later be triggered by a “Smart chatbot” style interface, so a sortof super evolved SM35 is my takeaway here ?, probably I would be understanding it better with a bit more detailed explanation, but hey-ho, enough of that.
Björns speech was wrapped up with an interview from one of Nasa’s youngest ever qualified astronauts which was surely inspirational, although I felt could probably have been tied into the overall presentation in a smarter way than it was. Really the new toolsets has the potential to do so much more than offer a smarter employment engine, perhaps for example supporting various other initiatives at NASA, it is rocket science after all… or possible future space agency work in sending rockets around the solar system, really the possibilities appear only limited by our own imaginations.
Still it was certainly fascinating to hear the experiences of this youngest of NASA astronauts and encouraging to us all how far anyone can come with the right approach to our own development.
Then we finished with Björn encouraging us all to pull up a chair to step up and reach beyond, which was surely a slight missed opportunity to get a room full of ABAPer’s and SAPtechs to stand on their chairs, still it set a good message to kick off the next 2.5 days of intense learning opportunities for which I am most grateful.
Fast forward on to the sessions then and a long moving floor assisted walk across the convention center. My current role is as a development architect which really means I had a hard time to pick one of the official journeys, as I find my role is fairly diverse, and I have an interest crossing many of the teched tracks, still I felt it a good idea to get a little more familiar with the cloud platform in ABAP, and see that it is fairly straightforward to add new ABAP apps as PaaS from the cloud although integration to the backend seems very dependent on an SAP restricted white list of APIs that (if I understood correctly) is currently limited to 400 or so APIs, I am hoping that I got this part wrong and that customer defined APIs can be added (otherwise it appears to be very restricted in options to backed integrations), deployment is all controlled via GIT code management for version management, with all ABAP development being eclipse based. Unfortunately not much time was allotted to questions in this session and so I did not get the chance to ask for more details here.
Probably the highlight of the day for me was attending Karl Kesslers CAN319 optimize your custom code for SAP HANA. One of the things that I found fascinating was to learn how many in the room had started using CDS – and really it seems just a handful of the room full of developers, and to me this spells out that Teched needs to also have a major focus on learning current technologies first, a lot of the time it appears to focus on the areas that SAP are most interested in selling next, and unfortunately not always what skills will most benefit users in making the steps necessary to reach these new techs.
That said the session with Karl I found really useful, and particularly appreciated his answer to my question, on whether we should be building APIs to make use of our CDS views, or if there was a better approach as APIs will loose some of the main benefits of the CDS views. Karl’s answer is that CDS should be layered with a Basic layer of CDS views nearest the DB layer (this layer of views you do not let all developers change, (more the senior devs and architects) or access as it is the base views that everything else is built on, then an intermediate interface layer and then finally a consumer view layer that is exposed to the UI. The CDS in the top two layers are essentially the API available to other developers. Most developers need only work with the interface and consumer layers, and so this seems to parallel fairly well the model view controller paradigm that we are all familiar with. The model being the basic view level, the controller being the interface level and the view the top consumer layer. I now plan to review our overall approach to database table APIs with a design thought to building up CDS views in this 3 layered design pattern (possibly a blog to follow on this topic). The current recommendation now from SAP appears to be first to write open SQL, then next use CDS and finally consider AMDP (which is database proprietary but you really need to be a master of SQL script and also it offers the most flexibility and power). CDS are really empowered with the correct usage of annotations, and so that will be a challenge for all development architects everywhere to ensure that the right approach to these annotations is followed by all devs.
The extending S/4 HANA was also an eye opener into how simple the in-app extensibility now appears to be both in S/4 and in the cloud. Machine learning in Enterprise applications once again chose to highlight that the algorithms behind the scenes that implement the learning are a “Done Deal” and not something that we should concern ourselves about, this appears to be the official SAP line, but reading the latest from MIT press on this topic, and they seem to take a different approach, they are not afraid to explain to the reader the inner workings of these algorithms and that machine learning algorithm development is very much a world in developmental progress… Then SAP are telling us that explaining AI is the key, but how do we explain what is inherently described to us by SAP as something that is essentially an API to the SAP developer… this all seemed a bit back to front to me, which is why I plan to be reading some more MIT press books. More details are needed here to ensure that we can explain the results of these algorithms, and so increase the business confidence in using them.
On day 2 one of the highlight sessions of teched was sitting in a hands-on where Rich Heilman was presenting advanced concepts in SQL Script whilst at the same time pretty much jogging around the room, this presentation technique, whilst being unconventional, I found extremely effective, and it was truly impressive to see somebody teaching a complex topic at breakneck speed whilst at the same time getting some solid exercise in, good on you Rich! 😉 It was a true pleasure to participate in this one, and the topic was an eye opener to me in the powers of Database Procedures and SQLScripts, and as Spider Man once said, with great power comes great responsibility, which I think is why these techniques are not the first you should reach for when designing an application, but when optimal performance is really needed with optimised code pushdown at the DB a necessity, then it is the place to go. One topic I know my colleague was interested in was that the use of imperative statements (for instance an IF statement) in your SQLScript logic can push the entire execution to serial execution (on older HANA versions like the one we have), but there are work arounds to manage this, that the presenters were happy to explain to my colleague. We both left this session with enhanced knowledge and that really was facilitated by both the main presenter and his excellent and somewhat very overqualified support staff (Thomas Ljung and Katharina Shell) a big thank you to all who participated in this memorable session.
Persistent memory in SAP was an eye-opener into the hardware world, where a new form of persistent memory will soon open up the possibilities of larger maximum sizing of SAP HANA footprints as well as new memory techniques for us developers (which was just hinted at – as this was not really a developer session). It should also help us to minimize our overall system downtime, something I know that my own project management and business representatives will greatly appreciate.
Wednesday evening brought on 4 clubs and 4 vibes, and this part of Teched Barcelona has not changed much over the years – and just like the clubs, it does not need to change much to be very successful. An entertaining night was had by all and I think I can still hear the drum beats in my head right now typing this.
We much appreciated the complementary bus ride back to our hotels, that surely saved us a looong walk across town, and help guaranty we made our early sessions the day after.
Those of us oldies that didn’t stay up too late, managed to make the 8:30 start to Data anonymisation, this topic to me is fairly critical, as one of the major concerns to many customers of SAP I would think is the ability to comply with GDPR regulations (particularly in the building of test systems).
To achieve a high quality of development it is imperative to have realistic test data in your test systems, but you cannot just copy production without violating GDPR the minute you open up your test systems to testers. Also this is not just as simple as scrambling identifying fields in the DB, as attributes can also give away identity, instead you need to use techniques such as K-anonymisation and differential privacy. I had heard of these techniques prior to teched, but what I had been confused about was how best to make use of them, this session really helped me to work out a better way for my company to use these techniques, I now think what you do is define the anonymising views in production, and then when you use a tool such as TDMS to build your test system, you point TDMS to the anonymised views and not the tables. This way you can get really good anonymisation with TDMS and without the slowdown of TDMS having to primitively scramble table rows one at a time.
Spending sometime in the community area was fun, and instructive, I also really appreciated the artwork by the resident artist in the community lounge area, who had a true talent in summarising teched in a single picture. It is true what they say, a picture says a thousand words. There were many activities to participate in, too many to list, and the coding area was fun to try out some of the new cloud development features.
After a long but very informative and instructive teched, I left wishing I was back at the start of the week again with all the sessions still to be done, this time I would pick a few alternatives and but certainly revisit some of my highlight sessions, it was that much fun. I have one simple wish for future techeds though, and that is for SAP to have a focus track on teaching us already existing technology that is an enabler to the newer solutions, so a stronger focus on CDS for instance, and back to basics with some of the areas, I think this kind of track many attendees would benefit from greatly.
So whilst its good to have our heads in the clouds, it is also important to remember that many of us are still on premiss HANA under the clouds or possibly pre-HANA, and needing the directions for making the next steps. Sure teched did cover this, but there appeared to me no obvious tracks to identify these sessions easily, they were spread out and a little hidden to the untrained eye, it was a bit hit and miss trying to find them. Overall though these are just minor gripes and this teched will keep me motivated in motivating our team of developers for years to come, so a big thank you goes out to all involved at SAP for organsing such a great conference!
(all photographs by me (a little hobby of mine), slides snippets by SAP).
Teched always seems to start off with a bang, and long may this trend continue. This year Björn Goerke (SAP’s CTO) once again led the presentation, at my previous teched he was trying to escape Mars (in the Martian themed keynote) so it came as a bit of a shock to see that he is now considering training as an astronaut, that seems somewhat back to front there Björn (no wonder you got stuck on Mars back in 2015 Björn ? ).
It was an entertaining presentation from all involved, that perhaps, considering the main audience, I felt, could have dealt with the topics in a little more technical depth. It would be great to see the inner workings of the intelligence suite, and Leonardo, which I think to many of us still appears as a wonderfully described magicians black box of tricks.
The keynote successfully set the clear themes for this years teched, Cloud based services featuring a key part fully supported by the somewhat frustratingly mysterious “intelligent suite”. The clear message is now the move towards a cloud platform really opens up SAPs API’s to connect with the outside world, already over 150 external APIs to other major cloud services that can be intelligently integrated with SAPs cloud platform to SAPs own backend, and can all be combined and enabled with CDS featuring a major role in the cloud platforms open programming model, with a business domain focus, introducing new implementation possibilities. Things do seem to be all coming together to mean that we end customers can soon just plug in the components that we need together with the tools and APIs that we have on our own backend systems to really produce innovative solutions, relevant to our own solutions, with relatively little development effort. So from a developer centric perspective it is an exciting time to be around that is for sure, and that is the main take home message I took with me from the keynote. SAP is truly rapidly becoming a modern and very agile development toolset.
We are finally being encouraged by SAP to “Keep the core clean”, which is a message I like to try to ensure our developers understand, and which is perhaps a message that would have been more useful about 20 years ago, but still better late than never I suppose. Now with many of us about to make the long and difficult journey to an S/4 Hana (possibly cloudy) world, through stormy seas of hundreds of user exits, enhancements and modifications, it definitely seems like the time to reemphasize this point, that world is a world we must leave well behind us, and move to the newer cleaner worlds of S/4.
The intelligent robotic process automation was the one SAP-LABs sneak preview we got of the intelligent suite, and it appears that this will enable us to record activities in FIORI almost as if it is some super modern SM35 transaction, but instead of backend GUI recordings, it is recording our behaviour within the FIORI front end and then able to generate automated scripts that can later be triggered by a “Smart chatbot” style interface, so a sortof super evolved SM35 is my takeaway here ?, probably I would be understanding it better with a bit more detailed explanation, but hey-ho, enough of that.
Björns speech was wrapped up with an interview from one of Nasa’s youngest ever qualified astronauts which was surely inspirational, although I felt could probably have been tied into the overall presentation in a smarter way than it was. Really the new toolsets has the potential to do so much more than offer a smarter employment engine, perhaps for example supporting various other initiatives at NASA, it is rocket science after all… or possible future space agency work in sending rockets around the solar system, really the possibilities appear only limited by our own imaginations.
Still it was certainly fascinating to hear the experiences of this youngest of NASA astronauts and encouraging to us all how far anyone can come with the right approach to our own development.
Then we finished with Björn encouraging us all to pull up a chair to step up and reach beyond, which was surely a slight missed opportunity to get a room full of ABAPer’s and SAPtechs to stand on their chairs, still it set a good message to kick off the next 2.5 days of intense learning opportunities for which I am most grateful.
Fast forward on to the sessions then and a long moving floor assisted walk across the convention center. My current role is as a development architect which really means I had a hard time to pick one of the official journeys, as I find my role is fairly diverse, and I have an interest crossing many of the teched tracks, still I felt it a good idea to get a little more familiar with the cloud platform in ABAP, and see that it is fairly straightforward to add new ABAP apps as PaaS from the cloud although integration to the backend seems very dependent on an SAP restricted white list of APIs that (if I understood correctly) is currently limited to 400 or so APIs, I am hoping that I got this part wrong and that customer defined APIs can be added (otherwise it appears to be very restricted in options to backed integrations), deployment is all controlled via GIT code management for version management, with all ABAP development being eclipse based. Unfortunately not much time was allotted to questions in this session and so I did not get the chance to ask for more details here.
Probably the highlight of the day for me was attending Karl Kesslers CAN319 optimize your custom code for SAP HANA. One of the things that I found fascinating was to learn how many in the room had started using CDS – and really it seems just a handful of the room full of developers, and to me this spells out that Teched needs to also have a major focus on learning current technologies first, a lot of the time it appears to focus on the areas that SAP are most interested in selling next, and unfortunately not always what skills will most benefit users in making the steps necessary to reach these new techs.
That said the session with Karl I found really useful, and particularly appreciated his answer to my question, on whether we should be building APIs to make use of our CDS views, or if there was a better approach as APIs will loose some of the main benefits of the CDS views. Karl’s answer is that CDS should be layered with a Basic layer of CDS views nearest the DB layer (this layer of views you do not let all developers change, (more the senior devs and architects) or access as it is the base views that everything else is built on, then an intermediate interface layer and then finally a consumer view layer that is exposed to the UI. The CDS in the top two layers are essentially the API available to other developers. Most developers need only work with the interface and consumer layers, and so this seems to parallel fairly well the model view controller paradigm that we are all familiar with. The model being the basic view level, the controller being the interface level and the view the top consumer layer. I now plan to review our overall approach to database table APIs with a design thought to building up CDS views in this 3 layered design pattern (possibly a blog to follow on this topic). The current recommendation now from SAP appears to be first to write open SQL, then next use CDS and finally consider AMDP (which is database proprietary but you really need to be a master of SQL script and also it offers the most flexibility and power). CDS are really empowered with the correct usage of annotations, and so that will be a challenge for all development architects everywhere to ensure that the right approach to these annotations is followed by all devs.
The extending S/4 HANA was also an eye opener into how simple the in-app extensibility now appears to be both in S/4 and in the cloud. Machine learning in Enterprise applications once again chose to highlight that the algorithms behind the scenes that implement the learning are a “Done Deal” and not something that we should concern ourselves about, this appears to be the official SAP line, but reading the latest from MIT press on this topic, and they seem to take a different approach, they are not afraid to explain to the reader the inner workings of these algorithms and that machine learning algorithm development is very much a world in developmental progress… Then SAP are telling us that explaining AI is the key, but how do we explain what is inherently described to us by SAP as something that is essentially an API to the SAP developer… this all seemed a bit back to front to me, which is why I plan to be reading some more MIT press books. More details are needed here to ensure that we can explain the results of these algorithms, and so increase the business confidence in using them.
On day 2 one of the highlight sessions of teched was sitting in a hands-on where Rich Heilman was presenting advanced concepts in SQL Script whilst at the same time pretty much jogging around the room, this presentation technique, whilst being unconventional, I found extremely effective, and it was truly impressive to see somebody teaching a complex topic at breakneck speed whilst at the same time getting some solid exercise in, good on you Rich! 😉 It was a true pleasure to participate in this one, and the topic was an eye opener to me in the powers of Database Procedures and SQLScripts, and as Spider Man once said, with great power comes great responsibility, which I think is why these techniques are not the first you should reach for when designing an application, but when optimal performance is really needed with optimised code pushdown at the DB a necessity, then it is the place to go. One topic I know my colleague was interested in was that the use of imperative statements (for instance an IF statement) in your SQLScript logic can push the entire execution to serial execution (on older HANA versions like the one we have), but there are work arounds to manage this, that the presenters were happy to explain to my colleague. We both left this session with enhanced knowledge and that really was facilitated by both the main presenter and his excellent and somewhat very overqualified support staff (Thomas Ljung and Katharina Shell) a big thank you to all who participated in this memorable session.
Persistent memory in SAP was an eye-opener into the hardware world, where a new form of persistent memory will soon open up the possibilities of larger maximum sizing of SAP HANA footprints as well as new memory techniques for us developers (which was just hinted at – as this was not really a developer session). It should also help us to minimize our overall system downtime, something I know that my own project management and business representatives will greatly appreciate.
Wednesday evening brought on 4 clubs and 4 vibes, and this part of Teched Barcelona has not changed much over the years – and just like the clubs, it does not need to change much to be very successful. An entertaining night was had by all and I think I can still hear the drum beats in my head right now typing this.
We much appreciated the complementary bus ride back to our hotels, that surely saved us a looong walk across town, and help guaranty we made our early sessions the day after.
Those of us oldies that didn’t stay up too late, managed to make the 8:30 start to Data anonymisation, this topic to me is fairly critical, as one of the major concerns to many customers of SAP I would think is the ability to comply with GDPR regulations (particularly in the building of test systems).
To achieve a high quality of development it is imperative to have realistic test data in your test systems, but you cannot just copy production without violating GDPR the minute you open up your test systems to testers. Also this is not just as simple as scrambling identifying fields in the DB, as attributes can also give away identity, instead you need to use techniques such as K-anonymisation and differential privacy. I had heard of these techniques prior to teched, but what I had been confused about was how best to make use of them, this session really helped me to work out a better way for my company to use these techniques, I now think what you do is define the anonymising views in production, and then when you use a tool such as TDMS to build your test system, you point TDMS to the anonymised views and not the tables. This way you can get really good anonymisation with TDMS and without the slowdown of TDMS having to primitively scramble table rows one at a time.
Spending sometime in the community area was fun, and instructive, I also really appreciated the artwork by the resident artist in the community lounge area, who had a true talent in summarising teched in a single picture. It is true what they say, a picture says a thousand words. There were many activities to participate in, too many to list, and the coding area was fun to try out some of the new cloud development features.
After a long but very informative and instructive teched, I left wishing I was back at the start of the week again with all the sessions still to be done, this time I would pick a few alternatives and but certainly revisit some of my highlight sessions, it was that much fun. I have one simple wish for future techeds though, and that is for SAP to have a focus track on teaching us already existing technology that is an enabler to the newer solutions, so a stronger focus on CDS for instance, and back to basics with some of the areas, I think this kind of track many attendees would benefit from greatly.
So whilst its good to have our heads in the clouds, it is also important to remember that many of us are still on premiss HANA under the clouds or possibly pre-HANA, and needing the directions for making the next steps. Sure teched did cover this, but there appeared to me no obvious tracks to identify these sessions easily, they were spread out and a little hidden to the untrained eye, it was a bit hit and miss trying to find them. Overall though these are just minor gripes and this teched will keep me motivated in motivating our team of developers for years to come, so a big thank you goes out to all involved at SAP for organsing such a great conference!
(all photographs by me (a little hobby of mine), slides snippets by SAP).
- SAP Managed Tags:
3 Comments
