DOAG Database 2015

By Franck Pachot

.

It was my first time at DOAG Datenbank in Dusseldorf. 

My understanding of German is very limited. I’m French, I’ve learned German at school but then didn’t practice. Now that I’m living in Switzerland it would help but when I tried to remember something from school time it failed with Ora-1555: Snapshot too old…

So I was there because I presented SQL Plan Directives – in English – and I really appreciate that DOAG has accepted a session in English in a German event. It seems that it was ok for the attendees as well: lot of people, interaction… language is not a barrier, good that we can overcome the Babel thing. In IT we have a lot of non-standards (keyboards, endianess, charactersets, CR+LF,…) but at least we all speak and understand the same language – a basic subset of English. Basic is not pejorative here. With less words we have to focus on the main points without litterature, and that makes good technical presentations.

That DOAG event is like the other DOAG event I know – the DOAG conference in November: very well organized in a very nice place. Near the airport makes it accessible. And the same thing that I really appreciate at the DOAG conference: rooms are like a circle around the place where people meet, talk, eat, etc. I don’t like those events where rooms are at the end of several corridors. The DOAG events, besides the quality of the sessions, are a greet place to meet. I’ve met in ‘real’ people I appreciate though their blogs or tweets. I’ve met people that have read my articles (so it seems that the German transation of my Docker article was good). I’ve met people that have migrated to 12c (and they encountered SQL Plan Directives, for the good and the less good consequences). People don’t hesitate to give you direct feedback about the good and bad points of your session and I like that. We are there to share and the places chosen by DOAG helps that.

I attended the sessions where I can understand something: slides in English, and demos, helps that. Here is quick shot at those I appreciated the most:

Martin Klier on In-Memory column store: I like when presentations go into the details. For example, everybody talks about vector processing (SIMD) when processing a column store, but rarely give an idea about the size of that vector. Martin did: current CPU can have 256 bits register.

Martin’s core message is: InMemory ist cool. Cool reicht nicht. UseCase muß passen.

Lothar Flatz has his own style which makes him a great speaker, presenting scientific approach to performance as if telling a story. And in an interactive way which forces us to think about the problem. Very interresting ‘middle ground in the buffer cache’. Looking at segments statistics and comparing the ratio between physical and logical reads to see who is using the cache efficiently or not.

Lothar’s first message is: Before we start tuning, let us tidy up the buffer cache.

So, thanks again DOAG and I hope to see everybody in Nürnberg on November.

Tweets about doagdb15