Fwd: Easy way for community to volunteer storage

[Sajith Sasidharan]

It looks like at some point there was interest in allowing storage nodes
behind NAT; not sure what, if anything, has happened since.

https://tahoe-lafs.org/trac/tahoe-lafs/ticket/445
https://tahoe-lafs.org/pipermail/tahoe-dev/2009-December/003331.html

In that mailing list thread, IPv6 was proposed as a solution.  I live in the
metro area of a first world country, and my ISP does not offer IPv6 out of the
box, twelve years later... looks like it remains unsolved? :^)

Regards,
Sajith.


On Sun, 2021-04-18 at 21:43 -0400, Sajith Sasidharan wrote:
> Hi Swami,
> 
> You are right, your email did not make it to the list.  I'm sorry about
> that.
> Our list is a little broken right now.  We're working on fixing it.
> 
> Thank you for not giving up on us though!  One challenge we have is not
> knowing enough about real world users or usage of Tahoe-LAFS, so I am very
> glad that you persisted.
> 
> The usage pattern you've proposed sounds like sounds like a variation of
> what
> is called a "friendnet" in Tahoe-LAFS parlance, where a group of friends
> share
> storage with each other.  Tahoe-LAFS could be appropriate for you, although
> not quite in its current shape, and not right now, just as you thought.
> 
> You have gathered some ideas from the linked Gridsync issue about
> difficulties
> of running Tahoe-LAFS storage nodes in people's home machines.  I am still
> new
> to Tahoe-LAFS, so I am not sure I have better ideas than that.  Hopefully
> more
> experienced people in the list will chime in.
> 
> Regarding point 1, Tahoe-LAFS is not super simple to install or use at the
> moment.  I suppose we should strive to improve that.  Regarding point 2, I'm
> not quite sure what kind of challenges there are when running a storage
> server
> node behind a home router -- I have not partaken in a friendnet
> myself.  About
> all your points, I'd love to heed to the community's wisdom.
> 
> To answer your question about interest in the ideas, I do think they are
> interesting.  It should be possible to create what you describe using Tahoe-
> LAFS as a building block.  I am just unclear about the amount of effort.
> 
> In any case, I suppose we should stay in touch. :)
> 
> Regards,
> Sajith.
> 
> 
> On Sat, 2021-04-17 at 23:38 +0530, Swami Kevala wrote:
> > Hi Sajith,
> > 
> > Not sure if my email to Tahoe lafs Dev list was getting through... So
> > sending to you directly. Hope you don't mind.
> > 
> > Kind Regards
> > Swami Kevala
> > 
> > ---------- Forwarded message ---------
> > From: Swami Kevala <swami.kevala at ishafoundation.org>
> > Date: Sat, 17 Apr 2021, 13:09
> > Subject: Easy way for community to volunteer storage
> > To: <tahoe-dev at tahoe-lafs.org>
> > 
> > 
> > Hello,
> > 
> > I am looking for a way to use Tahoe-LAFS to enable a large community of
> > non-
> > technical people to volunteer a fraction of the hard drive space on their
> > personal computers to collaboratively host an archive of digital content
> > in
> > perpetuity. (Sort of a cross between BitTorrent and the SETI at Home project
> > -
> > but for storage rather than processing)
> > 
> > We have an archive with about 1 PB of content, and we have a community of
> > over a million active supporters. We are looking at cost-effective and
> > resilient ways of hosting our archives on several redundant systems.
> > Sharding, encrypting and distributing it to our community seems like one
> > robust way of doing this.
> > 
> > Points to note:
> > 
> > 1. Software installation and configuration should be super-simple. One-
> > click 
> > install, which asks the user how much space he can volunteer, and which
> > local directory will be used   
> > 
> > 2. Most people will not have a static IP address, so the solution should
> > work for the average personal computer setup (dynamic IP assigned by ISP)
> > -
> > without needing any additional router configuration
> > 
> > 3. It is not a problem if many people have their computers switched off
> > frequently (e.g. during non-work hours). The primary purpose is to provide
> > a
> > durable long term archive; speed of restoring content is a secondary
> > factor. 
> > 
> > 4. The volunteers would not use the grid to share their own content. The
> > system would be strictly one-way. Our central archives would be pushed
> > onto
> > the distributed storage provided by the community. The motivation for the
> > community to participate is that they understand the archive holds content
> > which is of great value to humanity.
> > 
> > 5. The amount of storage per node could vary between say 10GB on a home
> > PC,
> > to say 1TB on a server type system 
> > 
> > For your reference, I initially posted this question on the Gridsync
> > Github
> > page, and was directed to this mailing list by the developer
> > 
> > https://github.com/gridsync/gridsync/issues/341
> > 
> > I understand that currently this may not possible as I have described.
> > Would
> > like to know if there is any interest in the ideas being proposed, and any
> > likelihood of this kind of arrangement being possible in the future.
> > 
> > Kind Regards
> > Swami Kevala
> >