The last day of tech events is usually the hardest to write about. Sometimes the grind, the travel home, and time pressures cause me to abandon ideas and half-written stories, leaving thoughts dangling. My friend jon.reed makes a disciplined habit of staying behind an extra day at most events to cogitate and produce fully-baked analyses, rather than off-the-cuff, gut reactions to press release snippets. I attended a few sessions today, a full calender on the schedule, but lightly attended compared to the prior days. I'm sure I'll have further thoughts and misplaced references, but here goes.
Getting to the KTPO Centre (yes) this morning, I finally observed these traffic barriers, of a sort, are the standard issue clerk or security guard station work stands. They are just missing a paper ledger book for completeness.
After looking at the uploaded photo, it also reminds me of the Beatles Abbey Road cover. I'm not quite sure why. Anyway, that's my entrance segue.
ITM210 Minimizing Downtime During Software Maintenance With Software Update Manager
I picked this session because it's usually the situation where business interruptions occur, a lot of time is spent preparing for the event (patching), and any technique to cut the total time improves IT services. And there's no HANA in the title. You may have heard "no downtime upgrades" promised by some software executives. This content proves that reality is not quite here yet.
Two major components were presented, as well as a customer success story (bonus points in my ledger) from a multinational firm. First, the near Zero Downtime Maintenance feature, which employs a shadow instance, business transaction logging ("change recording replication" - CRR), and more recently, the ability to do SGEN on the shadow copy instead of against the main system downtime. I didn't pick up more details than that, but SGEN is one of those unpredictable additions to downtime, so removing it from the timeline is a big win.
The second component are "customer transports", which sounds like a descendent of a technique ("customer specific upgrades") used by companies such as Dow Corning years ago to allow the shortest possible outages. As transports are another variable time component of upgrades, pulling as many steps out of the downtime upgrade phase can be a life saver (at the end of 40 hours of no sleep, staff tends to get sluggish and/or sloppy).
As these are customized to some extent, the recommendation was to use them only when required. If a weekend for a support pack is available, skip the extra work of these tools. The other caveat is that, while apparently free to licensed customers, they need to be requested. More in the note 1759080 - "Conditions for SUM including customer transport requests".
P&G project details were shared. Might be a great ASUG Annual conference topic. atul.patankar?
ITM260 How to Handle Landscape Data to Maintain and Manage Your System Landscape
I may have been the first student in the room, and one of the last to leave this session, but I achieved what I have rarely managed in SAP TechEd hands-on sessions - completing all of the exercises. Maybe they left enough time for easily distracted people like me ("Squirrel!), or perhaps this topic is one that I'm more familiar with than others, or perhaps this tool is better designed than a few I've used. The lessons were only about configuring system details, not actually pushing data around or monitoring them. But this needs to go first, and needs to be refreshed whenever the landscape changes.
Unlike Wednesday's hands-on exercises, which were done on virtual desktops located in Walldorf, these were done with standard SAP GUI and Business Client (web pages) on the local classroom PC connected to backend systems in Germany. Neither scenario is totally "cloud-based", but there is a high degree of virtualization that I'd like to understand better. During a cool tour of the "IT room" in Bangalore, we learned the PCs are high-end laptops, so that part of the IT world is still an ongoing maintenance issue (Windows patching, system imaging, etc.)
I took lots of "screen shots" with my cell phone camera, partly to reinforce the printed/online reference material, but also to show timing of the work (typing, thinking, waiting for server calculations). No HANA backend databases in evidence... I'll let the images speak for themselves.
slides
waiting
nearly done waiting
initial software system "visualization"
Inexplicable error (I ignored)
Success (manual said this might fail, so I call win).
Populated system configurations, via the visualization tabs.
Instructor (vinod.vijayakumaran, if I am not mistaken) giving me the green light.
Now, if we could only get one, stable, Solution Manager system working...
No one bugged me for the SAP TechEd 2007 Goo Goo Dolls T Shirt, so I handed it to Jason Lax, per his request on an SCN post.
The tiny badges from the land of SAP Gamification
Inside the TechEd "Nerve Center", getting a tour from IT Head Anand Jayakumar. Cool insights into how this thing works, not to mention real world crises such as a bus driving over and cutting a fiber cable (backhoe I've heard before, but not a bus). Of the 2 wireless networks "Old TechEd" ran about 30% of the total, with "TechEd" taking the rest, not counting the private fiefdoms. Anand said there's a 2GB pipe to the SAP Labs office, with a 450MB line to Walldorf (if I have my units correct).
you are here?
Shop Buddy, Second Opinion
With more time to review and understand this demo application, I went back to the pod. First, let's dismiss the concept that most people would voluntarily share intimate medical details with big-box retail chains. Some might, I guess, but this app was written to demonstrate how to use the technology, not because this is an approach that had great merit.
The number of lines of code, and the time required to build the app, seem meager. Looking at the table and data definitions, this is more like a back-of-the-envelope or front-of-the-napkin development rather than a scalable concept template waiting for customers to grab-and-go. The data is somewhat realistic, rather than totally hallucinatory, but still.
I asked about scaling. In other words, this has a few dozen records. What does it take to get to millions or billions? The answer: there's a HANA sizing guide. No doubt there is (and I've probably skimmed it and forgotten it already), but setting up the backend systems to run this type of application is only a part of a scale-out process. How are backups managed; what's the disaster tolerance like; how do you avoid attacks such as denial of service, ad infinitum.
The other aspect that needs details is the security side, whereby I learned a new term: "Aadhaar". Fellow conference delegates will probably consider me a not having done my homework, so bear with me briefly. This word represents an ID "card" that India uses, or will use (not sure yet). A quick search shows this is apparently intended for identification purposes only, so leveraging that system to mash-up with retail purchase history, medical data, or other profiling is, to my mind, problematic. A link for reference (OK, 2): e-Aadhaar by Unique Identification Authority of India
Check the FAQ: "How will the UIDAI protect against functional creep?". 🙂
SPK9717 Q&A: Post-Copy Automation for SAP NetWeaver BW System Landscapes
Since my schedule squeeze out the main presentation for this topic, I didn't have enough background to understand the base technology. We've used an evolving set of tools for copyback automation, so learning new tips was worth a try. Speaker vasanth.immanuel was quite kind in giving me the elevator speech version of the presentation, and allowing me to toss questions about how it could work in my shop. I promise to review the slides and ask him more intelligent questions than "how does it work?"
What doesn't seem to be in this tool yet are printer tweaks (making sure production devices don't get test output), nor our custom user clone set up for quality systems.
RDP952 Full-Text Search, Fuzzy Search ,Text Analysis, and InfoAccess in SAP HANA
radhika.satish and colleague ankit.kumar10 (?) demonstrating the iPad app called "InfoAcces", which seems like a slick reporting tool, available for free (to HANA licensees...) via iTunes. The earlier part of the session was what I was more interested in - the text searching features that HANA could potentially include. I did not quite put the pieces together about how much manual set up is required to deploy fuzzy searches (the example given was "SOP" instead of "SAP"). I do know Google has a full-fledged algorithm farm in this space (showing you "SAP" instead of "SOP" - click here if your really meant "SOP"). Is SAP trying to reinvent the wheel for HANA customers to roll?
This way to the egress.
People were taking pictures in front of the "TechEd Live" studio (and finally being chased away by the anti-photo-bomb-squad) as well as other locations. This one seemed to be a favorite. I also had mine taken by an attendee I met (and interviewed) during my 2011 TechEd visit to Bangalore.
Hopefully, these fuel cans are refilled and used again. The food was awesome at this event. I just wish I had had more time to enjoy it, rather than darting through a line and speeding back to sessions.
A short Mentor "selfie" video, uploaded while I waited for the hands-on session to begin. My apologies to fellow attendees for killing bandwidth for a few minutes!
John and Yoko. Give Peace A Chance.
- SAP Managed Tags:
1 Comment
