Hi,
I'm in the starting phase of my master thesis in computer science and I
think this would be a good opportunity to contributing to the open
source community. My supervisor at the universitysuggested Tahoe-LAFS
and from what I have read so far this is one interesting project.
The scalability problems caught my interest, in particular these
tickets: #68, #235, #444. I know that #68 has been solved by adding the
possibility to add several introducers to a grid. I want to make a
system that is independent from external services like the introducer. A
node should be able to join with the assistance of an arbitrary node
within the grid.
However, the basic idea is to remove the dependency of full-mesh and
implement some kind of a DHT solution.
I'm considering to support some kind of locality in the solution. This
could have negatively impact on the uniform distribution but in large
global grids locality is a desirable property for achieving good
performance.
The main goal is to increase my knowledge in this area and if the
implemented solution is successful it will be contributed.
Regards,
André
_______________________________________________
tahoe-dev mailing list
[email protected]
http://tahoe-lafs.org/cgi-bin/mailman/listinfo/tahoe-dev
