Hi Martin, Thanks for your answer. I'll take a look at the filesystem implementation from Hadoop.
Also, I've found jets3t [1] which seem to simplify development with Java on top of S3. We'll probably use it to build a less-featured alternative to Jackrabbit for now and will think about implementing a PM with it later on. Regards, Jorge [1] https://jets3t.dev.java.net/ On 7/2/07, Martin Perez <[EMAIL PROTECTED]> wrote:
I don't think anyone is currently looking at this. It was only a random idea that I dropped at the forum. However, seeing that another Apache project has already a S3 filesystem built in perhaps it would be interesting to look at it ( http://wiki.apache.org/lucene-hadoop/AmazonS3) Cheers, Martin On 7/2/07, Jorge Ferrer <[EMAIL PROTECTED]> wrote: > > Hi Jukka, Martin, > > Has there been any progress on this? I'm also interested in this and > would like to help. > > Related to Jerome's suggestions, do you think that the current PM > architecture perform well if using a FileSystem PM to a directory that > is actually a S3 FileSystem [1]? > > Regards, > Jorge > > [1] > http://developer.amazonwebservices.com/connect/thread.jspa?threadID=10271&tstart=30 > > On 6/9/07, Jerome B <[EMAIL PROTECTED]> wrote: > > > > > > I think there are projects which access S3 as a WebDAV store, as well > as S3 > > "filesystems". If there was a S3 Jackrabbit persistence manager being > > developed, I'd be interested as well. > > > > -- jerome > > > > > > Martin Perez wrote: > > > > > > Really glad to read that. Massive distributed storage propositions are > > > being > > > proven like very good solutions for massive scaling. I guess the key > > > challenge in this case is trying to minimize bandwidth usage as much > as > > > possible in addition to the traditional storage challenge. > > > > > > Would be really nice to test it when that global data store stuff you > are > > > working in is done. If at that time I have some web projec using > > > Jackrabbit > > > (hopefully) I would love to spend some time playing with that. > However, It > > > would be even better if someone already has experiences or is already > > > trying > > > to do it and shares his experience. > > > > > > Cheers, > > > Martin > > > > > > On 6/5/07, Jukka Zitting <[EMAIL PROTECTED]> wrote: > > >> > > >> Hi, > > >> > > >> On 6/4/07, Martin Perez <[EMAIL PROTECTED]> wrote: > > >> > Is there anyone developing something like a Amazon S3 persistence > > >> manager? > > >> > > > >> > If not, it seems to me like a good candidate for a new project ( it > is > > >> too > > >> > late for SOC ), or perhaps it is an insane idea? > > >> > > >> I haven't heard of anyone actively working on something like that, > but > > >> I agree that it would be a cool idea. > > >> > > >> Actually S3 was floating around in my mind when I wrote the NGP > > >> proposal (http://jackrabbit.apache.org/dev/ngp.html). S3 or another > > >> similar massively distributed storage mechanism would work very well > > >> with the NGP concepts. > > >> > > >> I'm currently working on a related more short-term idea of adding a > > >> global "data store" that would handle all binary properties (and > > >> possibly other data like item bundles as well). The data store > concept > > >> is designed to work very well with remote or distributed storage, so > > >> S3 would be an interesting option there as well. > > >> > > >> BR, > > >> > > >> Jukka Zitting > > >> > > > > > > > > > > -- > > View this message in context: > http://www.nabble.com/S3-persistence-manager-tf3864934.html#a11036005 > > Sent from the Jackrabbit - Dev mailing list archive at Nabble.com. > > > > > > > -- > Jorge Ferrer Zarzuela > http://www.jorgeferrer.com > http://www.liferay.com >
-- Jorge Ferrer Zarzuela http://www.jorgeferrer.com http://www.liferay.com
