Another idea would be to start by using something like Phoenix+HBase. As we start supporting Phoenix/Drill integration, you'd already be in a good position.
J On Wed, Jun 5, 2013 at 8:14 AM, Jacques Nadeau <[email protected]> wrote: > On Tue, Jun 4, 2013 at 1:04 PM, Alexandre BECHE > <[email protected]> wrote: >> Hi Everybody, >> >> I would like to use drill to interact with a large number of monitoring >> data (ie: order of 5 millions events a day). To do it, I would like to >> store this RAW events inside HDFS (json format) or HBase (I don't have any >> preference right now,I will try both anyway) and retrieve this statistics >> (json format) using drill. >> >> Moreover, by browsing the web, I found a nice web frontend for drill ( >> http://srvgal85.deri.ie/apache-drill/) and I would love to retrieve my data >> in such a way. >> >> First, I would like to know whether or not my goal is realistic and if I >> can >> achieve it in the current state of the project? (if not, what can I achieve >> today? and do you have any approximative timescale for the rest?) > > We're approaching the point where you could do much of this utilizing > the reference interpreter. However, it was designed as a reference > implementation rather than a performant one. We're probably a few > months away from you being able to do this in earnest with the fully > distributed execution engine. > >> >> If some developement is needed from me, I would love to help but I would >> first understand how much effort it represents. > > There is a decent amount of work to be done but a lot of people are > working on it. We need both code and non-code help. An initial help > would be writing up a set of requirements for your application in more > detail. Then we can discuss how/if we'll achieve them with Drill and > what steps will need to be taken to accomplish each. > >> >> Thanks in advance for your help, >> Cheers, >> Alex
