$1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Eugen Leitl

http://www.extremetech.com/extreme/122989-1-5-billion-the-cost-of-cutting-london-toyko-latency-by-60ms

$1.5 billion: The cost of cutting London-Tokyo latency by 60ms

By Sebastian Anthony on March 20, 2012 at 1:04 pm

Arctic Link submarine cable

Starting this summer, a convoy of ice breakers and specially-adapted polar
ice-rated cable laying ships will begin to lay the first ever trans-Arctic
Ocean submarine fiber optic cables. Two of these cables, called Artic Fibre
and Arctic Link, will cross the Northwest Passage which runs through the
Canadian Arctic Archipelago. A third cable, the Russian Optical Trans-Arctic
Submarine Cable System (ROTACS), will skirt the north coast of Scandinavia
and Russia. All three cables will connect the United Kingdom to Japan, with a
smattering of branches that will provide high-speed internet access to a
handful of Arctic Circle communities. The completed cables are estimated to
cost between $600 million and $1.5 billion each.

All three cables are being laid for the same reasons: Redundancy and speed.
As it stands, it takes roughly 230 milliseconds for a packet to go from
London to Tokyo; the new cables will reduce this by 30% to 170ms. This
speed-up will be gained by virtue of a much shorter run: Currently, packets
from the UK to Japan either have to traverse Europe, the Middle East, and the
Indian Ocean, or the Atlantic, US, and Pacific, both routes racking up around
15,000 miles in the process. It’s only 10,000 miles (16,000km) across the
Arctic Ocean, and you don’t have to mess around with any land crossings,
either.

Russian Optical Trans-Arctic Submarine Cable System (ROTACS) between UK and
JapanThe massive drop in latency is expected to supercharge algorithmic stock
market trading, where a difference of a few milliseconds can gain (or lose)
millions of dollars. It is for this reason that a new cable is currently
being laid between the UK and US — it will cost $300 million and shave “just”
six milliseconds off the fastest link currently available. The lower latency
will also be a boon to other technologies that hinge heavily on the internet,
such as telemedicine (and teleconferencing) and education. Telephone calls
and live news coverage would also enjoy the significantly lower latency. Each
of the fiber optic cables will have a capacity in the terabits-per-second
range, which will probably come in handy too.

Beyond the stock markets, though, the main advantage of the three new cables
is added redundancy. Currently, almost every cable that lands in Asia goes
through a choke point in the Middle East or the Luzon Strait between the
Philippine and South China seas. If a ship were to drag an anchor across the
wrong patch of seabed, billions of people could wake up to find themselves
either completely disconnected from the internet or surfing with dial-up-like
speeds. The three new cables will all come down from the north of Japan,
through the relatively-empty Bering Sea — and the Arctic Ocean, where each of
the cables will run for more than 5,000 miles, is one of the least-trafficked
parts of the world. That said, the cables will still have to be laid hundreds
of meters below the surface to avoid the tails of roving icebergs.

The ROTACS cable path

Each cable will be laid by a pair of ships: an ice breaker that leads the
way, and a cable ship. Until now it has been impossible to lay cables in the
Arctic Ocean, but the retreat of the Arctic sea ice means that the Northwest
Passage is now generally ice-free from August to October; a big enough window
that cable can be laid fairly safely. Existing cable ships (and there aren’t
many of them) are all outfitted for balmier climes, so all three cables will
require the use of a polar ice-rated ship that has been retrofitted to carry
cable-laying gear.

Read more about the secret world of submarine cables.

For more information on the Russian Optical Trans-Arctic Submarine Cable
System (ROTACS), check out the Polarnet Project (machine translated).

The Arctic Fibre and Arctic Link websites have information on the North
American cables.

[Image credit: New Scientist]



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Anurag Bhatia
Yeah this is super cool!

I hope ISPs will peer well once cable is ready!

(Sent from my mobile device)

Anurag Bhatia
http://anuragbhatia.com
On Mar 23, 2012 5:24 PM, "Eugen Leitl"  wrote:

>
>
> http://www.extremetech.com/extreme/122989-1-5-billion-the-cost-of-cutting-london-toyko-latency-by-60ms
>
> $1.5 billion: The cost of cutting London-Tokyo latency by 60ms
>
>By Sebastian Anthony on March 20, 2012 at 1:04 pm
>
> Arctic Link submarine cable
>
> Starting this summer, a convoy of ice breakers and specially-adapted polar
> ice-rated cable laying ships will begin to lay the first ever trans-Arctic
> Ocean submarine fiber optic cables. Two of these cables, called Artic Fibre
> and Arctic Link, will cross the Northwest Passage which runs through the
> Canadian Arctic Archipelago. A third cable, the Russian Optical
> Trans-Arctic
> Submarine Cable System (ROTACS), will skirt the north coast of Scandinavia
> and Russia. All three cables will connect the United Kingdom to Japan,
> with a
> smattering of branches that will provide high-speed internet access to a
> handful of Arctic Circle communities. The completed cables are estimated to
> cost between $600 million and $1.5 billion each.
>
> All three cables are being laid for the same reasons: Redundancy and speed.
> As it stands, it takes roughly 230 milliseconds for a packet to go from
> London to Tokyo; the new cables will reduce this by 30% to 170ms. This
> speed-up will be gained by virtue of a much shorter run: Currently, packets
> from the UK to Japan either have to traverse Europe, the Middle East, and
> the
> Indian Ocean, or the Atlantic, US, and Pacific, both routes racking up
> around
> 15,000 miles in the process. It’s only 10,000 miles (16,000km) across the
> Arctic Ocean, and you don’t have to mess around with any land crossings,
> either.
>
> Russian Optical Trans-Arctic Submarine Cable System (ROTACS) between UK and
> JapanThe massive drop in latency is expected to supercharge algorithmic
> stock
> market trading, where a difference of a few milliseconds can gain (or lose)
> millions of dollars. It is for this reason that a new cable is currently
> being laid between the UK and US — it will cost $300 million and shave
> “just”
> six milliseconds off the fastest link currently available. The lower
> latency
> will also be a boon to other technologies that hinge heavily on the
> internet,
> such as telemedicine (and teleconferencing) and education. Telephone calls
> and live news coverage would also enjoy the significantly lower latency.
> Each
> of the fiber optic cables will have a capacity in the terabits-per-second
> range, which will probably come in handy too.
>
> Beyond the stock markets, though, the main advantage of the three new
> cables
> is added redundancy. Currently, almost every cable that lands in Asia goes
> through a choke point in the Middle East or the Luzon Strait between the
> Philippine and South China seas. If a ship were to drag an anchor across
> the
> wrong patch of seabed, billions of people could wake up to find themselves
> either completely disconnected from the internet or surfing with
> dial-up-like
> speeds. The three new cables will all come down from the north of Japan,
> through the relatively-empty Bering Sea — and the Arctic Ocean, where each
> of
> the cables will run for more than 5,000 miles, is one of the
> least-trafficked
> parts of the world. That said, the cables will still have to be laid
> hundreds
> of meters below the surface to avoid the tails of roving icebergs.
>
> The ROTACS cable path
>
> Each cable will be laid by a pair of ships: an ice breaker that leads the
> way, and a cable ship. Until now it has been impossible to lay cables in
> the
> Arctic Ocean, but the retreat of the Arctic sea ice means that the
> Northwest
> Passage is now generally ice-free from August to October; a big enough
> window
> that cable can be laid fairly safely. Existing cable ships (and there
> aren’t
> many of them) are all outfitted for balmier climes, so all three cables
> will
> require the use of a polar ice-rated ship that has been retrofitted to
> carry
> cable-laying gear.
>
> Read more about the secret world of submarine cables.
>
> For more information on the Russian Optical Trans-Arctic Submarine Cable
> System (ROTACS), check out the Polarnet Project (machine translated).
>
> The Arctic Fibre and Arctic Link websites have information on the North
> American cables.
>
> [Image credit: New Scientist]
>
>


Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Aled Morris
On 23 March 2012 11:53, Eugen Leitl  wrote:

> All three cables are being laid for the same reasons: Redundancy and speed.
> As it stands, it takes roughly 230 milliseconds for a packet to go from
> London to Tokyo; the new cables will reduce this by 30% to 170ms. This
> speed-up will be gained by virtue of a much shorter run:




If they could armor the cable sufficiently perhaps they could drill the
straigh line path through the Earth's crust (mantle and outer core) and do
London-Tokyo in less than 10,000km.

Aled


RE: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Vitkovsky, Adam
That is why there's this neutrinos project
It's not faster than the speed of light though it can shoot through the Earth 
and no cables cost involved

So far the speed is 0.1 bit per sec

Can't wait for the neutrino SFPs :)

adam

-Original Message-
From: Aled Morris [mailto:al...@qix.co.uk] 
Sent: Friday, March 23, 2012 1:31 PM
To: Eugen Leitl
Cc: NANOG list
Subject: Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

On 23 March 2012 11:53, Eugen Leitl  wrote:

> All three cables are being laid for the same reasons: Redundancy and speed.
> As it stands, it takes roughly 230 milliseconds for a packet to go from
> London to Tokyo; the new cables will reduce this by 30% to 170ms. This
> speed-up will be gained by virtue of a much shorter run:




If they could armor the cable sufficiently perhaps they could drill the
straigh line path through the Earth's crust (mantle and outer core) and do
London-Tokyo in less than 10,000km.

Aled



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Phil Regnauld
Vitkovsky, Adam (avitkovsky) writes:
> 
> Can't wait for the neutrino SFPs :)

You know the shipping cost on a 2 light year thick lead SFP ?




RE: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Leigh Porter


> -Original Message-
> From: Vitkovsky, Adam [mailto:avitkov...@emea.att.com]
> Sent: 23 March 2012 12:57
> To: Aled Morris; Eugen Leitl
> Cc: NANOG list
> Subject: RE: $1.5 billion: The cost of cutting London-Tokyo latency by
> 60ms
> 
> That is why there's this neutrinos project It's not faster than the
> speed of light though it can shoot through the Earth and no cables cost
> involved
> 
> So far the speed is 0.1 bit per sec
> 
> Can't wait for the neutrino SFPs :)
> 
> adam
> 


Nooo, we just need Interocitors!

http://en.wikipedia.org/wiki/Interocitor

---
Leigh


__
This email has been scanned by the Symantec Email Security.cloud service.
For more information please visit http://www.symanteccloud.com
__



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Valdis . Kletnieks
On Fri, 23 Mar 2012 12:53:45 +0100, Eugen Leitl said:
> http://www.extremetech.com/extreme/122989-1-5-billion-the-cost-of-cutting-london-toyko-latency-by-60ms

Lower latency is good...

> The massive drop in latency is expected to supercharge algorithmic stock
> market trading, where a difference of a few milliseconds can gain (or lose)
> millions of dollars.

But it should be illegal to run a stock market that volatile.  This can't end 
well.



pgpf7PEVyqUE6.pgp
Description: PGP signature


RE: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Vitkovsky, Adam
Or
http://en.wikipedia.org/wiki/Ansible
That's what came to my mind when I first heard about quantum entanglement just 
to learn that there's really small chance we could ever use it for communication

adam

-Original Message-
From: Leigh Porter [mailto:leigh.por...@ukbroadband.com] 
Sent: Friday, March 23, 2012 2:20 PM
To: Vitkovsky, Adam; Aled Morris; Eugen Leitl
Cc: NANOG list
Subject: RE: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms



> -Original Message-
> From: Vitkovsky, Adam [mailto:avitkov...@emea.att.com]
> Sent: 23 March 2012 12:57
> To: Aled Morris; Eugen Leitl
> Cc: NANOG list
> Subject: RE: $1.5 billion: The cost of cutting London-Tokyo latency by
> 60ms
> 
> That is why there's this neutrinos project It's not faster than the
> speed of light though it can shoot through the Earth and no cables cost
> involved
> 
> So far the speed is 0.1 bit per sec
> 
> Can't wait for the neutrino SFPs :)
> 
> adam
> 


Nooo, we just need Interocitors!

http://en.wikipedia.org/wiki/Interocitor

---
Leigh


__
This email has been scanned by the Symantec Email Security.cloud service.
For more information please visit http://www.symanteccloud.com
__



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Joel jaeggli
On 3/23/12 14:47 , valdis.kletni...@vt.edu wrote:
> On Fri, 23 Mar 2012 12:53:45 +0100, Eugen Leitl said:
>> http://www.extremetech.com/extreme/122989-1-5-billion-the-cost-of-cutting-london-toyko-latency-by-60ms
> 
> Lower latency is good...
> 
>> The massive drop in latency is expected to supercharge algorithmic stock
>> market trading, where a difference of a few milliseconds can gain (or lose)
>> millions of dollars.
> 
> But it should be illegal to run a stock market that volatile.  This can't end 
> well.

Notwithstanding how bad an idea high speed trading from the vantage
point of those who don't participate in it, 60ms would place you at a
competitive disadvantage to traders that are collocated at or near the
exchange, such that if you're engaged in an arbitrage activity between
two markets someone can frontrun your front-running.





Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Nick Hilliard
On 23/03/2012 15:16, Joel jaeggli wrote:
> Notwithstanding how bad an idea high speed trading from the vantage
> point of those who don't participate in it, 60ms would place you at a
> competitive disadvantage to traders that are collocated at or near the
> exchange, such that if you're engaged in an arbitrage activity between
> two markets someone can frontrun your front-running.

I'd be quite interested in seeing the MTTR for a sub-ice cable break which
happened in late october.

Nick



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Brandon Butterworth
> I'd be quite interested in seeing the MTTR for a sub-ice cable break which
> happened in late october.

More fun too when we get global warming under control and there's no
longer any way to reach it

brandon



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Barry Shein

What about those -- I assume successful -- experiments to fire
neutrinos straight through the earth as a communications medium?

Not sure what the bandwidth of a neutrino stream is.


On March 23, 2012 at 12:31 al...@qix.co.uk (Aled Morris) wrote:
 > On 23 March 2012 11:53, Eugen Leitl  wrote:
 > 
 > > All three cables are being laid for the same reasons: Redundancy and speed.
 > > As it stands, it takes roughly 230 milliseconds for a packet to go from
 > > London to Tokyo; the new cables will reduce this by 30% to 170ms. This
 > > speed-up will be gained by virtue of a much shorter run:
 > 
 > 
 > 
 > 
 > If they could armor the cable sufficiently perhaps they could drill the
 > straigh line path through the Earth's crust (mantle and outer core) and do
 > London-Tokyo in less than 10,000km.
 > 
 > Aled

-- 
-Barry Shein

The World  | b...@theworld.com   | http://www.TheWorld.com
Purveyors to the Trade | Voice: 800-THE-WRLD| Dial-Up: US, PR, Canada
Software Tool & Die| Public Access Internet | SINCE 1989 *oo*



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread ポール・ロラン
Hello,

On Fri, 23 Mar 2012 15:52:21 +
Nick Hilliard  wrote:

> I'd be quite interested in seeing the MTTR for a sub-ice cable break which
> happened in late october.

Maybe that's the reason they want to build three with different paths ;)

Paul


signature.asc
Description: PGP signature


Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Valdis . Kletnieks
On Fri, 23 Mar 2012 15:56:46 -, Brandon Butterworth said:
> > I'd be quite interested in seeing the MTTR for a sub-ice cable break which
> > happened in late october.
>
> More fun too when we get global warming under control and there's no
> longer any way to reach it

Submarines.  It's allegedly been done before, and will probably be done again.


pgpyqwLA1DggE.pgp
Description: PGP signature


Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread George Herbert
On Fri, Mar 23, 2012 at 10:21 AM,   wrote:
> On Fri, 23 Mar 2012 15:56:46 -, Brandon Butterworth said:
>> > I'd be quite interested in seeing the MTTR for a sub-ice cable break which
>> > happened in late october.
>>
>> More fun too when we get global warming under control and there's no
>> longer any way to reach it
>
> Submarines.  It's allegedly been done before, and will probably be done again.

No allegedly about it, though it's not officially acknowledged.  See
Sontag's "Blind Man's Bluff".

However, that was in shallow water, where the sub could rest on the
seabottom next to the cable and divers could exit and work on the
cable outside the sub to tap it.  The subs involved can't dive to the
depths that the cables in question will be mostly laid at, and those
exceed reasonable diver operations depths as well.

One could fix this situation, but it would probably have to be a (low
end) nuclear power plant and a very custom deep submergence hull, and
probably on the order of as expensive as the combined cables cost to
lay.  Probably easier to run redundant cables and fix it come the next
spring...


-- 
-george william herbert
george.herb...@gmail.com



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Marshall Eubanks
On Fri, Mar 23, 2012 at 8:53 AM, Vitkovsky, Adam
 wrote:
> That is why there's this neutrinos project
> It's not faster than the speed of light though it can shoot through the Earth 
> and no cables cost involved
>
> So far the speed is 0.1 bit per sec
>

I bet for $ 1.5 billion neutrino communication (anywhere on Earth) to
its antipode in about 40 msec one way) could be
developed (i.e., the bit rate improved), and I could see some real
market advantages to anyone who had access to it, even
at 100 kbps type bit rates.

Given that, I wouldn't be too surprised to see some physicists and
networking people quietly being hired away by an
obscure new venture...

Regards
Marshall

> Can't wait for the neutrino SFPs :)
>
> adam
>
> -Original Message-
> From: Aled Morris [mailto:al...@qix.co.uk]
> Sent: Friday, March 23, 2012 1:31 PM
> To: Eugen Leitl
> Cc: NANOG list
> Subject: Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms
>
> On 23 March 2012 11:53, Eugen Leitl  wrote:
>
>> All three cables are being laid for the same reasons: Redundancy and speed.
>> As it stands, it takes roughly 230 milliseconds for a packet to go from
>> London to Tokyo; the new cables will reduce this by 30% to 170ms. This
>> speed-up will be gained by virtue of a much shorter run:
>
>
>
>
> If they could armor the cable sufficiently perhaps they could drill the
> straigh line path through the Earth's crust (mantle and outer core) and do
> London-Tokyo in less than 10,000km.
>
> Aled
>



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Jeroen van Aart

valdis.kletni...@vt.edu wrote:

The massive drop in latency is expected to supercharge algorithmic stock
market trading, where a difference of a few milliseconds can gain (or lose)
millions of dollars.


But it should be illegal to run a stock market that volatile.  This can't end 
well.


The average consumer gets a 15 minute artificial delay in trading, why 
not implement for all trades...


--
Earthquake Magnitude: 4.8
Date: Friday, March 23, 2012 14:35:31 UTC
Location: Tonga
Latitude: -16.2478; Longitude: -174.0706
Depth: 119.50 km



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread George Herbert
The physics is not conducive to improving the situation a lot.

There's probably $1.5 billion in the ground already in neutrino
detectors; the total combined detector bit rate is pretty poor.  One
experiment looking at neutrinos coming off the Fermilab accelerator
had 473 million accelerator pulses with under 1.1 million detected
neutrinos.



On Fri, Mar 23, 2012 at 10:45 AM, Marshall Eubanks
 wrote:
> On Fri, Mar 23, 2012 at 8:53 AM, Vitkovsky, Adam
>  wrote:
>> That is why there's this neutrinos project
>> It's not faster than the speed of light though it can shoot through the 
>> Earth and no cables cost involved
>>
>> So far the speed is 0.1 bit per sec
>>
>
> I bet for $ 1.5 billion neutrino communication (anywhere on Earth) to
> its antipode in about 40 msec one way) could be
> developed (i.e., the bit rate improved), and I could see some real
> market advantages to anyone who had access to it, even
> at 100 kbps type bit rates.
>
> Given that, I wouldn't be too surprised to see some physicists and
> networking people quietly being hired away by an
> obscure new venture...
>
> Regards
> Marshall
>
>> Can't wait for the neutrino SFPs :)
>>
>> adam
>>
>> -Original Message-
>> From: Aled Morris [mailto:al...@qix.co.uk]
>> Sent: Friday, March 23, 2012 1:31 PM
>> To: Eugen Leitl
>> Cc: NANOG list
>> Subject: Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms
>>
>> On 23 March 2012 11:53, Eugen Leitl  wrote:
>>
>>> All three cables are being laid for the same reasons: Redundancy and speed.
>>> As it stands, it takes roughly 230 milliseconds for a packet to go from
>>> London to Tokyo; the new cables will reduce this by 30% to 170ms. This
>>> speed-up will be gained by virtue of a much shorter run:
>>
>>
>>
>>
>> If they could armor the cable sufficiently perhaps they could drill the
>> straigh line path through the Earth's crust (mantle and outer core) and do
>> London-Tokyo in less than 10,000km.
>>
>> Aled
>>
>



-- 
-george william herbert
george.herb...@gmail.com



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Jay Ashworth
- Original Message -
> From: "Phil Regnauld" 

> Subject: Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms
> Vitkovsky, Adam (avitkovsky) writes:
> >
> > Can't wait for the neutrino SFPs :)
> 
> You know the shipping cost on a 2 light year thick lead SFP ?

Ah... *here's* the Whacky Weekend thread.  I was wondering where it was.

Cheers,
-- jra
-- 
Jay R. Ashworth  Baylink   j...@baylink.com
Designer The Things I Think   RFC 2100
Ashworth & Associates http://baylink.pitas.com 2000 Land Rover DII
St Petersburg FL USA  http://photo.imageinc.us +1 727 647 1274



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Valdis . Kletnieks
On Fri, 23 Mar 2012 13:16:59 -0700, George Herbert said:
> The physics is not conducive to improving the situation a lot.
>
> There's probably $1.5 billion in the ground already in neutrino
> detectors; the total combined detector bit rate is pretty poor.  One
> experiment looking at neutrinos coming off the Fermilab accelerator
> had 473 million accelerator pulses with under 1.1 million detected
> neutrinos.

Note that each pulse was probably millions or even billions of neutrinos, so
the detection rate was even worse than you'd think.  I saw a statistic that
every second, 50 trillion neutrinos pass through your body.  And the number
that will interact is well into the single digits.



pgp1dGSYhiBGu.pgp
Description: PGP signature


Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Simon Lyall


You guys joke but here is n little article from last week on the current 
state of Neutrino communications:


http://www.economist.com/node/21550242

"The neutrinos themselves are created by smashing bunches of protons into 
a target made of graphite. They are detected roughly 1km away by researchers 
[..] . By modulating the pulses of protons the group was able to send a 
message in binary that, when translated, read “neutrino”.   "



--
Simon Lyall  |  Very Busy  |  Web: http://www.darkmere.gen.nz/
"To stay awake all night adds a day to your life" - Stilgar | eMT.


Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Robert Bonomi

Jeroen van Aart  wrote:
> valdis.kletni...@vt.edu wrote:
> >> The massive drop in latency is expected to supercharge algorithmic stock
> >> market trading, where a difference of a few milliseconds can gain (or lose)
> >> millions of dollars.
> > 
> > But it should be illegal to run a stock market that volatile.  This can't 
> > end well.
>
> The average consumer gets a 15 minute artificial delay in trading, why 
> not implement for all trades...

Virtually any consumer can get true real-time trading data if they're willing
to pay some relatively modest fees for that access -- Last I knew, the most
expensive 'real-time' fee charged by any exchange was under $200/mo.  For
everything traded on that exchange.  For anybody doing short-term, 'tactical',
trading, that is a "petty cash" expense.

Imposing the 15-minute delay on 'everybody', would simply give the 'floor
traders' the -exclusive' edge on those trading strategies.



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread George Herbert
>From the abstract:  "The link achieved a decoded data rate of 0.1
bits/sec with a bit error rate of 1% over a distance of 1.035 km,
including 240 m of earth."

http://arxiv.org/pdf/1203.2847v1.pdf

For practical communications, at longer distances, you probably lose
beam intensity as a 1/R^2 function (the neutrino beam isn't precisely
collimated), so 1,000 km away it will be 1 millionth as strong, or
0.001 baud, 1 bit per 115.74 days.  At 2,000 km it would be less
than 1 bit per year.

Sure you want to do this? 8-)


On Fri, Mar 23, 2012 at 2:44 PM, Simon Lyall  wrote:
>
> You guys joke but here is n little article from last week on the current
> state of Neutrino communications:
>
> http://www.economist.com/node/21550242
>
> "The neutrinos themselves are created by smashing bunches of protons into a
> target made of graphite. They are detected roughly 1km away by researchers
> [..] . By modulating the pulses of protons the group was able to send a
> message in binary that, when translated, read “neutrino”.   "
>
>
> --
> Simon Lyall  |  Very Busy  |  Web: http://www.darkmere.gen.nz/
> "To stay awake all night adds a day to your life" - Stilgar | eMT.



-- 
-george william herbert
george.herb...@gmail.com



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Christopher Morrow
On Fri, Mar 23, 2012 at 11:52 AM, Nick Hilliard  wrote:
> On 23/03/2012 15:16, Joel jaeggli wrote:
>> Notwithstanding how bad an idea high speed trading from the vantage
>> point of those who don't participate in it, 60ms would place you at a
>> competitive disadvantage to traders that are collocated at or near the
>> exchange, such that if you're engaged in an arbitrage activity between
>> two markets someone can frontrun your front-running.
>
> I'd be quite interested in seeing the MTTR for a sub-ice cable break which
> happened in late october.

hopefully it's harder to drag an anchor then ... so it won't happen? :)



RE: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread George Bonser
> If they could armor the cable sufficiently perhaps they could drill the
> straigh line path through the Earth's crust (mantle and outer core) and
> do London-Tokyo in less than 10,000km.
> 
> Aled

I suggested this once when it was decided that the latency from California to 
the UK was too high and that I should reduce it.  The company wouldn't go for 
it, though.

G




RE: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread George Bonser
> 
> I'd be quite interested in seeing the MTTR for a sub-ice cable break
> which happened in late october.
> 
> Nick

Well, you won't have to worry about people dragging anchor across the cable.  
Other than earthquake or volcanic eruption, I can't imagine what would damage a 
cable that time of year in that location.

It would be interesting if they put some sensors on those cables to monitor 
ocean salinity and temperature at those depths, too.





Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Marshall Eubanks
On Fri, Mar 23, 2012 at 5:14 PM,   wrote:
> On Fri, 23 Mar 2012 13:16:59 -0700, George Herbert said:
>> The physics is not conducive to improving the situation a lot.
>>
>> There's probably $1.5 billion in the ground already in neutrino
>> detectors; the total combined detector bit rate is pretty poor.  One
>> experiment looking at neutrinos coming off the Fermilab accelerator
>> had 473 million accelerator pulses with under 1.1 million detected
>> neutrinos.
>
> Note that each pulse was probably millions or even billions of neutrinos, so
> the detection rate was even worse than you'd think.  I saw a statistic that
> every second, 50 trillion neutrinos pass through your body.  And the number
> that will interact is well into the single digits.
>

Small detection numbers are not, per se, fatal to communication. What
fraction of the photons generated by a GPS satellite are captured by
your phone?

The neutrino interaction rate increases with neutrino energy, and sea
water makes a good neutrino detector. You could, for a billion
dollars, do
a LOT better than they did.

By the way, here is the original paper : http://arxiv.org/pdf/1203.2847v1.pdf

Regards
Marshall



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread George Herbert
On Fri, Mar 23, 2012 at 7:11 PM, Marshall Eubanks
 wrote:
> On Fri, Mar 23, 2012 at 5:14 PM,   wrote:
>> On Fri, 23 Mar 2012 13:16:59 -0700, George Herbert said:
>>> The physics is not conducive to improving the situation a lot.
>>>
>>> There's probably $1.5 billion in the ground already in neutrino
>>> detectors; the total combined detector bit rate is pretty poor.  One
>>> experiment looking at neutrinos coming off the Fermilab accelerator
>>> had 473 million accelerator pulses with under 1.1 million detected
>>> neutrinos.
>>
>> Note that each pulse was probably millions or even billions of neutrinos, so
>> the detection rate was even worse than you'd think.  I saw a statistic that
>> every second, 50 trillion neutrinos pass through your body.  And the number
>> that will interact is well into the single digits.
>>
>
> Small detection numbers are not, per se, fatal to communication. What
> fraction of the photons generated by a GPS satellite are captured by
> your phone?

Much higher fraction than with neutrinos.  Remember their MFPs are
measured in light-years...

> The neutrino interaction rate increases with neutrino energy, and sea
> water makes a good neutrino detector. You could, for a billion
> dollars, do
> a LOT better than they did.

On the detector end, sure.  On the transmitter end, it's just not a
well collimated beam due to physics, and no matter how hard you try
the generation of neutrinos is a low-efficiency process.

> By the way, here is the original paper : http://arxiv.org/pdf/1203.2847v1.pdf

Yep.  I meant to include the URL but forgot.



-- 
-george william herbert
george.herb...@gmail.com



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Joel jaeggli
On 3/23/12 19:45 , Jeroen van Aart wrote:
> valdis.kletni...@vt.edu wrote:
>>> The massive drop in latency is expected to supercharge algorithmic stock
>>> market trading, where a difference of a few milliseconds can gain (or
>>> lose)
>>> millions of dollars.
>>
>> But it should be illegal to run a stock market that volatile.  This
>> can't end well.
> 
> The average consumer gets a 15 minute artificial delay in trading, why

in data, not trading... and that really only applies to the sort of free
feeds you're getting.

Even the average consumer gets their ecn cleared market order filled in
seconds inclusive of order entry.


> not implement for all trades...






Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-23 Thread Joel jaeggli
On 3/24/12 01:32 , George Bonser wrote:
>> If they could armor the cable sufficiently perhaps they could drill the
>> straigh line path through the Earth's crust (mantle and outer core) and
>> do London-Tokyo in less than 10,000km.

Current record depth of a borehole is under 12,500 meters which is a bit
short of the goal.

>> Aled
> 
> I suggested this once when it was decided that the latency from California to 
> the UK was too high and that I should reduce it.  The company wouldn't go for 
> it, though.

Bandwidth delay product has undone many a well laid plan.

> G
> 
> 
> 




Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-24 Thread Joly MacFie
Hey $1.5Bn would get you less than half of Spotify right now, so it seems
like a good deal.

-- 
---
Joly MacFie  218 565 9365 Skype:punkcast
WWWhatsup NYC - http://wwwhatsup.com
 http://pinstand.com - http://punkcast.com
 VP (Admin) - ISOC-NY - http://isoc-ny.org
--
-


Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-24 Thread Nick Hilliard
On 24/03/2012 00:32, George Bonser wrote:
> I suggested this once when it was decided that the latency from
> California to the UK was too high and that I should reduce it.  The
> company wouldn't go for it, though.

I assume they had a practical alternative to your proposition?  Perhaps
making light go faster?

Nick



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-24 Thread Marshall Eubanks
On Sat, Mar 24, 2012 at 12:51 AM, George Herbert
 wrote:
> On Fri, Mar 23, 2012 at 7:11 PM, Marshall Eubanks
>  wrote:
>> On Fri, Mar 23, 2012 at 5:14 PM,   wrote:
>>> On Fri, 23 Mar 2012 13:16:59 -0700, George Herbert said:
 The physics is not conducive to improving the situation a lot.

 There's probably $1.5 billion in the ground already in neutrino
 detectors; the total combined detector bit rate is pretty poor.  One
 experiment looking at neutrinos coming off the Fermilab accelerator
 had 473 million accelerator pulses with under 1.1 million detected
 neutrinos.
>>>
>>> Note that each pulse was probably millions or even billions of neutrinos, so
>>> the detection rate was even worse than you'd think.  I saw a statistic that
>>> every second, 50 trillion neutrinos pass through your body.  And the number
>>> that will interact is well into the single digits.
>>>
>>
>> Small detection numbers are not, per se, fatal to communication. What
>> fraction of the photons generated by a GPS satellite are captured by
>> your phone?
>
> Much higher fraction than with neutrinos.  Remember their MFPs are
> measured in light-years...

Actually, at the energy they used it's more like 0.1 light seconds.

>
>> The neutrino interaction rate increases with neutrino energy, and sea
>> water makes a good neutrino detector. You could, for a billion
>> dollars, do
>> a LOT better than they did.
>
> On the detector end, sure.  On the transmitter end, it's just not a
> well collimated beam due to physics, and no matter how hard you try
> the generation of neutrinos is a low-efficiency process.
>

The beam width was < 2 meters after 1 km, equivalent to ~12 km after 1
Earth radius. The beam can be made tighter by going to
higher energy and using more or better post collision focusing
magnets. The  KM3NeT detector in the Mediterranean will be more
sensitive, 3 km across and will cost order 200 million  euros. With
better magnets, the existing beam could be made to be the size of that
detector at 1 Earth radius. So, existing technology could certainly
communicate across the Atlantic or the Pacific. The real question,
again, would be what it would take to get the bit rate up.

Regards
Marshall


>> By the way, here is the original paper : http://arxiv.org/pdf/1203.2847v1.pdf
>
> Yep.  I meant to include the URL but forgot.
>
>
>
> --
> -george william herbert
> george.herb...@gmail.com



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-26 Thread Tei
On 23 March 2012 13:31, Aled Morris  wrote:
> On 23 March 2012 11:53, Eugen Leitl  wrote:
>
>> All three cables are being laid for the same reasons: Redundancy and speed.
>> As it stands, it takes roughly 230 milliseconds for a packet to go from
>> London to Tokyo; the new cables will reduce this by 30% to 170ms. This
>> speed-up will be gained by virtue of a much shorter run:
>
>
>
>
> If they could armor the cable sufficiently perhaps they could drill the
> straigh line path through the Earth's crust (mantle and outer core) and do
> London-Tokyo in less than 10,000km.
>
> Aled

I imagine a easier solution.  Use a random number generator in both
sides, with the same seed.  Then use a slower way to send "packets
re-sync" that will contain the delta from the generated number, to the
real actual number.

I suppose this speeds are needed for some "fast speed transaction",
that are leeching money from the background noise on the market.

This is not like the Roman empire, where you could make a lot of money
buying wheat wen theres a dry year in egypt.

note: I could be wrong.






-- 
--
ℱin del ℳensaje.



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-26 Thread Joe Loiacono
Tei  wrote on 03/26/2012 06:16:53 AM:

> I imagine a easier solution.  Use a random number generator in both
> sides, with the same seed.  Then use a slower way to send "packets
> re-sync" that will contain the delta from the generated number, to the
> real actual number.
> 
> I suppose this speeds are needed for some "fast speed transaction",
> that are leeching money from the background noise on the market.
> 
> This is not like the Roman empire, where you could make a lot of money
> buying wheat wen theres a dry year in egypt.
> 
> note: I could be wrong.

Noted.

Joe


Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-26 Thread Valdis . Kletnieks
On Mon, 26 Mar 2012 12:16:53 +0200, Tei said:

> I imagine a easier solution.  Use a random number generator in both
> sides, with the same seed.  Then use a slower way to send "packets
> re-sync" that will contain the delta from the generated number, to the
> real actual number.

Congrats. You've just re-invented the crytpo method called "xor with a
pseudorandom bitstream".  And no, it doesn't minimize your round-trip
latency at all.

> I suppose this speeds are needed for some "fast speed transaction",
> that are leeching money from the background noise on the market.

Unfortunately, you are correct on that point.


pgp9boe3wybW3.pgp
Description: PGP signature


Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-26 Thread Rodrick Brown
On Mar 23, 2012, at 2:45 PM, Jeroen van Aart  wrote:

> valdis.kletni...@vt.edu wrote:
>>> The massive drop in latency is expected to supercharge algorithmic stock
>>> market trading, where a difference of a few milliseconds can gain (or lose)
>>> millions of dollars.
>> But it should be illegal to run a stock market that volatile.  This can't 
>> end well.
> 
> The average consumer gets a 15 minute artificial delay in trading, why not 
> implement for all trades...

The average consumer shouldn't be day trading with shit market data thats 
delayed or worse with level 1 depth of the markets they're just asking to be 
taken by the heavy quant firms. 

HIgh frequency trading does provide a service to the financial markets as a 
whole despite what the media and government politicians will have you think. 

Transaction cost has plummeted over the years and do has the barrister to enter 
the markets.

> -- 
> Earthquake Magnitude: 4.8
> Date: Friday, March 23, 2012 14:35:31 UTC
> Location: Tonga
> Latitude: -16.2478; Longitude: -174.0706
> Depth: 119.50 km
> 



Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-26 Thread Valdis . Kletnieks
On Mon, 26 Mar 2012 08:59:34 -0400, Rodrick Brown said:
> HIgh frequency trading does provide a service to the financial markets as a
> whole despite what the media and government politicians will have you think.

OK, I'll bite. What benefit does the market *as a whole* get from the ability
to do trades in 60ms rather than 120ms?  (Hint - the fact you can extract
money via more arbitrage transactions per minute is a benefit to the trader,
not the market as a whole - if you extract $100M from the market, it came
from somewhere)


pgplz1y40kyC5.pgp
Description: PGP signature


Re: $1.5 billion: The cost of cutting London-Tokyo latency by 60ms

2012-03-26 Thread Rodrick Brown
On Mar 26, 2012, at 9:32 AM, valdis.kletni...@vt.edu wrote:

> On Mon, 26 Mar 2012 08:59:34 -0400, Rodrick Brown said:
>> HIgh frequency trading does provide a service to the financial markets as a
>> whole despite what the media and government politicians will have you think.
> 
> OK, I'll bite. What benefit does the market *as a whole* get from the ability
> to do trades in 60ms rather than 120ms?  (Hint - the fact you can extract
> money via more arbitrage transactions per minute is a benefit to the trader,
> not the market as a whole - if you extract $100M from the market, it came
> from somewhere)

In its core very liquid markets and market efficiency. 

The faster a trade executes is the faster the market can recover and self 
correct themselves from trading mishaps. 

Faster speeds has provided higher volume of shares traded which has directly 
lead to higher profits, higher tax income for governments, and the less 
likely-hood of being front-run by dishonest brokers and most of all lower 
transaction cost for the average market precipitant. 

HFT like anything else in the modern world is prime for abuse for anyone 
looking to manipulate the markets by taking advantage of certain rules or 
loopholes that legislation has not yet discovered or plugged. That being said I 
strongly believe high speed trading has done more good than bad for the 
financial markets as a whole. 

Lowering access time to the markets will only open up a new can of worms which 
will easily be circumvented by other loopholes and abused by those with the 
know how! 

Sorry for the off topic rant!