[volunteergrid2-l] Hello!

[Shawn Willden]

On Sat, Mar 31, 2012 at 11:49 PM, Benjamin <masterme120 at gmail.com> wrote:

> Thanks Shawn!
>
> My upstream bandwidth is only around 0.3 Mbps.  Would that be a problem?


Wow, that is slow.

That is going to pose two problems, one for you (which isn't a problem for
everyone else) and one for everyone else.

The problem for you is that your backups are going to be slow.  You would
definitely want to use a helper to mitigate that, but it'll be slow
regardless if you have very much to back up.  (Without the helper, your
node would do the Reed-Solomon coding and then upload the shares, so the
redundancy added by the error correction would all have to be uploaded over
your little pipe.  With a helper, you'd just upload your encrypted data to
the helper, and it would do the coding and distribution for you).  The big
problem with slow uploads is that the way "tahoe backup" works, it doesn't
created the dirnodes which make all of your files accessible until the
backup completes.  And without those dirnodes your data can get
garbage-collected.  So it's possible to get into a situation where you can
never finish your backup before Tahoe deletes your data.

There is a workaround to that problem, but it makes things a little more
painful.

The problem for us is that our restores are going to be slow if we get into
a situation where we really need the share you're holding.  Our backups
should be reasonably fast based on your downstream rate, but our restores
will be hampered by slow shares from your box -- but only if we actually
need the share from your box.

I think it might be okay, but I would want to reserve judgment and wait to
see how it actually performs in practice.

-- 
Shawn
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://tahoe-lafs.org/cgi-bin/mailman/private/volunteergrid2-l/attachments/20120401/1ef4cd9f/attachment.html>