Tech Ed Barcelona, Not Much Detecting, 2 Days Out Of 3.
Getting to Barcelona for TechEd 2018 was perhaps more travel and planning than flying across the US to Las Vegas, yet much more satisfying. I had a too full agenda planned. Alas, left my printed paper agenda with hand-scrawled updates in the hotel room. At check-in they wanted a photo ID and confirmation “paperwork”. While I didn’t carry my US driver’s license as I was taking public transit or walking, I could not get into the SAP Mentor breakfast without a wifi password. I persisted, making it to breakfast while it was still warm.
Conference app was needed for orientation, no paper copy available. Posted maps were helpful but after downloading the app I had several crashes when trying to drill down to an “all day activity” detail [SS1]. Went through the recommended escalation path via the app feedback button, which simply launched a gmail pre-populated with the third party (non-SAP) app team support address. Their reply was not especially helpful (“uninstall and reinstall”). I didn’t do that as the next time I tried those links worked. Should SAP roll their own code? Perhaps. Or just buy that company and make it another division.
Some slides were only partial decks (not playing with the full deck) and the link for the remainder takes you to SAP.com, which I didn’t have set up on my phone, only my laptop. [SS2]
The major other glitch I noticed was when zooming in or out on the maps. The wide view was needed to tell how far from the building corners anything was; narrow view to find the room numbers. Mainly needed that for small spaces such as road maps, as there were large posted signs for the primary lecture (“L”) and hands on sessions (“H”). SS3 shows the partly zoomed map with only pins, no numbers. Other legibility issues were noticeable, perhaps owing to the Android version, or perhaps a raster image problem.
I got several pictures taken by the SAP community team, who then forwarded 3 that they liked of those. One is now adorning my SAP community profile. She made me smile, which I think is not my natural state (Hi, Eddy).
Usability session are always enjoyable as they turn the tables around and let you be an active audience member instead of a passive consumer of a slide deck push. Mine was on the API Data Hub,though confidentiality agreements bar me from sharing any further details. I can say I had signed up to observe, not type, given my lack of hands-on experience in this space. When the second person who signed up did not appear, I was given a field promotion and sat in the hot seat. At first I thought being a novice would make me less of a useful tester; after speaking with Steve Rumsby I came to the conclusion that people unfamiliar with the application area are just as useful, since who knows when you might be called on to take over a role should someone leave their job unexpectedly.
For a few minutes, I lurked around the edges of the Community Influencer panel session with Daniel Graversen , Florian Henninger , Karin Tillotson / Karin Tillotson , and Anne Kathrine Petterøe / Anne Kathrine Petteroe
i ran into a former contact from IBM Germany, now with Lenovo, whom I’ve worked with in the past on system sizing and related infrastructure challenges. Good to catch up and get a level set on the changing world of enterprise software.
SESSION – CGE107 The Future Data Center: Solutions for Intelligent Enterprise Applications
The conference application showed one speaker; two presented though. I had expected this session to be more general than it was, talking about trends and projections for what future paths may be. Instead it was rather more of a presales talk about a specific set of offerings. Probably useful for those in a point of making decisions on impending changes, not as useful for those in a comfortable steady state, or already mapped out directions.
One useful takeaway was around concerns for disaster recovery sites, such than in country data gdprly replication might be legally required. Conscientious types have already addressed this; the rest had better catch on quickly.
When I found myself with a few minutes between planned sessions I stopped by to chat with @Qmacro (DJ) for app testing. He asked if I had an SAP Cloud ID? What I didn’t know since these things tend to have a different name every so often or depending to whom you speak was he meant my P number! I tried to connect, the screen said my ID was already in use, which was the clue stick. Then I asked for a password reset. Which has never arrived. Waste of time there. My fault, I should have dragged out my PC for a connection lookup. Ah well, not my cuppa.
Craig Cmehil Community session (no pictures, sorry). I asked one question about “topic areas” versus “top level tags” and learned that topic areas have curated pages that are being improved (“as we speak”) and that tags are chosen by posters, who may need to be nudged into the correct space, as well as new spaces (such as Basis) needing to be created, after the SAP internal train has been put together in the way of clue.
SAP On Oracle Cloud was an interesting session to attend, given the somewhat competitive areas they and SAP are in. My main SAP initial experience was as an Oracle DBA; this topic is quite different. For an online commentary, check starting with my tweet.
DAT108 An Outlook on Data Integration Tools from SAP: When to Use What
A short review of this session – the speaker was very sharp, the content was broad and deep, and I left not knowing much more than when I started. I took many notes, and could probably describe a few key concepts if I could peek at the slides. Otherwise, not much memorable to me. A dry topic, for sure, so not my finest moment.
One takeaway, or lightbulb moment, was the comparison of 2 primary data moving tools – SDI (SAP Data Integration?) reads database transaction logs, while SLT (SAP Landscape Transformation) uses database triggers. Probably my rusticated DBA skills understood those differences more clearly than the cute imagery so prevalent in modern slide decks, plus the speaker dove into the fundamental impacts of each,where log scraping has no discernible performance hits (though is only applicable to some areas), and triggers may have measurable application interference (but have broader applicability).
I missed a road map session on WIndows (er,”SAP Analysis for Microsoft Office: Road Map at SAP Tech Ed”) as the room was full. Tammy Powlas said it was repeated at least 3 times in Las Vegas.
I did get to attend a road map on a topic I’m more familiar with (in passing, not live), the SAP Landscape Manager. I subsequently learned there’s more than one use of the term landscape (i.e., also in transport management) so one must be careful to discern which one is in play. The session was good, forward looking statements were pretty clear (“anyone actually using IPV6 in an SAP production instance please step forward”) and the questions were answered clearly and cleanly (though in this as in all other sessions I attended the speakers failed to repeat the question for the rest of us…).
Last screenshot of the day [SS4] and a major gripe: don’t put ads into the conference app. When turned my device sideways (in order to, you know, see the content), one-third of the screen was blocked by a pop-up (the greyed-out area). One third was already eaten by title bars, leaving only one line of legible but useless text. Just don’t.
Back of the house
This day started with a hotel breakfast, followed by a 30-ish minute walk with Steve Rumsby . We had walked on Tuesday, but with daylight an hour later this one seemed easier (and I now knew the route, which always helps.
My first session was going to be:
A Technical Deep Dive into the ABAP RESTful Programming Model
But when Fred Verheul and I arrived, the prior session wasn’t emptying out, many were staying in their seats for a followup from the previous one, and as I had a 10:00 meeting i didn’t want to miss I didn’t want to occupy a chair, denying the chance to someone who clearly needed to consume the content more than I did. Alas, rules. Can’t stand up to see the session. I tried the “I just want to take some pictures gambit”. Didn’t have the requisite big lens, photo vest, and a gaggle of colored badges hanging around my neck, just a Nikon Coolpix (Craig said it is a relic – it isn’t really, it has GPS tagging features).
Manage and Govern Your Business Rules Centrally in a Hybrid Landscape
Quick twitter convo cut/paste:
hmm. Looks easy enough but guessing there's a lot to do outside the app.
Phil Gleadhill Retweeted Jim Spath
Agree, easy to say, very hard to execute… Will require a lot of governance and support in a hybrid multi-component landscape. Any predictions anyone?
I had to leave early for another session, but my first impression was migration/conversion of “legacy” rules to the cloud repository/framework is a well built path. Pushing from a single source to run time instances seems workable.
Thx Jim, much appreciated…
My takeaway from this sessions is primarily that SAP appears to have a handle on pulling together a long legacy of business rules and the tools used to capture and distribute them, with a coherent strategy. Unlike some hand waving or back flipping I’ve perceived in other enterprise software zones, this seemed concrete. Okay, the demo was contrived with shopping cart, which is not new ground in any sense, but the examples of discounts (or campaigns) lines up pretty well with the oddly constructed condition records malarkey we dealt with 20 years ago rolling out SAP 3.x.
I moved onto the the Mentor session on Data Integration.
[redacted]. You can see a picture, though, of how intense we look.
When Size Does Matter: Data Aging and More in SAP
This session initially sounded like a data archiving/data prevention topic, yet that was perhaps a small fraction of the content. Perhaps too much time was spent dancing around the latest “how may CPUs and how much RAM does it take to plug in a HANA light bulb” but for the general technical audience probably not a bad idea to set the stage for “now what do you do if it’s still dark?” Some of the answers made pretty good sense, except when the answer came back with “oh, just build a cluster”. This direction came back to the Tru64 clusters I worked with a couple decades back, and discussions of “NUMA” – non-uniform memory access; in other words, everything’s great if the data I need is on the cluster the CPU is on, not so great if on a different node. The technical solution (keep the data close to where it’s needed, don’t do joins across the clusters) is strongly reminiscent of the table buffers used for a very long time in SAP R/3 application servers, with the associated cache coherency concerns.
I’m hoping hardware advances in individual node sizes preclude customers from needing to go to a cluster route, as the human touches and tuning go against the grain.
On the plus side, using clusters for accessing read-only data appears to be more elegant, where the application server can just pick up the data on the non-primary node and deliver it to the results flow with only a SQL hint (also a human touch, except less often I’d expect.
A couple more community sessions – one with Matt Fraser and Christian Braukmüller , and one with female influencers. I attended the former, telling a story about me and Blag and Marilyn Pratt . I guess you had to be there. Some good followups later. I promised Christian more details on the Algonquin Round Table (of Dorothy Parker fame). [ https://www.biography.com/news/algonquin-round-table-members ]. How does that relate to SAP Inside Track, you might ask?
Witalij Rudnicki talked about graph theory in a mini-code-jam-hands-on. I demurred, while learning quite a bit with hands-off. Just as with the subsequent JSON in HANA session, I wonder how much performance hit there might be from straying outside the carefully (cough) engineered column or row store database models, time will tell if these are best cases. Or if another approach may be better. I lied the example with the ski routes. Vitaliy’s pisture had to be tweaked a bit for the ski googles to even be visible; hope it looks good.
Konigsberg bridges = Graph Theory, Knight’s Tour. (Solve in HANA!?)
DAT114 Discover the JSON Document Store in SAP HANA
My last session of the day, I expect to be as crowded as the first, as I’ve heard JSON mentioned about as much as REST in various non-SAP contexts. Not that I was actually counting, and my intuition was either way off, or the time difference of the sessions was a factor, or maybe the speakers themselves. Or REST really is a bigger deal than I knew.
Anyway, the session took a while to get to a point. The syntax really reminded me of Perl; maybe that’s the developers background, maybe it’s just a coincidence (parallel evolution), or who knows. The demo was canned; good for repeating later, bad for proving realtime performance in a demo.
That’s it for the main event. A few more gripes and grips at the end.
Big crowd for REST
Mentors in session.
Scale out to parallelism
Graph theory and practice
JSON in the house
Last but not least.
Screenshot 5 shows an apparently available presentation that wasn’t. The error message was curt and not so useful. So I don’t know if it’s me, or the system, or what.
Screenshot 6 is more subtle. In the conference app you can add a session to your calendar (google, in my case). Sort of. The “Save” button just goes ahead and saves whatever. If you don’t look, you won’t notice that the time of the event is not the time it shows in the conference app, it’s whatever the current time is. You can see i ended up with 3 appointments for one session, at 2 different hours. That happened too many times, as I was in a hurry to either get somewhere, or trying to pay attention to the one I was in but make a note for a later session I would go to next or maybe a past one I wanted to review the slides for before they slipped my mind.
Hint: the appointment with “DAT114” in the title came from the agenda builder “MyAgenda”, the others came from the dark side “Session Catalog”.
p.s. Thanks to the black hats in the back of the house for serving food and cleaning up after.