[open-science] BioTorrents

[Rufus Pollock]

On 17 April 2010 23:28, John Wilbanks <wilbanks at creativecommons.org> wrote:
> Also, be sure to include the Tranche Project in anything on this topic. I
> think they're a long ways towards the goal.
> https://proteomecommons.org/tranche/

Thanks John, we'd already got that on the listing :)

<http://wiki.okfn.org/p/Distributed_Storage/Research#Tranche>

They seemed to have adopted a hybrid centralized/p2p model
<http://code.google.com/p/tranche/wiki/WhitePaper#Introduction>:

<quote>
Second, the repository is a hybrid model that uses concepts from both
the peer-to-peer networks and centralized servers, so it can best be
described as a "distributed server model". Because we are frequently
dealing with very large data sets that must be available at all times,
a pure peer-to-peer model would not be feasible since nodes
continuously enter and leave the network. Furthermore, a centralized
model presents other issues such as scalability, and places the entire
burden of hosting hardware on one group or organization. Our model
accommodates a reasonable amount of individual server downtime, as
well as permitting the replacement of old servers and the addition of
new servers to meet growing space requirements.
</quote>

I'd be interested to hear of any experience people have had with Tranche ...

Rufus