Re: [PAPER] Juggling with packets: floating data storage

[Nicholas Weaver <nweaver () CS berkeley edu>]

On Wed, Oct 08, 2003 at 08:18:12PM +0200, Michal Zalewski composed:
> > The higher level juggling is still pointless.  Lets assume I still have
> > the 1 Gb link to the outside world, and the data round trip latency is a
> > minute (the amount of time data can be stored externally before it comes
> > back to me).  Thats still just 60 Gb of data, or 7.5 GB.  And I'm having
> > to burn a 1 Gb link to do it!
>
> Should you actually read the paper, it would be easier to comprehend the
> idea. The paper proposes several approaches that do not require you to
> send the data back and forth all the time; in some cases, you can achieve
> near-infinite latency with a very low "sustaining traffic" necessary to
> prevent the data from being discarded. Besides, the storage size is not
> the main point, the entire approach is much more interesting from the
> limited deniability high-privacy storage point of view.

You have some refresh time required, just like DRAM or any other
lossy/decaying medium.  Without the refresh, you will have
uncorrectible data decay.  Its simply the refresh bandwidth and
refresh rate.

So lets assume a 1 DAY refresh time, and a 1 Gb refresh bandwidth.

That's still a maximum of ~10 TB of storage (3600 * 24 / 8), and you
are going to have to be saturating the link to maintain it.

With a 100 Mb refresh bandwidth, thats 1 TB.  I can go out and buy a
box which holds nearly 1 TB for <$2000.  Heck, I can buy a >2 TB RAID 
array, FROM APPLE (hardly a low price vendor) for $11k!


In terms of serving data, CDN's reportedly charge around $20/GB served
or so, or you can always construct a BitTorrent-like network to use
the edge-hosts, which is what Valve wants to do for selling/patching
HalfLife.



Even given a low/moderate-user (~100 user) network, the data cost is
going to be pretty extreem to maintain refresh, and as a result,
rather/very unstealthy.  Lets assume 1 TB shared between 1k users, and
a 1 hour refresh time (probably still way high).  

Thats still 200 Kb/s/user continuous bandwidth.  I don't know about
you, but I can't get that reliably on my cable modem upload, and if I
DID use that much bandwidth, ALL THE TIME, the cable company would
probably come knocking.

Why not just build a distributed filesystem for those 1K users?



The only real advantage is a monocrum of stealth, but it isn't really
stealthy if you are storing a non-trivial amount of data: the network
traffic is "Storage volume / refresh time", and the refresh time is
going to have to be fairly short to have a monocrum of reliability.
Burning even just 10Kb/s/user isn't going to be "stealthy" in
practice, due to the continual load.


For storing a small quantity of data (eg, ~1-2 GB or less), there are
much better covert repositories one could consider.

> The second observation is that not all the data has to be send back and
> forth all the time. Interesting.

The refresh rate is required for parasitic storages, to prevent data
decay and to recompute checksums etc to handle data loss.

Even given a 1 WEEK refresh cycle, a 100 Mb continual refresh
bandwidth only gets you 7 TB of storage, 1 Gb aggregate bandwidth gets
70 TB.

-- 
Nicholas C. Weaver                                 nweaver () cs berkeley edu

_______________________________________________
Full-Disclosure - We believe in it.
Charter: http://lists.netsys.com/full-disclosure-charter.html