Hi, thank you both for the links and advices. Looking at the work in progress concerning security, it seems that currently the only way to have authentication is to implement it at the Web Server level. In my case, my web server is deployed under Tomcat, therefore I think I should manage clients authentication with Tomcat, and then, in case, forward the obtained credentials in some way from my service downstream to hadoop, when I could define service level authorization and/or users permissions. Is that right? Could you also suggest me some other useful links, maybe with examples if any, on how to implement such a mechanism?
Thank you again, Giovanni 2009/7/22 Andrey Pankov <apan...@iponweb.net> > Maybe this would be interesting for you, > > > http://www.hadoop.iponweb.net/Home/hdfs-over-webdav/authentication-and-permissions > > On Wed, Jul 22, 2009 at 19:00, Giovanni Tusa<giovan...@gmail.com> wrote: > > Hi all, > > > > on my company we recently began some projects where we would like to > > actively use hadoop. > > For one of them, basically we have implemented a feed service where feed > > entries (there could be a lot of data because many different users can > post > > feeds), are actually stored in the server filesystem (for the moment only > > one server) by using Hadoop.S > > > > Such a system needs some security improvements, and I'm wondering what > kind > > of security mechanism is currently provided by Hadoop. > > For example, only allowed users should be able to access, from the > browser, > > to our service to read data (because they will normally be sensitive > data), > > provided that they have access credentials (authenticated and authorized > > users). Also, the same data should be secured by using some forms of > > encryption for the server-to-client data transfer. > > I read some old posts and discussions inside the hadoop project > concerning > > security, but I am a little confused on what is actually implemented > > concerning authentication, authorization and encryption, on which version > of > > hadoop, and if something currently provided can match my needs. > > > > If someone could point me to the right direction, your help will be very > > appreciated. > > > > Giovanni > > > > > > -- > Andrey Pankov >
