YODA version 1.1 are out

Version 1.1 are considered as a "base" version. This means
I'd implemented most of the planned YODA functionality at this
point. Respective user will get truly distributed objectstore,
where the single objectstore can live on multiple instances.
Each instance can replicate there objects to multiple peers.
All updates are reflected in METADIRECTORY. The newly implemented
"YODA interface appication" provides you transparent access to the
information stored in distributed objectstore without knoweledge
which physical repository holds it. If the one of repository failed
and object can be found someplace else (in the replica repository),
YODA will find and extract this information for you "at no extra cost".
The same rule applies to the all update operations, you just set the
value of the attribute and YODA will take care about the rest.
Actually, it it a pretty heavy job going on in the background. YODA
interface detects which repositories are alive and which is not by sending
event broadcast, than it's propagate updates on the live servers.
YODA interface provides two ways to access to the objectstore:
- native YODA interface (which means Pyro: http://pyro.sourceforge.net\)
- XMLRPC. All YODA objectstore operations such as get, set, drop_attr, drop
add_class, drop_class are implemented for XMLRPC. This valuable addition
will allow users to create YODA clients using almost any programming
language (Java, C++, C, Perl and so on. Of course, you can create XMLRPC
clients using Python). This is open door to using YODA in non-pythonic
environments.
As minor feature, YODA now heavily depends on Pyro event service. It uses it
for status broadcasting and for sending notification to the servers in the
YODA network about node status.