Article is meh. Most storage technologists have known for years that putting a bunch of hard disks in the back of a truck and driving it somewhere is faster than transferring over a wire (for large amounts of data).
> The second thing that changed my mind was a new law in the UK that makes it illegal to not hand over encryption keys if the police want to decrypt your data.
Except that the law is not new, and was passed in 2000. The law linked from the page linked in the above quote is entitled "Regulation of Investigatory Powers Act 2000." The author plainly didn't even look at that page, however ...
Being pedantic here, transmitting one byte (8 bits) of data requires 10 "bit times" (called bauds) because there is a start bit and a stop bit. This is sometimes called 8b/10b encoding since you get 8 bits delivered out of 10 bits sent. In packet systems (like ethernet) it gets a bit weirder since you might have 1500 bytes of data (represented by 15,000 bauds) but you add 36 bytes of network header information (another 360 bauds)) 1500 bytes is now 15,360 bauds [1].
The biggest challenge of shipping drives around though is that hard drives are really really cranky about g-shock so you can easily destroy all of your data if you drop, or even suddenly bang your drive as you are moving it around.
I think it was mostly interesting from the perspective of someone who isn't used to dealing with large data sets suddenly having that 'ah ha' moment when confronted with the challenges of managing data with consequences.
[1] This is nominally incorrect since the protocol layer overhead isn't 'bauds' like the physical access layer which uses the term strictly for signalling structures, but it works well for this example.
What we need is a dual-key encryption system. One key keeps your real data encrypted. The other key "decrypts" the drive to a benign (mostly blank) disk. If the cops ask, hand over the alternate key and stand back, bemused.
Yes, and this already exists. It's called "plausible deniability". Truecrypt has it. You can use one password and unlock the harmless data. And nobody can prove that there is another hidden, evil, volume in there, which can only be unlocked with another password.
The brain excerpt is very odd. One billion neurons? It's more like 100 billion. One trillion synapses? It's more like 400 trillion to 1,000 trillion. That's one trillion to three trillion synapses per cubic centimeter of brain on average (obviously areas have various synaptic density).
It appears the cited article is only talking about "memory neurons" and related synapses, but that's extremely unstable ground.
12 comments
[ 4.4 ms ] story [ 35.7 ms ] threadExcept that the law is not new, and was passed in 2000. The law linked from the page linked in the above quote is entitled "Regulation of Investigatory Powers Act 2000." The author plainly didn't even look at that page, however ...
A 2TB drive contains 2,199,023,255,552 bytes, or 17,592,186,044,416 bits.
17,592,186,044,416 / (60sec x 60min x 6hrs) = 814,453,058 bits/sec, or roughly 814Mbps.
http://www.google.com/search?sugexp=chrome,mod=8&sourcei...
The biggest challenge of shipping drives around though is that hard drives are really really cranky about g-shock so you can easily destroy all of your data if you drop, or even suddenly bang your drive as you are moving it around.
I think it was mostly interesting from the perspective of someone who isn't used to dealing with large data sets suddenly having that 'ah ha' moment when confronted with the challenges of managing data with consequences.
[1] This is nominally incorrect since the protocol layer overhead isn't 'bauds' like the physical access layer which uses the term strictly for signalling structures, but it works well for this example.
That's the beauty of math. Sorry police!
It appears the cited article is only talking about "memory neurons" and related synapses, but that's extremely unstable ground.