On Thu, Jul 23, 2009 at 6:46 PM, Meredith Gregory <[email protected]>wrote:
> Viktor, > > Yes. For example, in the biotech case the data is coming in from a > device-based origin. This is interesting. Analyzing the AST of the query for the data, then identify the different data sources and then query each for all relevant data since last refresh, then push that into partitions and then schedule the Jobs and execute SQL and then merge/reduce the results. > > > Best wishes, > > --greg > > On Thu, Jul 23, 2009 at 2:35 AM, Viktor Klang <[email protected]>wrote: > >> >> >> On Wed, Jul 22, 2009 at 11:52 PM, Meredith Gregory < >> [email protected]> wrote: >> >>> Alex, Viktor, >>> >>> i think write semantics could get complicated quickly, actually. However, >>> i was initially responding to the idea that trad business object models >>> don't give way to analytics. Being able to make read-only queries against >>> large volumes of data using the original business object schema seems to me >>> like a win -- even if it's only used to populate a db that's sliced up in a >>> different way for further analytics processing. >> >> >> So basically, what's needed on top of HadoopDB is a service that updates >> data as needed from external data sources. >> >> >>> >>> >>> Best wishes, >>> >>> --greg >>> >>> On Wed, Jul 22, 2009 at 2:44 PM, Alex Cruise <[email protected]> wrote: >>> >>>> >>>> Viktor Klang wrote: >>>> > Absolutely, perhaps I'm tainted by write-heavy systems and perhaps I'm >>>> > just failing to see the overhead we're talking about. >>>> > Perhaps I overlooked it, but the paper didn't mention performance for >>>> > small writes and potentially multiple nodespanning transactions. >>>> HadoopDB makes no claim to any support for writes at all, I'm just >>>> speculating that It Should Be Possible given my understanding of its >>>> architecture, which is admittedly limited and based solely on reading >>>> the paper and a bit of the code. :) >>>> > I'm inclined to believe that some sort of immutable records storage >>>> > would simlify the semantics (analytic queries are IMHO very seldom >>>> > demanding real-time snapshots) >>>> Analytical queries against static data are exactly what it's for. I >>>> have no experience with its competition, namely parallel/distributed >>>> column-oriented databases, so I can't say whether they're any happier >>>> with writes. >>>> >>>> FYI I brought up HadoopDB on the NoSQL list too but so far not too many >>>> takers... >>>> >>>> -0xe1a >>>> >>>> >>>> >>> >>> >>> -- >>> L.G. Meredith >>> Managing Partner >>> Biosimilarity LLC >>> 1219 NW 83rd St >>> Seattle, WA 98117 >>> >>> +1 206.650.3740 >>> >>> http://biosimilarity.blogspot.com >>> >>> >>> >> >> >> -- >> Viktor Klang >> >> Rogue Scala-head >> >> Twttr: viktorklang >> >> >> > > > -- > L.G. Meredith > Managing Partner > Biosimilarity LLC > 1219 NW 83rd St > Seattle, WA 98117 > > +1 206.650.3740 > > http://biosimilarity.blogspot.com > > > > -- Viktor Klang Rogue Scala-head Twttr: viktorklang --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Lift" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/liftweb?hl=en -~----------~----~----~----~------~----~------~--~---
