Re: [Sks-devel] IPv6 peering; keydumps annoyingly large

[Jason Harris]

On Thu, Jun 02, 2011 at 12:26:41AM +0100, Xian Stannard wrote:

> If I could host only part of the entire key collection (or the
> collection was smaller) I would. My guess is there are other people in a
> similar situation. Many many more servers hosting only part of the
> collection could still achieve a higher redundancy than we currently
> have. If an admin could dictate the proportion of the entire collection,
> or just specify the minimum amount of disk to keep free, life would be
> easier. For me anyhow.

Then look into distributed/NoSQL databases, e.g., MongoDB.

On Thu, Jun 02, 2011 at 03:14:17PM +1000, Matthew Palmer wrote:
> On Wed, Jun 01, 2011 at 06:14:34PM +0100, Xian Stannard wrote:
> [redundancy through partitioning]
> > Servers could carry multiple subsets to make sure that no particular
> > subset lacked in redundancy?
> 
> But who coordinates which servers carry which subset(s)?  Without that, I

AIUI, MongoDB does all that partitioning under the hood.  You just need
to make sure you don't lose too many nodes.  You do have to trust the
master/node(s) updating/merging the key material not to remove any, and
the node(s) storing data not to tamper with it, of course.  A good
keyserver design would add digital signatures from the master(s) and
be able to drop malicious storage nodes.

On Wed, Jun 01, 2011 at 03:44:08PM -0400, Jeff Johnson wrote:

> 4Gb is large for a data store? Get a grip please, peta-byte databases
> are almost routine these days and no amount of growth in the SKS key
> stores is going to suddenly explode exponentially.

From a growth in actual users - we wish.  From a malicious adversary,
either script kiddies or a virus-driven DDoS via botnet - totally easy
to imagine:  continually generate keys, send them to any SKS keyserver,
watch the network fill up/try to keep up/whatever.

> "sneaker net" transport any time soon. There's only, what 200?
> 500? people in the world who care (and know) what a SKS dump is.

Likely less, and most of them are hopefully reading this list.

On Wed, Jun 01, 2011 at 03:10:01PM -0400, Phil Pennock wrote:

> Note that we've already lost one valued keyserver operator in Germany
> because he was unable to comply with a privacy request from a user to
> delete their key and he, quite reasonably, did not want to be sued, so
> shut down.

Unfortunately, I don't think any of us knew enough OCaml to help in time...

-- 
Jason Harris           |  PGP:  This _is_ PGP-signed, isn't it?
address@hidden _|_ Got photons? (TM), (C) 2004

pgpMXPy6GnzTS.pgp
Description: PGP signature