[tahoe-dev] GSoC

[Zooko O'Whielacronx]

Hello Lalit Bharat:

Thank you for your interest in contributing to Tahoe-LAFS in the
Google Summer of Code!

You mentioned two ideas in your initial message. Both of these could
potentially be good Google Summer of Code projects. Let's talk about
each of them and then you go to the GSoC web site
(http://socghop.appspot.com/ ) and submit a Proposal for the one that
you like most. Remember that you can always update your project
Proposal "in place" on the GSoC web site after you have submitted your
Proposal, so the best strategy is to submit it as soon as you know
which project you are going to work on, and then iteratively improve
it in-place. That way you can't miss the deadline.

(The Google Summer of Code organizers are utterly inflexible about
deadlines. If your house burns down the day before the deadline and
you write to them saying can you please have an extra 24 hours to dig
your proposal out of the ashes and submit it they will write back and
say No -- you shouldn't have waited until a day before the deadline in
the first place. :-))

In fact, you are allowed to even submit more than one project Proposal
if you have finished polishing up the first one and then you start
getting increasingly interested in another project. :-) Of course
Google will not sponsor a student to work on more than one project in
a summer, so it is useful to propose more than one project only in
case something prevents the first project from happening.

By the way there is another student also interested in the RAIC
project. For now I guess both of you should continue to expore both
RAIC and alternate proposals.

On Wed, Mar 31, 2010 at 3:32 AM, Lalit Bharat
<lalit.bharat.apm06 at itbhu.ac.in> wrote:
>    1. The first one is regarding the idea "Redundant Array of Independent
> Clouds"

This is a good project. It comes in a few discrete subprojects so that
you can tell how you are making progress. The first subproject is to
abstract out the interface between the storage server logic and the
backend, as you mentioned, and as is described in #999.

The next subproject is to make an Amazon S3 backend.

Once you've done those two then the project can already be considered
a success. (Of course, it has to be done in the Tahoe-LAFS development
style which means thorough tests, docs, code-review, and backward- and
forward- compatibility analysis.)

However, hopefully you will already finish the first two subprojects
early in the summer and then you can do the "bonus" subprojecs: make a
Rackspace Cloud Files backend and then a Google Docs ("GDrive Does Not
Exist") backend.

Frankly it isn't until you get more than one alternative backend that
the project begins to be really interesting to me, because once you
have several independent clouds then you get the phenomenon of your
availability and reliability being as good as the best availability
and reliability of the several cloud services that you use [1].

Okay, so the next step on that project is to submit a Proposal on the
GSoC web site saying that you want to work on RAIC and describing the
project. This doesn't mean that you can't actually switch to a
different project if you prefer a different project and another
student prefers RAIC.

I will write a separate letter about your other idea.

Regards,

Zooko

http://tahoe-lafs.org/trac/tahoe-lafs/ticket/666# Accounting: limit
storage space used by different parties

[1] http://allmydata.org/~zooko/RAIC.png