h a l f b a k e r yThink of it as a spell checker that insults you, as well.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Back in the 19somethings, geomagnetic surveys provided the first
strong evidence for continental drift. The mid-oceanic sea-floor
was
found to have alternating stripes of north/south magnetisation,
running parallel to deep-sea trenches.
These alternating stripes were caused by periodic reversals
in the
Earth's magnetic field, combined with the continual emergence of
new volcanic rock spreading from the deep-see trenches. As new
rock emerged, its iron-rich minerals took on the prevailing field
direction before
solidifying and the continuing to migrate outwards, with its
magnetic field locked in place.
Maps of these patterns are beautiful, showing a series of
alternating
stripes (north/south polarity). I imagine that a close-up
magnetometric view of a recording tape would look roughly
similar.
But I digress.
We've now mapped out this field pattern pretty well, and all the
data is stored away somewhere safe and sound. We no longer
need
the magnetic archive of the rocks themselves.
So, MaxCo Data and Livestock, Inc., is launching a scheme to re-
use
the magnetic storage capabilities of these deep-sea rocks. We
have
developed a rover capable of travelling slowly across the vast
pointless expanses of the ocean floors. This rover (the MaxMag) is
capable of generating magnetic pulses sufficiently strong to re-
polarize the iron-containing elements even in long-solidified sea-
floor rocks.
Due to the need for the field to penetrate through several metres
of
overlying silt before reaching reasonably solid and stable rock, we
estimate that patches of ocean floor can be re-magnetized with a
chosen polarity on a resolution of about 10 metres. This means
that
the Pacific sea-floor will have a storage capacity of about 1.5
Terabits, whilst the smaller Atlantic sea-floor can be used for
checksums.
Although this storage capacity is not huge compared to modern
hard
drives, it is very secure against fire and flood. Data written in this
way should endure for millions, if not billions of years. Companies
and governments needing the ultimate in secure, disaster-proof
data-storage are queueing up to avail themselves of this facility.
For an additional charge, MaxCo will perform regular data-
checking
and servicing on a millennial cycle. Data which is approaching
subduction zones will be read, and re-encoded in the freshly-
produced sea-floor rocks alongside mid-ocean trenches.
For short-term storage, special rates are available for regions
around
Hawaii, Iceland and so forth, where data can be guaranteed stable
only for a few millennia.
Data read-out is, of course, a concern. Current read-out rates are
approximately 0.01 bits per second, using a sensitive
magnetometer
mounted on the MaxMag rover itself. However, we are now
developing a fleet of eight nimble reading-only-rovers, and expect
to
attain readout rates of over 2 bytes per minute in the near future.
However, access times will be dependent on the location of your
data, and on seasonal factors.
Magnetic patterns in the seafloor
http://pubs.usgs.gov/of/1999/ofr-99-0132/ [MaxwellBuchanan, Feb 27 2011]
Reed-Solomon coding
http://www.cs.cmu.e..._solomon_codes.html Use this for error correction and redundancy. [neelandan, Mar 04 2011]
STOP PRESS: Joint Venture Agreement speeds data access!
Kiritimaticentrifugomobile [MaxwellBuchanan, Mar 04 2011]
Please log in.
If you're not logged in,
you can see what this page
looks like, but you will
not be able to add anything.
Annotation:
|
|
you failed to mention the write rate. Pray tell what sort of speed we can anticipate. Also a little curious about the track width re: the continuing warping of the media itself will effect the life of the data. My money is still in scribbling messages on the moon. |
|
|
But the Earth's crust is recycled and melted down through tectonic plate activity ever few hundred million years. To avoid data loss you'll need backup storage on another plant - essentially a
trans-galactic RAID array. |
|
|
The moon, with no tectonic plate activity, should be suitable for a backup. There is scope for a low-res magnetic record and a higher res optical one. |
|
|
You could have solar powered rovers roaming the surface, writing the data in. |
|
|
Readout could be by lunar satellites - or by telescope from earth. |
|
|
[neelandan] Yes, but you're forgetting that the sun will grow to a red giant star in a few billion years' time, engulfing the Earth and the Moon. You need the cosmic equivalent of an off-site backup. |
|
|
So, how to interfere in the first few microseconds of universe formation, to modify the patterns of cosmic background radiation? |
|
|
[WcW] Write rate is currently 1bpm, or slightly faster than
the read rate. Others see this as a limitation. We prefer
to
think of it as a means to ensure that available capacity is
not
filled too quickly. |
|
|
[hippo] see para.9 (it's between paras 8 and 11, and closer
to
the former) |
|
|
[vernon] reactionless drive is still not going to work. |
|
|
[neelandan] you're not looking at this from a practicality
point of view, are you? |
|
|
It would be a glorious coincidence if the natural
magnetic data was 42 in binary. At some point it
must have been 0101010. |
|
|
I think that, by definition, an alternating series of 1s and 0s
must be 0101010 quite frequently. |
|
|
//an error correction scheme// |
|
|
See the last bit of para 7. It is in between the middle part of
para 7, and the start of para 8. |
|
|
You sure? I mean on error correction? A simple one-bit
checksum at least lets you identify an error; a two-but
checksum would allow pretty good error correction, n'est ce
pas? |
|
|
Well, then we're in business. I figure the Atlantic is good for
four bits for every Pacific byte. |
|
|
Any thoughts to the location of the file system? Mediterranean perhaps? Also, I am not sure if you want any Thomas, Richard or Harold reading this data off the sea bed with their own read head, so pehaps some encryption. At last a possible use for the Yucca Mountain OTP, stored safely as faults in the crystaline structure of a plasma deposition created diamond. |
|
|
You might use some of the block convolution codes for the purpose of error correction and redundancy. Sort of that used for recording on CD and DVD's. |
|
|
Let's consider the head for a little: |
|
|
Since GPS doesn't work down there, the head will perhaps
need to be zeroed (parked?) from time to time.
It may need parking if you should want to carry the data
somewhere, and there should be a little diagram on it to
explain the effect of smoke particles should they get
between the head and the platter. |
|
|
Let's say the MTBF of the head is 100,000hrs. That's a new
head every 12MB or so. |
|
|
[Ling] you raise a valid point, but our engineers have already
begun to address this problem. We have got in contact with
the developers of the Kiribaticentrifugomobile <link>, with
which this idea has an almost preternatural affinity. I'm sure
the details are self-inexplicatory. |
|
| |