On Jan 29, 2013, at 7:10 AM, Michael Hausenblas <[email protected]> wrote:
> >> Like it. > > Thanks. Glad you find it useful. > > >> You should probably add jdbc and odbc as input and output interfaces > > > Agreed, direct RDBMS access should be there. Would a further data source at > the bottom capture that or do you mean (also) in the top right corner > (extending REST, etc.)? Unsure how to express this best ... > Both. > Cheers, > Michael > > -- > Michael Hausenblas > Ireland, Europe > http://mhausenblas.info/ > > On 29 Jan 2013, at 15:01, Jacques Nadeau <[email protected]> wrote: > >> Like it. You should probably add jdbc and odbc as input and output >> interfaces >> >> >> >> On Jan 29, 2013, at 5:06 AM, Michael Hausenblas >> <[email protected]> wrote: >> >>> All, >>> >>> Motivated by the fact that I find myself explaining Drill more and more >>> often lately I thought it might be useful to provide a high-level view on >>> the architecture of Drill. So, I gave it a shot and you can find the result >>> at [1] - note that also the OmniGraffle source is attached to this page in >>> case someone wants to tweak it. >>> >>> This high-level architecture diagram was created with the following in mind: >>> >>> * Providing a quick overview of the features, interfaces and ecosystem ... >>> * ... while aiming not *only* at a technical audience >>> * Summarising the various discussions around desired data sources and query >>> languages supported >>> >>> One aspect I find very important and that I tried to highlight is the human >>> in the loop: being an interactive system, this should clearly be in the >>> centre of attention? >>> >>> Cheers, >>> Michael >>> >>> >>> [1] >>> https://cwiki.apache.org/confluence/display/DRILL/High-level+Architecture >>> >>> -- >>> Michael Hausenblas >>> Ireland, Europe >>> http://mhausenblas.info/ >
