wrote:
wondering how the PC 'Atomic Clock' sync utility pgm from
worldtimeserver.com works? How can it sync over the ethernet when
xmit/rcv times are random variables? Or, can it only sync to nearest
second? Anyone know the algorithm used?



Before I read the other replies, and strictly off the top of my head I
would think computer A & B would 'ping'* each other a few times to get
the lag time established (averaged) and then A would send B the time,
adjusted for the trouble getting through the routers and wire along the way.

Dems me thinks.

TBerk
*In this context ping refers to a command sent from one device to
another with a corresponding reply; "are you there?, Yes I am, and it
took x amount of milliseconds to get your answer..."
Think of it like those sonar pings in the submarine biz.

"T" wrote in message
m...
wrote:
wondering how the PC 'Atomic Clock' sync utility pgm from
worldtimeserver.com works? How can it sync over the ethernet when
xmit/rcv times are random variables? Or, can it only sync to nearest
second? Anyone know the algorithm used?



Before I read the other replies, and strictly off the top of my head I
would think computer A & B would 'ping'* each other a few times to get the
lag time established (averaged) and then A would send B the time, adjusted
for the trouble getting through the routers and wire along the way.

Dems me thinks.

TBerk
*In this context ping refers to a command sent from one device to another
with a corresponding reply; "are you there?, Yes I am, and it took x
amount of milliseconds to get your answer..."
Think of it like those sonar pings in the submarine biz.

If you add the current timestamp to the message, you
are getting close. This explains in a lot more detail:

http://www.eecis.udel.edu/~mills/dat.../arch/arch.pdf

George

"George Dishman" wrote in message ...
"T" wrote in message
m...
wrote:
wondering how the PC 'Atomic Clock' sync utility pgm from
worldtimeserver.com works? How can it sync over the ethernet when
xmit/rcv times are random variables? Or, can it only sync to nearest
second? Anyone know the algorithm used?



Before I read the other replies, and strictly off the top of my head I
would think computer A & B would 'ping'* each other a few times to get the
lag time established (averaged) and then A would send B the time, adjusted
for the trouble getting through the routers and wire along the way.

Dems me thinks.

TBerk
*In this context ping refers to a command sent from one device to another
with a corresponding reply; "are you there?, Yes I am, and it took x
amount of milliseconds to get your answer..."
Think of it like those sonar pings in the submarine biz.

If you add the current timestamp to the message, you
are getting close. This explains in a lot more detail:

http://www.eecis.udel.edu/~mills/dat.../arch/arch.pdf

George

Hi
This stuff is way beyond me technically although I have downloaded the
paper
but I have had a idea for a while that the perfect sort of privacy
encryption could be based on how long it takes the message from your
home pc to the bank
that would supply a time amount that would be used as the encryption
key.
Anyone else trying to tap in would always have a different time path
being at a different physical location. In a nutshell your pc sends a
message to the bank server with a time sent as part of the key. Your
pcs clock would have to be very accurate . Then the banks pc would
measure how long it took and use that as the encryption key and
encrypt the message and send it back to you.
Then your pc would then work out what the time difference is (as long
as you knew the delay time at the banks pc for send/recieve.etc.
Anyways the internet probably makes this not feasable as its always
diferent routes during a message
Sean

In sci.physics sean wrote:

snip

Hi
This stuff is way beyond me technically although I have downloaded the
paper
but I have had a idea for a while that the perfect sort of privacy
encryption could be based on how long it takes the message from your
home pc to the bank
that would supply a time amount that would be used as the encryption
key.
Anyone else trying to tap in would always have a different time path
being at a different physical location. In a nutshell your pc sends a
message to the bank server with a time sent as part of the key. Your
pcs clock would have to be very accurate . Then the banks pc would
measure how long it took and use that as the encryption key and
encrypt the message and send it back to you.
Then your pc would then work out what the time difference is (as long
as you knew the delay time at the banks pc for send/recieve.etc.
Anyways the internet probably makes this not feasable as its always
diferent routes during a message
Sean

Nope, usually it is the same route, but it wouldn't work anyway since
the time over the same physical route is normally not a constant either.

--
Jim Pennino

Remove .spam.sux to reply.