h a l f b a k e r yIt might be better to just get another gerbil.
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,
|
|
|
The internet is a tricky thing to consider in its entirety. On
a basic level, you might have a GHz microwave WiFi
connection to a local access point, that's connected by
copper ethernet cable through a few switches and routers
to a fiber optic connection. At some point, this goes
through to
the fiber optic backbone that spreads through
major population centers. Because data moves through
shared connections, it travels as packets with address
information like some electronic equivalent of mail. The
internet is discontinuous, unlike the traditional telephone
system for example.
The discontinuity offers opportunities; a data packet can
move by various means with diverse speed and bandwidth
characteristics. Fiber optics are zippy, my ping time to NY
is only ~4ms, even better is internet over HF, you can span
the US at practically light speed. Sadly, reliability and
bandwidth are dreadful. Currently, high-bandwidth
internet requires costly and vulnerable fiber optics that
must be laid underground.
One way to increase bandwidth is to increase the size of
the data packet. More data for the same amount of
routing/checking information. As illustrated by the quote,
"Never underestimate the bandwidth of a station wagon
full of tapes hurtling down the highway". Now, station
wagons require drivers, oil changes and a
highway/bathroom infrastructure. They also move at the
speed of mail, and that idea has already been had <link>.
So, let's use artillery. A 155mm shell can fit a couple of
petabytes using modern SSD densities. They average about
1mile/sec with a node to node rage of about 15miles,
which is just fine for well-populated regions. The shells
care not about roads, and artillery doctrine is so mature at
this point, that all the creases caused by temperature,
weather, etc. were ironed out decades ago. Artillery, if
not automated can be controlled by labor of negligible
quality.
The tricky part would be catching the shell, but with a
little optimization, I reckon a shell could be aimed at a
cone-shaped hole with some clever magnetic
guidance/braking and read or re-packaged onward
transmission. As an example, with a 20s write/read time I
could get my 458gigs of imaging data to NY in <5 minutes.
A vast improvement.
The military will love the system: they can build an ad-hoc
network anywhere in the world with minimal
infrastructure. Better yet, artillery shells in flight are very
difficult to intercept by foreign nose-pokerinners.
Especially if they want to look at the data without
detection, with a little work, military engineers could build
a layered system, perhaps using ICBMs with MIRV data
heads as a high capacity backbone, or .50 cal rounds with
USB-c plugs on the back for tactical comms.
High Bandwidth Data Transfer
https://what-if.xkcd.com/31/ [bs0u0155, Oct 29 2019]
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:
|
|
It's going to be an awfully loud ping. |
|
|
//It's going to be an awfully loud ping// |
|
|
I prefer "robust network utilization monitoring system"
which has a scale ranging from 0-Verdun. |
|
|
// can be controlled by labor of negligible quality. // |
|
|
For Despatch, fine - but hiring for Goods Inwards may be challenging. |
|
|
// The tricky part would be catching the shell, // |
|
|
A terminal guidance system that deploys a drogue chute for the "last mile" will do fine. |
|
|
I need a bit more re-useability built in. I'm going for a sabot
round so the same round can be re-packaged and fired
again. Maybe some rigid airbrakes? |
|
|
Sounds like you need to read up on Dr. Gerald Bull's work on martlets. |
|
|
I hope he has pictures. I've always found them very difficult
to distinguish from swifts. |
|
|
It's not too difficult. Swifts build mud nests under the eaves of houses, whereas martlets smash through the roof and settle on the ground floor in a heap of wreckage. |
|
|
Could machine-gun fire be modulated to carry a data steam? |
|
|
Great aerodynamics and terminal ballistics on those swifts
though.... variable geometry wings for sustained 5g turns
AND great glide ratios AND dive performance. |
|
|
Surely part of the problem is that fibre connections typically
have only a single fibre. It can't be that hard to make a fibre
bundle and associated hardware with, say, 2^20 fibres. |
|
|
//associated hardware with, say, 2^20 fibres// |
|
|
You can fire several signals down one fiber, you can mess
with color and phase to get many. 2^20 fibers might be OK,
but the same number of switchs/ amplifiers/ multiplexers? |
|
|
Sure. Why not? There are plenty of optics-based chips that
work with millions of elements. |
|
|
<Obligatory Blade Runner reference/> |
|
|
"Eyes ... I just do eyes ..." |
|
|
// Could machine-gun fire be modulated to carry a data steam? // |
|
|
Yes, certainly. The message is limited to "Now we are trying to kill you ... now we're not" from the mark/space of bursts, but yes. |
|
|
But what about those pneumatic tubes... A kind of hyperloop for data
between large data centers. |
|
|
Maybe the answer is less data. My computer now has 2 million
times the memory of my first computer. My data files are
now typically 1 million times bigger than stuff I used to work
with. My internet connection is about 850 times faster than
the first dial-up connection. Yet, strangely, I am perhaps only
two or three times happier than I used to be. This suggests
that we have reached a point of diminishing returns. |
|
|
Not entirely nonsensical.. |
|
|
A fully (10^-9 torr) evacuated tube, using MagLev for suspension and propulsion, actually has no theoretical upper limit to velocity other than the speed of light. It's a linear accelerator (a mature technology ) but for a physical object; and that object requires neither an internal energy source, nor reaction mass. |
|
|
Now, let us consider BUCHANAN, STOP THAT, IT'S DISGUSTING the energetic of such a system. If an evacuated tube were constructed from, say, Frederickton to Vancouver, that distance is a little over 4000km. As a test system this is ideal, as the only thing between those two points is Canada which contains nothing of interest or value so in the event of a failure there's nothing to damage. |
|
|
The "shuttle" spends the first 2000 km accelerating, and the last 2000 km undergoing regenerative braking. |
|
|
Given an acceleration of 10g, what is the peak velocity ? |
|
|
Using Newton's laws, what is the transit time ? |
|
|
What is the peak kinetic energy of a 1kg shuttle ? |
|
|
19,809m/s peak. That's pretty fast.
403.9s total (assuming 10g deceleration as well).
196.2MJ. That's also a lot... |
|
|
If you're going that far, you could do a "straight line" through
the earth, and (since it's the "physicists favourite"
frictionless vacuum), a little kick at the start and it will
"fall" all the way to the end. |
|
|
//Yet, strangely, I am perhaps only two or three times
happier than I used to be. This suggests that we
have reached a point of diminishing returns.// |
|
|
However, if you were to return to those dialup
speeds, I can almost guarantee that you will not be
at the same level of satisfaction as you were when
you originally had them. |
|
| |