[extropy-chat] and the nano/holo fun goes on...

Dan Clemmensen dgc at cox.net
Thu Aug 19 02:03:00 UTC 2004

duggerj1 at charter.net wrote:

>Dan Clemmensen wrote:
>>[Claimed a current cost of $1000/TB]
>Well, not quite so cheap as that. Having just bought 1000 GB of disk space in the last three months, I confidently state it costs a little more than $1000. Dan's estimate might prove right for a stationary set-up. 
>As portable computers turn into wearables, and as cell phones do the same, we'll see demand for ever-larger portable storage. In particular, I suspect as medicine improves medical monitoring will become biotelemetry. This might increase demand for very reliable, compact and high-density data storage. If black-boxes serve well in vehicles, could they serve as well in biological systems? People, children, valuable animals, crops and their symbiotes, all might need data storage like this. As it grows cheaper, more applications turn economical. 
Portable storage is "temporary." You eventually get back to your home 
base, or to a place from which you can communicate with your base for 
"free." therefore, the cost for massive storage should be at a fixed 
location, not a portable location. I built a 1TB store at my house last 
year for <$1000. To evaluate the requirement for portable storage, you 
need to compute the average data rate for data accumulation times the 
amount of time you will be cut off from a "free" connection to your 
permanent store. note that the examples you cite have trivial data 
rates, <<10Kbps average, and therefore have inconsequential storage 
costs. The costs of the battery overwhelm the cost of a Flash ram at 
this rate. (10Kbps=36Mbps/hr=108MB/day, or <$20.)

>>The only way I know to use this much disk at home is to store DVDs. 
>>Storing audio
>>CDs hardly counts: my entire CD collection fits on <<300GB with lossless 
>You can see other ideas for this kind of storage space in Microsoft's MyLifeBits project. <http://research.microsoft.com/barc/mediapresence/MyLifeBits.aspx>
>Video swallows disk space, and with wearable cameras one can very well have a personal Wayback Machine for one's own life. Today it is very easy to have a personal Wayback Machine for the Internet--just use wget as your default web browser.
>Unfortunately, the real problem with this volume of data storage is the user interface. Already a professional society of organizers exists for possessions. <http://www.napo.net/> Any bets on how long before these branch out into organizing your digitized life, into ordering one's N-Bytes of data? I'll go even odds on 27 months, +/- 9 months.
>So how do you eventually handle this? Abandon the hierachical file system to mere googling? Hire a butler equal parts remembrance agent, OpenCyc, and ALICE to explain where you left what and why? Or do you build a memory palace in a gibsonian-style cyberspace to remind one's self of everything
This is all very interesting, but the organization of the data has at 
most a minor impact on the size of the data, except to the extent that a 
sophisticated system can reduce the amount of storage.

You need to run the numbers again. A 1Mbps continuous video is likely to 
be more than enough, after compression. 1Mbps is 3.6Gb/hr, or 
3.6Gb*24hr*1B/8b=10.8GB per day. If you can get home once a day (where 
"home" == "free" access to home) then you only need 10.8 GB of portable 
store. Portable store is currently a LOT more expensive than 
non-portable, but that is not what we were discussing: your portable 
store is a temporary buffer.

Continue to do the math: 10GB/day is 3650GB/yr, or 3.65TB/yr. That's 
$3650/yr at today's prices, but the prices is dropping at 50%/yr. (Yes, 
counterintuitivly, it's faster than Moore's Law.) That says that by the 
time you can get around to implementing a fully-documented life, the 
lifetime cost will be <$3,000. That's lifetime cost, not yearly cost.
Of course, as your life becomes richer (due to augmented intelligence) 
you will need to increase the bit rate required to capture it, unless of 
course your data compression capabilities increase in proportion.

Just for the heck of it, let's consider the cost of the portable store.: 
10.8GB. Let's go for the expensive approach: flash. Looking at 
www.pricewatch.com, we see that flash (in the form of USB flashs drives) 
costs $100/GB, retail, quantity 1. So we can store 10.8GB for $1100, 
plus the rest of he cost of your wearable, and we can confidently expect 
this cost to come down by half every 18 months or better. Size of the 
buffer goes up with the acquisition bit rate, but it goes down with the 
"cost" of connectivity with home. A reasonable first approximation would 
be that the size will less than double each year and the cost per GB 
will be less than half/GB each year, so the yearly cost of temporary 
(protable, buffer) storage will decrease each year.

Conclusion: for leading-edge geeks, a continuous life-record is already 
easily in economic reach. The current cost is primarily the system 
development cost, not the HW cost.

>>You can store about 5 audio CDs with lossless compression (flac) in one 
>>GB, for a storage
>>cost of $.20/CD. Storage cost for a DVD would be about $4.00, which is a 
>>small fraction
>Double these numbers for a back-up copy of valuable data. :)
I did :-)
I'm keeping two live copies of my CD collection, one active, one backup. 
Ripping about 1000 CD's took a lot of time, and my time is valuable. It 
would have been cheaper to store the result onto DVD-RW, except that the 
time to do so was more expensive than doubling the amount of hard disk.

Note that "double" is really a brute-force approach, based on laziness. 
I'm using a brute-force backup scheme. If I used RAID-5, I could get the 
same level of backup at a 25% increment instead of a 100% increment.

More information about the extropy-chat mailing list